Superfibronectin, a Multimeric Form of Fibronectin, Increases HIV Infection of Primary CD4⁺ T Lymphocytes¹

During the initial phase of HIV infection, viral particles in lymph nodes are predominantly found intracellularly (1). Following this phase, HIV becomes sequestered extracellularly and is found in association with the follicular dendritic cell (FDC)⁵ network (1, 2, 3). This pool of virus makes up the major reservoir of HIV (4, 5). Presentation of viral particles by the FDC network is thought to facilitate de novo infection of T lymphocytes (6, 7, 8, 9, 10). The extracellular environment of lymph node tissues also contains multimeric FN, a major component of the extracellular matrix. Following HIV infection, extensive lymph node remodeling occurs, and FN deposition is increased in lymph node tissues of HIV-infected patients (11, 12). We hypothesized that matrix FN may play a role in the extracellular trapping of HIV particles and facilitate de novo infection of T lymphocytes circulating through the lymphoid tissues.

In vivo, FN is secreted into the plasma as soluble dimeric fibrils. Fibronectin fibrils can then be incorporated into the extracellular matrix as insoluble multimers. The assembly of FN into a matrix is a complex process that requires interaction of FN fibrils with cell surface receptors (13, 14, 15, 16, 17). Fibronectin is a modular protein, and the first type III repeat (III₁-C) plays an important role in matrix assembly in vivo (18, 19). The addition of a recombinant III₁-C fragment to soluble FN results in spontaneous cross-linking of FN dimers in vitro (18, 19) to form superfibronectin (sFN), a multimeric form of FN. The matrix (multimeric) form of FN is structurally and functionally different from plasma (dimeric) FN. sFN closely resembles matrix FN and is functionally different from dimeric plasma FN in that it has enhanced cell-adhesive properties, inhibits cell migration, and displays antimetastatic properties (19, 20, 21). For these reasons, we used sFN, the in vitro generated multimeric form of FN, to study the role of matrix FN in HIV infection of T lymphocytes.

We show that multimeric sFN, but not soluble FN, increased HIV infection of primary CD4⁺ T lymphocytes by 10- to 15-fold. The observed increase in the presence of multimeric FN required binding of HIV particles to sFN. Viral particles were shown to bind efficiently to sFN, and the HIV-1_IIIB/sFN complex had increased adhesion and internalization in T lymphocytes. Interestingly, the III₁-C fragment alone also bound HIV particles effectively and enhanced HIV infection by severalfold. Additionally, we show that the III₁-C epitope of FN is exposed in lymph node tissue and thus is available to interact with extracellular viral particles in vivo. Together, our data support a role for matrix FN in enhancing HIV infectivity in vivo by facilitating and stabilizing the interaction of HIV with its target cell.

PBMC were isolated from whole blood by Ficoll-Hypaque density gradient centrifugation and were depleted of monocytes by several rounds of adherence to tissue culture plastic. CD4⁺ T lymphocytes were isolated from this population using magnetically labeled anti-CD4 Abs (Miltenyi Biotec, Auburn, CA) and were maintained in RPMI 1640 medium supplemented with 5% FCS, penicillin-streptomycin, and 50 U/ml IL-2 (Proleukin, Chiron, Emeryville, CA). Cells were verified as >95% CD4⁺ and CD3⁺ T lymphocytes by flow cytometric analysis. In some experiments CD4⁺ T lymphocytes were activated for 2 days with PHA (5 μg/ml). Unless otherwise noted, all experiments were performed using unstimulated primary CD4⁺ T lymphocytes.

HIV-1_IIIB was purchased from Advanced Biotechniques (Rockville, MD). A 1-μl aliquot of this stock corresponded to ∼6 ng of p24 Ag with a multiplicity of infection of 0.01 when titrated on 1 × 10⁶ human PBMC.

sFN was generated as described by incubating human FN with recombinant III₁-C fragment at a ratio of 1 μM/1 μg/ml FN (18, 19). Purified human FN and recombinant III₁-C fragment and III₁₁-C (control fragment) were obtained as previously described (19, 21). Vitronectin (VN) was prepared from human plasma according to the methods of Yatohgo et al. (22).

Abs P5D2 (anti-β₁) (23) and P1F6 (anti-α_vβ₅) (24) were gifts from Dr. Elizabeth Wayner (University of Minnesota, Minneapolis, MN). L230 (anti-α_v) (24) was prepared from hybridoma cells obtained from the American Type Culture Collection (Manassas, VA). CD18 (anti-β₂) was purchased from Caltag (Burlingame, CA). P1B5 (anti-α₃) (25, 26) was obtained from Telios (San Diego, CA), and P3D10 (anti-α₅) was a gift from Dr. William Carter (Fred Hutchinson Cancer Center, Seattle, WA). All of the above Abs are function-blocking Abs. Abs were used at concentrations that block cell adhesion to matrix proteins. The 12G5 Ab (anti-CXCR4) was a gift from Dr. James Hoxie (University of Pennsylvania, Philadelphia, PA) (27). FITC-conjugated Ab to MHC class II Ag was obtained from Becton Dickinson (Mountain View, CA), FITC-conjugated Ab to CD25 was obtained from PharMingen (San Diego, CA). Ab to CD38 (OKT10) was a gift from Dr. Karen Zier (Mount Sinai School of Medicine, New York, NY).

The anti-III₁-C polyclonal Ab was made by immunizing mice with a histidine-tail III₁-C fusion protein bound to Ni beads (Qiagen, Valencia, CA) (28) to facilitate slow release and Ag presentation by macrophages. Six injections were given biweekly, i.p. and s.c. The first injection and the boosters used 500 μg of protein. The immune response was evaluated by ELISA and Western blot using FN, the III₁-C fragment, and other FN fragments (including III₁₁-C) as controls. Anti-III₁-C polyclonal serum was affinity purified on a protein G column.

Full-length HIV-1_IIIB gp120 and anti- HIV-1_IIIB gp120 Ab were obtained from the AIDS Research and Reference Reagent Program, Division of AIDS, National Institute of Arthritis and Infectious Diseases, National Institutes of Health. HIV-1_IIIB gp120 peptide 295–328(295–328) and HIV-1_IIIB gp120 peptide (418–441) were obtained from the AIDS Research and Reference Reagent Program, Division of AIDS, National Institute of Arthritis and Infectious Diseases, National Institutes of Health, and contributed by Dr. Seth Pincus. Heparan sulfate (bovine intestinal mucosa; m.w., 7500), GRGDSP peptide, heparatinase, and chondroitinase ABC were purchased from Sigma (St. Louis, MO). Endo-β-galactosidase was purchased from Seikagaku (Tokyo, Japan).

To determine HIV infectivity in the presence of immobilized matrix proteins, 24-well tissue culture plates were coated overnight at 4°C with 1% BSA, VN (10 μg/ml), FN (10 μg/ml), sFN (10 μg/ml FN plus 10 μM III₁-C), or III₁-C (10 μM) and blocked with 1% BSA for 30 min at 37°C. Coating efficiencies for FN, III₁-C, and sFN were equal, as determined by ELISA (R. Pasqualini, unpublished observations). HIV-1_IIIB was added to each well in 1-ml aliquots containing ∼6 ng of p24 Ag and were allowed to attach for 15 min at 37°C. Following incubation, wells were washed three times with PBS to remove unbound virus. PHA-stimulated, primary CD4⁺ T lymphocytes were added at 1 × 10⁶ cells/well. Cell-free supernatants were collected at 1, 3, 10, and 17 days postinfection, and the amount of virus was quantitated by p24 ELISA. Data are reported as the mean p24 Ag levels of triplicate wells ± SD (nanograms per milliliter).

In some experiments PHA-stimulated CD4⁺ T lymphocytes were infected at a multiplicity of infection of 0.01. Three days after infection, cells were plated onto 24-well plates coated with III₁-C, sFN, FN, VN, or BSA. Six days after plating, virus production was quantitated in supernatants by p24 ELISA.

Virus was incubated with FN (50 μg), III₁-C fragment (25 μM), or sFN (50 μg FN plus 25 μM III₁-C) for 15 min at room temperature, then mixed with PHA-stimulated CD4⁺ T lymphocytes and incubated at 37°C. After 2 h, cells were washed and incubated in 24-well tissue culture plates. Cell-free supernatants were collected at 1, 4, and 7 days postinfection, and the amount of virus was quantitated by p24 ELISA. Data are reported as the mean p24 Ag levels of triplicate wells ± SD (nanograms per milliliter).

Cells treated as described for HIV infection studies were tested for HIV DNA at 1 and 4 days postinfection. Total cellular DNA was obtained using a Qiagen DNA isolation kit. DNA concentrations were determined by spectrophotometric analysis of samples at 260 nm and ethidium bromide staining after gel electrophoresis. Approximately 500 ng of DNA was amplified from each sample. The presence of HIV-1_IIIB viral DNA was determined using the following primers: sense, 5′-GTGACTCTGGTAACTAGAGA-3′ (nt 477–497); and antisense, 5′-CCACAGATCAAGGATATCTTGTC-3′ (nt 539–516), which amplify a 120-bp product in the long terminal repeat (LTR) coding region of circularized, extrachromosomal HIV DNA. PCR reactions were conducted in 50-μl reaction mixtures containing PCR buffer (Perkin-Elmer, Palo Alto, CA), 2 mM MgCl₂, 0.2 mM dNTPs, 50 pmol of each primer, and 1.25 U of Taq DNA polymerase (Taq Gold, Perkin-Elmer). The reaction mixture was incubated for 10 min at 95°C and subjected to 40 cycles, consisting of 1 min at 94°C, 1 min at 62°C, and 1 min at 72°C, followed by a single cycle of 10 min at 72°C. The PCR-amplified products were analyzed by 2% agarose Tris-acetate gel electrophoresis. Known concentrations of plasmid DNA containing the PCR-generated fragment (10³, 10², 10¹, and 10⁰ copies/reaction) were amplified simultaneously to generate a standard curve. Agarose gel was transferred to Nytran (Schleicher & Schuell, Keene, NH) and then hybridized with a radiolabeled PCR fragment (amplified from control plasmid and labeled with Ready-To-Go dCTP (Pharmacia, Piscataway, NJ)) overnight at 65°C. After the membrane was washed, autoradiography was performed. Densitometry was performed using ImageQuant version 1.11 (Becton Dickinson, Mountain View, CA), and a standard curve was generated by simple linear regression using StatView 4.01 (Abacus Concepts, Berkeley, CA) (r = 0.994). In addition, nested amplification was performed on DNA samples using original primers, sense 5′ primer (CCACAGATCAAGGATATCT TGTC), and antisense 5′ primer (GTGACTCTGGTAACTAGAGA). To verify equal loading of DNA, β-globin sequences were simultaneously amplified from the DNA samples (sense 5′ primer, ACACAACTGTGT TCACTAGC; antisense 5′ primer, CAACTTCATCCACGTTCACC).

Binding of HIV-1_IIIB to immobilized proteins was assessed as follows. Ninety-six-well plates were coated overnight at 4°C with 1% BSA or with increasing concentrations (0.01–180 μg/ml) of FN, sFN, or III₁-C fragment. Before adding virus, all wells were blocked with 1% BSA in PBS. HIV-1_IIIB was added in 100-μl aliquots containing ∼1.2 ng of p24 HIV Ag and was allowed to attach for 20 min at 37°C. Wells were then washed three times with PBS to remove unbound virus. Attached virus was lysed in PBS containing 1% Triton X-100 and was quantitated by p24 ELISA. Data are reported as the mean amount of p24 (nanograms per milliliter) ± SD measured in triplicate samples minus the mean p24 level measured in wells coated with BSA.

To confirm that the III₁-C region was involved in binding of sFN to HIV, adhesion assays were performed as described above, except that wells were incubated with increasing concentrations of III₁-C polyclonal Ab or nonimmune serum for 1 h at 37°C before addition of virus.

To determine attachment of HIV-1_IIIB to cells, HIV-1_IIIB was incubated with FN (10 μg), sFN (10 μg FN plus 10 μM III₁-C), or III₁-C fragment (10 μM) for 15 min at room temperature. The HIV-1_IIIB/protein mix was then added to primary CD4⁺ T lymphocytes and incubated for 30 min at 0°C. This temperature allows attachment of virus to the cell surface, but not internalization (29). Following incubation, cells were washed three times with PBS and lysed in 1% Triton X-100. Virus levels were quantified by p24 ELISA. Results are shown as the fold increase in p24 levels ± SD compared with p24 levels detected on cells exposed to HIV alone.

Virus was incubated with FN (10 μg), III₁-C fragment (10 μM), or sFN (10 μg FN plus 10 μM III₁-C) for 15 min at room temperature, then mixed with CD4⁺ T lymphocytes and incubated at 37°C for 1–2 h. In some experiments, CD4⁺ T lymphocytes were incubated with FN, III₁-C fragment, or sFN for 15 min at room temperature, then mixed with virus and incubated for 1–2 h at 37°C. Following incubation, cells were washed in PBS and treated with trypsin (25 μg/ml) for 20 min at room temperature to remove attached, noninternalized viral particles (29). Cells were lysed in 1% Triton X-100, and the amount of internalized virus was quantitated by p24 ELISA. Results are presented as the fold increase in p24 levels ± SD compared with p24 levels detected in cells exposed to HIV alone.

The cell line 1G5, a Jurkat derivative, contains a stably integrated HIV-LTR-luciferase construct and was obtained from the AIDS Research and Reference Reagent Program, Division of AIDS, National Institute of Arthritis and Infectious Diseases, National Institutes of Health, and contributed by Drs. Estuardo Aguilar-Cordova and John Belmont. Cells were incubated for 4 h in wells coated with FN, sFN, and III₁-C. The preparation and assay of cell extract was performed using the luciferase assay system with reporter lysis buffer (Promega, Madison, WI). Luciferase activity was measured by luminometer (Bio-Orbit, Turko, Finland).

HIV-1_IIIB (∼6 ng) was mixed with increasing concentrations of heparan sulfate (0.001, 0.01, 0.1, and 1 μg/ml) before incubation with sFN (10 μg FN plus 10 μM III₁-C). The HIV-1_IIIB/sFN mix was added to 5 × 10⁵ CD4⁺ T lymphocytes and allowed to attach for 30 min at 0°C. Cells were washed three times in PBS and then lysed in 1% Triton X-100. The number of viral particles bound to the cell surface was quantitated by p24 ELISA. Results are shown as the percentage of viral attachment detected in cells incubated with the HIV-1_IIIB/sFN mix alone.

Ninety-six-well microtiter plates were coated overnight at 4°C with 1% BSA or with increasing concentrations of FN, sFN, or III₁-C fragment. Wells were blocked with 1% BSA in PBS. Full-length HIV-1_IIIB-gp120 (1 μg/ml) was added and allowed to attach for 3 h at 37°C. Unbound protein was removed by washing with PBS/1% BSA. Anti-gp120 Ab (1:1000) was added to wells, allowed to attach for 1 h, and then washed three times with PBS. HRP-conjugated goat anti-mouse Ab (1:1000; Dako, Carpenteria, CA) was added for 2 h at room temperature and then washed with PBS. Color development was performed using the TMB Microwell Peroxidase Substrate System (Kirkegaard & Perry Laboratories, Gaithersburg, MD) and was read at 450 nm. The data were reported as the mean absorbance of triplicate wells ± SD.

To evaluate inhibition by gp120 peptides, HIV-1_IIIB was mixed with increasing concentrations of HIV-1_IIIB gp120 peptide aa 418–441 or peptide aa 295–328 (1.25–125 μg/ml) and then added to 96-well plates coated with III₁-C peptide. Virus particles were allowed to adhere for 1 h at 37°C. Following incubation, wells were washed five times with PBS to remove unbound virus. Attached virus was lysed in PBS containing 1% Triton X-100 and quantitated by p24 ELISA. Data are shown as the mean p24 levels (nanograms per milliliter) ± SD of triplicate samples.

Primary CD4⁺ T lymphocytes (1 × 10⁶) were incubated for 15 min at room temperature with the indicated anti-integrin Abs, GRGDSP peptide (5 mM), or EDTA (10 mM). Cells were then mixed with HIV-1_IIIB/sFN and incubated for 1–2 h at 37°C, washed, treated with trypsin (25 μg/ml) for 20 min at room temperature, and lysed in 1% Triton X-100. The amount of internalized virus particles was quantitated in cell lysates by p24 ELISA. Results are shown as the percentage of viral entry detected in cells incubated with HIV-1_IIIB/sFN mix alone.

Primary CD4⁺ T lymphocytes were washed in PBS, resuspended at 2 × 10⁶/200 μl PBS containing 0.1% glucose and 0.1 mg/ml BSA, and treated with 50 mIU/ml heparitinase, 50 mIU/ml chondroitinase ABC, or 15 mIU/ml endo-β-galactosidase for 1 h at 37°C. Following treatment, cells were washed and incubated with HIV-1_IIIB/sFN mix for 30 min at 0°C, and viral attachment was measured as described.

Dissected rat lymph nodes were fixed in 4% paraformaldehyde followed by immersion in 15% sucrose. Tissues were embedded in OCT (Miles, Elkhart, IN) and frozen in liquid nitrogen-chilled isopentane. Cryostat sections (5 μm) of tissue were air-dried and washed in PBS. Endogenous peroxidase activity was blocked by incubation with Peroxoblock (Zymed, San Francisco, CA) for 45 s. Sections were incubated in 3% BSA for 10 min at room temperature and then incubated with the anti-III₁-C polyclonal mouse serum (1:200) for 1 h at 37°C. As a control, sections were incubated with preimmune serum (1:200). After washing in PBS, sections were incubated with the biotinylated horse anti-mouse preabsorbed in rat (Vector, Burlingame, CA) for 30 min at 37°C, followed by 30-min incubation with avidin-biotinylated HRP (Vectastain ABC kit; Vector Laboratories, Burlingame, CA). The reaction product was visualized by incubation of the sections with diaminobenzidine (DAB Plus kit, Zymed). Sections were counterstained with hematoxylin, sequentially dehydrated in ethanol solutions, transferred to xylene, and mounted with Permount (Fisher, Fairlawn, NJ).

Data were initially analyzed by one-way ANOVA to test for differences involving three or more treatment protocols. Differences between individual conditions were assessed using a post-hoc analysis with Fisher’s protected least significant difference test. The level of statistical significance was set at p < 0.05.

We evaluated the role of immobilized and soluble multimeric FN on the infection of T lymphocytes with HIV. To test the effect of immobilized proteins, 24-well tissue culture plates were coated with various matrix proteins, including sFN, the multimeric form of FN, and were incubated with HIV-1_IIIB followed by washing and the addition of PHA-stimulated CD4⁺ T lymphocytes. Cell-free supernatants were taken at various days postinfection and tested for virus levels by p24 ELISA. By days 10 and 17, wells coated with sFN showed significantly higher p24 levels than wells coated with FN (Fig. 1). No significant levels of p24 were detected in wells coated with VN or BSA. Thus, immobilized multimeric FN significantly enhanced HIV levels. Interestingly, the III₁-C peptide alone, the FN fragment used to generate multimeric FN, also showed a significant increase in virus levels compared with soluble FN on days 10 and 17 postinfection.

To test the effects of soluble proteins, we incubated HIV with soluble sFN, FN, BSA, or III₁-C proteins for 15 min and added the mixture to PHA-stimulated lymphocytes. Accordingly, cells were exposed to equal amounts of virus. Cell-free supernatants were taken on various days postinfection and tested for virus levels by p24 ELISA. By day 4, significantly higher p24 levels were found in sFN-treated cells than in cells treated with soluble FN (Fig. 2 A). p24 levels were undetectable in wells coated with BSA. By day 7, sFN and III₁-C both enhanced infection of lymphocytes by 3- to 4-fold compared with FN or BSA. Thus, incubation of HIV with multimeric FN significantly enhanced the productive infection of lymphocytes compared with the effect of BSA or FN.

To examine whether the internalized viral particles underwent RT under single-cycle conditions, we extracted DNA in parallel with the above infectivity assay on day 1 following infection. We performed PCR with primers corresponding to the HIV-LTR region. These PCR primers are specific for the presence of extrachromosomal, circularized DNA and will not amplify DNA derived from attached, uninternalized virions (30). This was important because virions can contain virus-specific DNA (31, 32). Therefore, amplification products from these primers indicate that viral entry, RT, and nuclear translocation occurred (33). On day 1, a band of the appropriate size (120 bp) was detected only in sFN-treated samples (Fig. 2,B). Approximately 14 copies of viral DNA were present in the sFN sample, based on the standard curve generated from amplification of known copy numbers of control plasmid. To increase the sensitivity, PCR products were reamplified using nested primers. Again, a product was only visualized in the sFN-treated sample; no bands were detected in the BSA-, FN-, or III₁-C-treated samples (data not shown). PCR performed on DNA extracted from non-PHA-stimulated lymphocytes also showed a product only in sFN-treated samples (data not shown). By day 4, the 120-bp PCR product was detected in all samples (Fig. 2,B, right panel). However, the sFN- and III₁-C-treated samples contained ∼2- to 4-fold more copies than FN- or BSA-treated samples. Detectable levels of p24 Ag by ELISA correlated with the presence of HIV-extrachromosomal DNA as analyzed by PCR (Fig. 2). However, PCR analysis detected viral DNA at earlier time points than p24 ELISA assay; PCR results from day 1 predicted p24 detection on day 4, and PCR results from day 4 predicted p24 detection on day 7.

We speculated that the observed increase in virus levels in the presence of immobilized sFN and III₁-C resulted from an initial difference in the number of viral particles presented to T lymphocytes. Therefore, we compared the ability of HIV-1_IIIB particles to bind to surfaces coated with sFN, III₁-C, FN, or BSA. To test this, plates coated with increasing concentrations of proteins were incubated with equal amounts of HIV-1_IIIB. Unbound viral particles were removed by washing, and p24 Ag levels were measured as an index of the number of adherent viral particles. We found a significant increase in the number of viral particles that bound to sFN compared with FN (Fig. 3,A). This difference was evident at concentrations as low as 1 μg/ml and persisted at higher coating concentrations of matrix proteins. HIV-1_IIIB also adhered to III₁-C fragment significantly better than FN (Fig. 3 A).

We hypothesized that the III₁-C region exposed in sFN was involved in virus binding, since the III₁-C peptide supported the adhesion of viral particles. We asked whether an Ab directed against III₁-C region could block the interaction of HIV with sFN. The specificity of the Ab was verified by ELISA. The antisera recognized sFN and the III₁-C peptide, but not FN or BSA (Fig. 3,B). Preimmune serum did not recognize any proteins (data not shown). Ab to III₁-C peptide or nonimmune serum was added to wells precoated with different matrix proteins. Equal amounts of virus were added to wells and allowed to adhere for 1 h. Unbound virus particles were removed by washing, and p24 levels were measured. We found that Ab directed against III₁-C epitope was able to significantly inhibit adhesion of HIV to sFN (and III₁-C) in a dose-dependent manner (Fig. 3 C). We found an 80% inhibition of virus binding to sFN at the highest Ab concentration tested. No significant inhibition of viral adhesion to FN was observed in the presence of III₁-C Ab.

Next, we tested whether the increase in viral infection was the result of increased binding of viral particles complexed with sFN and III₁-C to the cell surface. To test this, HIV-1_IIIB premixed with FN, sFN, or III₁-C fragment was incubated with primary CD4⁺ T lymphocytes for 30 min at 0°C. This temperature allows viral attachment but not internalization (29). Preincubation of virus with sFN significantly increased the number of viral particles attaching to T lymphocytes compared with preincubation of virus with dimeric FN (8- to 10-fold; p < 0.001; Fig. 4 A). Preincubation of virus with III₁-C fragment also resulted in a significant increase in viral attachment compared with dimeric FN (2- to 3-fold; p < 0.04). Thus, despite equal amounts of viral particles presented to cells, exposure of cells to virus in the presence of sFN and III₁-C resulted in increased viral adhesion to the cell surface. As an additional control, HIV/sFN complex did not attach to 293 cells, a human embryonic kidney cell line (data not shown).

To determine whether internalization of virus followed increased viral adhesion, we measured viral uptake by T lymphocytes that were exposed to equal amounts of viral particles in the presence of sFN, III₁-C fragment, and FN. Viral particles were mixed with matrix proteins in solution and then added to lymphocytes for 1–2 h at 37°C to allow viral attachment and internalization. Cells were washed and treated with trypsin to remove uninternalized, attached virus (29). Subsequently, cells were lysed and tested for p24 levels by ELISA to determine the number of internalized viral particles (Fig. 4 B). As shown, incubation of virus with sFN or III₁-C fragment resulted in a significant increase in viral internalization (14 ± 5- and 9 ± 4-fold, respectively). Thus, the increase in the number of viral particles attaching to the cell surface correlated with an increase in viral internalization in the presence of sFN and III₁-C fragment.

In the previous experiments virus was incubated with matrix proteins before incubation with T lymphocytes. We asked whether prior incubation of cells with sFN or III₁-C resulted in enhanced HIV internalization or infection. However, preincubation of cells with either sFN or the III₁-C fragment did not result in enhanced levels of viral internalization (data not shown). Thus, exposure of cells alone to sFN or the III₁-C fragment did not render cells more susceptible to infection with HIV-1_IIIB.

To determine whether sFN or III₁-C was exerting its effect by inducing the activation of CD4⁺ lymphocytes, we asked whether incubation of CD4⁺ lymphocytes with BSA, FN, sFN, or III₁-C resulted in up-regulation of surface markers of T cell activation such as MHC class II (34), CD25 (IL-2Rα, Tac Ag) (35), or CD38 (36). We incubated lymphocytes with BSA, FN, sFN, or III₁-C for 2–4 h and then performed flow cytometry on the cells 24 and 48 h after treatment. As a positive control, cells were treated with PHA alone. We observed low level activation in all cell populations, most likely a result of positive selection and culture conditions. However, the surface profiles of class II, CD25, and CD38 were identical on all cells regardless of whether they were treated with BSA, FN, sFN, or III₁-C (Table I). Thus, the differences we found in infection cannot be attributed to differences in the activation state of the cells.

Another possibility that could contribute to the observed increase in viral production was that adhesion of lymphocytes to matrix proteins increased viral replication by increasing HIV-LTR promoter activity. Cell-extracellular matrix interactions have previously been shown to influence viral replication. For example, monocytes plated on laminin showed increased viral replication, whereas FN had no effect (37). To evaluate the role of sFN on promoter activity, we used a lymphocyte cell line (Jurkat cells) stably transfected with the luciferase reporter gene under control of the HIV-1-LTR promoter. Cells were plated onto various matrix proteins and allowed to interact for 4 h, after which luciferase activity was measured as an indicator of HIV-LTR activity. Interaction of cells with sFN, III₁-C, or FN did not result in up-regulation of HIV-LTR activity (data not shown). As a positive control, cells plated on wells coated with anti-CD3 Ab, which is known to up-regulate HIV-LTR activity, showed a 3-fold up-regulation of luciferase activity.

In a recent study the HIV envelope protein (gp160) was shown to bind with strong affinity to the heparin binding domain of FN (38). Therefore, we tested whether heparan sulfate could compete with viral binding to sFN and as a result inhibit the sFN-mediated increase in viral attachment. Virus was mixed with increasing concentrations of heparan sulfate before incubation with sFN and then incubated with CD4⁺ T lymphocytes for 30 min at 0°C. The number of adherent viral particles was quantitated as described. Incubation of HIV with heparan sulfate inhibited the sFN-mediated increase in viral attachment in a dose-dependent manner (Fig. 5).

To determine whether the interaction of HIV and sFN was mediated through viral gp120, we asked whether full-length gp120 could bind to sFN and III₁-C peptide, using an ELISA-based adhesion assay. We found that full-length gp120 of HIV-1_IIIB bound significantly better to III₁-C peptide and sFN than to FN or BSA at concentrations as low as 0.4 μg/ml (Fig. 6,A). The V3 region of gp120 envelope has been implicated in the attachment of HIV to cell surface proteoglycans (39, 40). Because FN contains several heparin binding domains, and heparan sulfate inhibited sFN-mediated increased viral attachment, we asked whether the V3 region might be involved in binding. However, neutralizing Abs raised against the V3 region had no effect on HIV adhesion to III₁-C or sFN (data not shown). In addition, a synthetic peptide (aa 295–328), that contains the principal neutralizing domain of gp120 (GPGRAF) did not inhibit adhesion of HIV-1_IIIB to III₁-C (Fig. 6,B). However, a nonoverlapping gp120 peptide (aa 418–441) that includes the CD4 binding region almost completely inhibited the adhesion of HIV to III₁-C at 125 μg/ml (Fig. 6 B). Thus, the gp120 envelope protein of HIV is involved in binding of virus to the III₁-C region of sFN. The binding does not involve the V3 region as originally hypothesized, but instead involves the CD4 binding region within gp120.

Our data suggested that the sFN-mediated increase in HIV infectivity required binding of virus to sFN and subsequent binding of HIV-1_IIIB/sFN complex to the cell surface. Integrins are the predominant class of receptors that mediate binding of cells to FN (15, 16, 41). Primary CD4⁺ T lymphocytes express several FN-binding integrins, including α₄β₁, α₅β₁, and α_v-containing integrins (42). To test whether integrins mediated binding of the HIV-1_IIIB/sFN complex to the cells, we incubated cells with various anti-integrin Abs before incubation with HIV-1_IIIB/sFN complex. Anti-integrin Abs were not able to block the sFN-mediated increase in viral entry (Fig. 7). Similarly, incubation of cells with an RGD-containing peptide, the major motif through which integrins bind FN (41), did not block the sFN-mediated increase in viral entry. Because integrin binding requires the presence of divalent cations, we also tested the effect of the chelating agent EDTA. Treatment of cells with EDTA also failed to block the sFN-mediated enhancement in viral entry. In contrast, an Ab directed at CXCR4 (anti-fusin), a coreceptor used by T lymphotropic HIV strains (27), blocked viral entry by 50%. This is consistent with previous studies that show that the anti-fusin Ab only partially blocks T lymphocyte infection with certain HIV strains (43). Thus, the sFN-mediated increase in viral adhesion involved the interaction of HIV-1_IIIB/sFN complex with a class of receptors other than integrins on the cell surface.

Cell surface proteoglycans have been implicated in adherence of HIV to cells. HIV has been shown to bind to heparan sulfate proteoglycans, and enzymatic removal of heparan proteoglycans interferes with the initial attachment of HIV to T lymphocyte cell lines (44, 45). We investigated whether enzymatic removal of proteoglycans interfered with the attachment of the HIV-1_IIIB/sFN complex to cells. Treatment of cells with heparitinase I had no effect on adhesion of HIV-1_IIIB/sFN complex (Fig. 8). In contrast, treatment of cells with endo-β-galactosidase or chondroitinase ABC resulted in 25 and 32% decreases (p < 0.002), respectively. A combination of enzymes resulted in a 50% decrease in p24 levels (p < 0.001; Fig. 8). Thus, removal of cell surface proteoglycans interferes with the attachment of HIV-1_IIIB/sFN complex to the cell surface of T lymphocytes.

The exposure of cryptic sites within FN after matrix assembly may contribute to functional differences between matrix and soluble FN. Since it was not known whether the III₁-C epitope is exposed in lymph node tissue, we performed immunohistochemistry using an Ab that recognizes III₁-C peptide. This Ab does not recognize dimeric FN even when immobilized in a solid phase (Fig. 3,B). Using this Ab, we show immunoreactivity in rat lymph node tissue (Fig. 9,A). No immunoreactivity was detected in lymph nodes incubated with preimmune serum (Fig. 9 B). We also detected III₁-C immunoreactivity in FN matrix secreted and assembled in vitro by rat embryo fibroblasts (data not shown). Thus, the III₁-C epitope is exposed in the matrix of lymph nodes and may contribute to viral binding in vivo.

In this study we show that sFN, a multimeric form of FN, enhances HIV infection of CD4⁺ T lymphocytes. In contrast, dimeric FN, even when immobilized in a solid state phase, and other matrix proteins tested did not affect the efficiency of HIV infection. sFN (and III₁-C) resulted in productive infection of lymphocytes, as measured by increasing p24 production over time and detection of extrachromosomal viral DNA. Levels of extrachromosomal viral DNA are associated with ongoing viral replication and pathogenicity (46). We demonstrate that the interaction of HIV and sFN is mediated in part by III₁-C epitope in sFN and gp120 envelope protein of HIV-I_IIIB. We hypothesize that the difference between plasma FN and sFN relates to the exposure of the III₁-C epitope after multimerization. In sFN, the III₁-C region is available to interact with the gp120 envelope protein of HIV. The HIV/sFN complex adheres to lymphocytes via cell surface proteoglycans, followed by internalization of HIV and productive infection.

We speculate that our findings may be relevant to HIV infection in vivo. Following the initial viremic phase of infection, HIV accumulates in lymph node tissue, which serves as the major viral reservoir (4, 5). During the chronic phase of infection, the majority of viral particles are located extracellularly, associated with the FDC network (2, 3, 47). Presentation of HIV by FDCs is thought to aid in bringing viral particles and T lymphocytes in closer proximity and thereby enhancing the efficiency of HIV transfer into T lymphocytes (6, 7, 8, 9, 10). Matrix FN may work in concert with the FDC network to provide the optimal environment for trapping of HIV particles and enhance de novo infection of T lymphocytes by facilitating contact between HIV and the CD4 and chemokine receptor. Similarly, HIV entering the host in the presence of genital lesions is likely to encounter matrix FN as a result of tissue injury and wound healing (48, 49, 50). Matrix FN may provide a surface for viral particles to adhere to and promote their uptake by T lymphocytes.

To our surprise, III₁-C fragment, the recombinant peptide used to induce in vitro polymerization of FN, also increased HIV infection. FN is a modular protein that contains multiple type III repeats. Although these repeats show homology, the ability to enhance HIV infection appeared to be specific for the first type III repeat (III₁-C), because a peptide corresponding to the 11th type III repeat had no effect on HIV infectivity (data not shown). Interestingly, under certain circumstances the III₁-C fragment shares other properties with sFN, including antitumorigenicity (R. Pasqualini and E. Ruoslahti, unpublished observations) and the ability to trigger cell signaling (28). When FN is assembled into a matrix, conformational changes may increase the accessibility of cryptic sites within FN. A recent study showed that application of tension to FN fibrils by cells exposed a cryptic site in FN that included the III₁-C region (51) We show that polyclonal serum raised against III₁-C recognized the III₁-C epitope in rat lymph node tissue, but not in dimeric FN. Thus, it appears that the III₁-C epitope is cryptic in soluble FN, becomes exposed during FN matrix assembly, and is available to interact with viral particles. A recent report shows that the pathogenic bacteria, Streptococcus pyogenes, is able to discriminate between matrix and soluble FN and adheres preferentially to matrix FN (52). This may be important in the initial establishment of infection. Likewise, HIV appears to preferentially adhere to matrix FN.

It is also possible that proteolysis of FN during matrix remodeling produces III₁-C-containing fragments in vivo. During HIV infection, lymph nodes undergo extensive remodeling. This may lead to the proteolysis of FN, with the release of III₁-C-containing peptides. Proteolytic degradation products of several other matrix proteins are known to possess biological activity distinct from that of the parent protein (53, 54, 55, 56, 57, 58, 59). For example, angiostatin and endostatin, fragments derived from proteolytic degradation of plasminogen and collagen XVIII, respectively, inhibit angiogenesis (55, 56, 58, 59). We show that the III₁-C fragment bound HIV as effectively as sFN, but was less effective at increasing HIV infection. This could be due to less efficient binding of cells to III₁-C, which is supported by data from our laboratory showing that CD4⁺ T lymphocytes adhere less efficiently to III₁-C than to sFN (data not shown).

Binding of HIV to FN is thought to be mediated by the envelope protein of HIV (38). We examined whether this was also true for binding of HIV to sFN or the III₁-C peptide. We showed that the full-length gp120 envelope protein of HIV-1_IIIB adhered significantly better to sFN and III₁-C peptide than to FN (Fig. 6 A). We initially hypothesized that the V3 loop of the gp120 envelope protein may be involved in binding, because this region carries a positive charge and has been implicated in binding of virus to the proteoglycan-anchored heparan sulfate (40, 60, 61). However, neither an Ab raised against the V3 loop nor a synthetic peptide based on the V3 loop was able to inhibit adhesion of HIV-1_IIIB to sFN or III₁-C peptide. However, a peptide that included the CD4 binding region of gp120 inhibited adhesion of HIV-1_IIIB to III₁-C in a dose-dependent manner. Thus, the binding of HIV to III₁-C appears to involve the CD4 binding domain of gp120 and not the V3 loop. Others have shown specific binding of the HIV-1_IIIB envelope protein (gp160) to the heparin binding domain located in the FN carboxyl terminal (38). We speculate that the heparin cell binding domain that contains the III₁-C region mediates binding of HIV to sFN. Conformational changes induced by multimerization of FN may increase the exposure of the III₁-C region, explaining our finding that HIV-1_IIIB bound more efficiently to sFN than to FN. This is supported by the ability of an Ab specific for III₁-C to reduce binding of HIV to sFN.

Increased HIV uptake in the presence of sFN is dependent on cell surface proteoglycans, not integrins. Although FN and FN-derived fragments enhanced the efficiency of retroviral transfer in several studies (62, 63, 64, 65), the enhancement required interaction of FN with integrins, because removal of the RGD and LDV integrin binding sites from the FN-derived fragments resulted in loss of activity (63). In contrast, the ability of sFN to mediate enhanced HIV infection did not depend on the interaction of sFN with integrins. Treatment of cells with various anti-integrin Abs, EDTA, or an RGD-containing peptide did not affect the sFN-mediated increase in viral infection. Thus, the mechanism behind the increased viral uptake in the presence of sFN differs from that in FN fragments. Other studies have suggested that a class of receptors besides integrins may contribute to the adhesion of cells to sFN (19, 21). Our data support a role for cell surface proteoglycans in the adhesion of sFN-viral complex to the cell, because enzymatic removal of proteoglycans decreased the sFN-mediated increase in viral uptake. Interestingly, heparan sulfate-containing proteoglycans did not contribute to binding of HIV/sFN complex to the cells. Previous studies have shown that heparan sulfate proteoglycans can mediate the initial attachment of HIV to lymphocytic cell lines (60). In addition, binding of HIV to HeLa cells transfected with CD4 depends strongly on the presence of cell surface heparan sulfate proteoglycans and not CD4 (66). In our study removal of heparan sulfate proteoglycans did not affect the attachment of HIV-1_IIIB/sFN complex to cells. However, we used primary T lymphocytes that express minimal cell surface heparan sulfate proteoglycans compared with the lymphocyte cell lines used in the other studies (60).

Taken together, our data suggest that the sFN-mediated increase in HIV infection was due to enhanced binding of HIV-1_IIIB/sFN complex to proteoglycans on the cell surface of T lymphocytes, followed by internalization of virus particles and complete RT of HIV DNA. Additional mechanisms may also have contributed. Internalization of receptors occupied by sFN could lead to simultaneous internalization of HIV bound to sFN, promoting viral entry in addition to viral attachment. Another possibility is that interaction of sFN with cell surface receptors triggered intracellular signaling, affecting the expression and activity of CD4 and CXCR4. However, this is less likely, because preincubation of cells with sFN before exposure to virus did not lead to increased viral uptake (data not shown). Similarly, a role for transcriptional regulation by sFN is less likely, because adhesion of a lymphocyte cell line to sFN had no effect on HIV-LTR promoter activity.

The interaction of HIV and cells with matrix FN may be an important modulator of HIV infection in vivo. The ability of pathogenic organisms to discriminate between forms of FN adds an additional level of complexity to host-pathogen interactions. We propose that the extracellular environment of the lymph nodes may contribute to HIV pathogenicity. Pharmacological approaches targeted at minimizing this interaction may be beneficial in HIV infection prevention and treatment.

We thank Yan Wang for her technical support and S. Bourdoulous for sharing unpublished data and for critical reading of the manuscript. We also thank Drs. Karen Zier and Paul Klotman for thoughtful suggestions.