Cellular Entry of Hantaviruses Which Cause Hemorrhagic Fever with Renal Syndrome Is Mediated by β₃ Integrins

Abstract

Hantaviruses replicate primarily in the vascular endothelium and cause two human diseases, hemorrhagic fever with renal syndrome (HFRS) and hantavirus pulmonary syndrome (HPS). In this report, we demonstrate that the cellular entry of HFRS-associated hantaviruses is facilitated by specific integrins expressed on platelets, endothelial cells, and macrophages. Infection of human umbilical vein endothelial cells and Vero E6 cells by the HFRS-causing hantaviruses Hantaan (HTN), Seoul (SEO), and Puumala (PUU) is inhibited by antibodies to α_vβ₃ integrins and by the integrin ligand vitronectin. The cellular entry of HTN, SEO, and PUU viruses, but not the nonpathogenic Prospect Hill (PH) hantavirus (i.e., a virus with no associated human disease), was also mediated by introducting recombinant α_IIbβ₃ or α_vβ₃ integrins into β₃-integrin-deficient CHO cells. In addition, PH infectivity was not inhibited by α_vβ₃-specific sera or vitronectin but was blocked by α₅β₁-specific sera and the integrin ligand fibronectin. RGD tripeptides, which are required for many integrin-ligand interactions, are absent from all hantavirus G1 and G2 surface glycoproteins, and GRGDSP peptides did not inhibit hantavirus infectivity. Further, a mouse-human hybrid β₃ integrin-specific Fab fragment, c7E3 (ReoPro), also inhibited the infectivity of HTN, SEO, and PUU as well as HPS-associated hantaviruses, Sin Nombre (SN) and New York-1 (NY-1). These findings indicate that pathogenic HPS- and HFRS-causing hantaviruses enter cells via β₃ integrins, which are present on the surfaces of platelets, endothelial cells, and macrophages. Since β₃ integrins regulate vascular permeability and platelet function, these findings also correlate β₃ integrin usage with common elements of hantavirus pathogenesis.

Hantaviruses are enveloped viruses which define a unique genus within the Bunyaviridae and contain a tripartite negative-stranded RNA genome. The gene segments L, S, and M encode the viral RNA polymerase, the nucleocapsid protein (N), and two integral membrane surface glycoproteins (G1 and G2), respectively. Hantavirus glycoproteins are cotranslationally cleaved from a polyprotein precursor and define the highly ordered surface structure of the virion envelope (49).

Hantavirus infections are known to cause two human diseases, hemorrhagic fever with renal syndrome (HFRS) and a highly lethal (>40%) hantavirus pulmonary syndrome (HPS). In 1993, cases of adult respiratory distress in the southwestern United States led to the discovery of Sin Nombre virus (SN) as the etiologic agent of HPS (27, 41). Since then, HPS-associated viruses have been identified in 30 states, Canada, and South America (17, 25, 34, 37, 54). In contrast to HPS-associated strains, Prospect Hill (PH) is a distinct North American hantavirus which is not associated with any human disease (61).

There is a long history of HFRS derived from hantavirus infections. Hantaan virus (HTN), Seoul virus (SEO), Puumala virus (PUU), and Dobrava virus (DOB) are prominent causes of HFRS. HTN is the prototype HFRS-causing hantavirus and is the etiologic agent of Korean hemorrhagic fever (49). HTN is carried by the field mouse, Apodemus agraricus, and causes severe and sometimes fatal HFRS in humans. DOB is also carried by a discrete species of Apodemus (A. flavicolus) and causes severe HFRS, while SEO, carried by rats, and PUU, carried by bank voles, cause less severe forms of HFRS (49).

The means by which specific hantaviruses cause pulmonary or renal diseases is obscure (22, 27, 42, 64). In both animals and humans, hantavirus replication occurs predominantly in endothelial cells and macrophages (22, 27, 30, 42, 43, 62, 64). However, there is no evidence that hantaviruses cause disease while they persistently infect their primary small-mammal hosts (20, 49). In humans, hantaviruses cause acute pulmonary edema (HPS) (42, 64) or vascular hemorrhage and kidney dysfunction (HFRS), although both diseases are associated with acute thrombocytopenia and either disease may have pulmonary or renal components. However, there is little immune cell recruitment or damage to hantavirus-infected endothelial cells in either HPS or HFRS (42, 62, 64).

We have recently reported that β₃ integrins mediate the cellular entry of HPS-associated hantaviruses (SN and NY-1) (21). In contrast, the cellular entry of PH was blocked by antibodies to β₁- rather than β₃-specific integrins. We further demonstrated that SN and NY-1 entry occurred through an RGD tripeptide-independent integrin interaction. Although determinants of pathogenesis have not been defined for any hantavirus, unique integrin usage by HPS-associated hantaviruses and nonpathogenic PH suggested that integrin-specific interactions could contribute to the pathogenesis of HPS-associated hantaviruses.

Integrins are heterodimeric receptors composed of a combination of α and β subunits which mediate cell-cell adhesion, platelet aggregation, Ca²⁺ channel activation, and extracellular matrix protein (ECM) recognition (3–5, 23, 29, 31, 38, 39, 44, 51, 55, 59). Integrin-ligand interactions mediate the activation and regulation of intracellular signaling pathways within cells, which further control both transcriptional and ligand binding cell functions (7, 14, 50, 52). Integrins are also linked to intracellular cytoskeletal elements which facilitate cellular migration on the ECM. ECM proteins such as vitronectin and fibronectin contain Arg-Gly-Asp (RGD) tripeptides, which are recognized by specific cellular integrins including α_IIbβ₃, α_vβ₃, and α₅β₁ (55).

α_vβ₃ and α_IIbβ₃ integrins are abundant surface receptors of endothelial cells and platelets, respectively, and α_vβ₃ integrins are also present on macrophages (5, 29, 44, 51). α_IIbβ₃ integrins are present only on platelets, where they regulate platelet activation and participate in thrombus formation. α_vβ₃ integrins on endothelial cells also play central roles in maintaining capillary integrity through ligand binding interactions (5, 14, 28, 29, 36, 50, 55, 63). Further, α_vβ₃ and α₅β₁ differentially regulate arteriolar smooth muscle, resulting in vasodilation or vasoconstriction, respectively, through the intracellular activation of calcium channels (3, 23, 38, 39, 57, 59).

Integrin usage and the requirements for the cellular entry of HFRS-causing hantaviruses have not been addressed and are likely to contribute to our understanding of HFRS pathogenesis. In this study, we investigated cell surface interactions which are required for the entry of HFRS viruses into endothelial, CHO, and Vero E6 cells. We demonstrated that ligands and antibodies to β₃ integrins block the cellular entry of HTN, SEO, and PUU and that recombinant β₃ integrins facilitate the entry of HFRS-causing hantaviruses into CHO cells. Our results indicate that β₃ integrins are central elements in the entry of pathogenic HPS- and HFRS-causing hantaviruses and implicate a common role for β₃ integrins in hantavirus pathogenesis.

MATERIALS AND METHODS

Cell and virus.

Biosafety level 3 facilities were used throughout these experiments for hantavirus cultivation. Vero E6 and CHO cells were grown in Dulbecco’s minimal essential medium supplemented with 10% fetal calf serum, l-glutamine, and penicillin-streptomycin (GIBCO). CHO cells were supplemented with 100 mM nonessential amino acids (GIBCO). Human umbilical vein endothelial cells (HUVECs) were grown in EBM-2 (Clonetics) supplemented with 0.1% endothelial cell growth factor. CHO cells stably transfected with integrins α_IIbβ₃ (CHO-A5) or α_vβ₃ (CHO-VNRC) were described previously (10, 35). HTN (76-118), SEO (SR-11), PUU (K-27), SN (CC107), NY-1, and PH were cultivated as previously described (11, 48, 49). SN and NY-1 are serotypically distinct hantaviruses associated with HPS (11, 21, 24, 48). PH has not been associated with human disease (61, 62).

Ligands, peptides, and antibodies.

Vitronectin, fibronectin, laminin, fibrinogen, heparin, phytohemagglutinin, osteopontin, dextran sulfate, chondroitin sulfate A and B, and bovine serum albumin (BSA) were from Sigma. GRGDSP and GRGESP peptides were from GIBCO. Polyclonal rabbit antisera to β₁, β₃, β₄, α₁, α₂, α₅, and α_V, polyclonal goat antiserum to α₅β₁ (blocking), and blocking monoclonal antibodies to β₂ (MAb1962) and α_vβ₃ (MAb1976) were from Chemicon. β₃-specific monoclonal antibodies M15 and LIBS were described previously (15, 16). Integrin-associated protein (IAP; CD47)-specific monoclonal antibodies B6H12 and 2D3 were described previously (18, 56). Rabbit sera and monoclonal antibody 7A12, specific for rotavirus proteins, were used as negative control sera in these studies (53). Antibodies were used at a range of concentrations (20 ng/ml to 50 μg/ml) in addition to those shown in the figures. After cells were washed, a 1:2,000 dilution of anti-mouse or anti-rabbit sera was incubated with the cells for 1 h at 37°C. N-protein-specific polyclonal rabbit serum has been described previously and was made to recombinant N-protein from NY-1 expressed in Escherichia coli (21). N-protein-specific sera cross-react with N-proteins from all tested hantaviruses (21a).

Ligand and antibody pretreatment of cells.

Vero E6 cells or HUVECs were pretreated with antibodies or potentially competitive ligands for 1 h at 37°C. Antibodies (20 ng/ml to 50 μg/ml), ligands (1 to 50 μg/ml), and peptides (1 to 500 μg/ml) were preadsorbed to cells in 50 μl of phosphate-buffered saline (PBS)–1% BSA in duplicate wells of a 96-well plate. The sera or ligands were removed, and the monolayers were washed three times with PBS. Then 200 to 800 focus-forming units (FFU) of respective hantavirus inocula was adsorbed to the monolayers for 60 to 90 min. The viral inocula were removed, and the monolayers were washed three times and incubated 24 or 36 h prior to methanol fixation.

Quantitation of hantavirus-infected cells.

Methods for immunoperoxidase staining of hantavirus antigens in infected cells were previously described (53). Briefly, cell monolayers were fixed in 100% methanol for 10 min at 4°C and incubated with polyclonal rabbit anti-nucleocapsid sera (1/2,000) for 1 h at 37°C. The monolayers were washed three to five times with PBS and incubated with a 1/5,000 dilution of a goat anti-rabbit horseradish peroxidase conjugate (Kirkegaard & Perry Laboratories). The monolayers were then washed three times with PBS and stained with 3-amino-9-ethylcarbazole (0.026%) in 0.1 M sodium acetate (pH 5.2)–0.03% H₂O₂ for 5 to 30 min (53). They were then washed once with distilled water, immunoperoxidase-stained infected cells were quantitated, and duplicate wells were compared. Quantitation of experimentally infected cells was compared to that of mock-infected or untreated infected cell controls.

RESULTS

To determine whether integrins or additional cell surface receptors mediate the cellular entry of HFRS-causing hantaviruses, we assessed the ability of pretreating cells with ligands to inhibit infection by HTN, SEO, PUU, and PH. A number of ligands to cell surface receptors were tested for their ability to block infectivity, including fibrinogen, heparin, vitronectin, fibronectin, osteopontin, laminin, phytohemagglutinin, dextran sulfate, and chondroitin sulfate A (5 ng/ml to 50 μg/ml). We found that SEO and PUU (Fig. 1B and C) infectivity was inhibited by >70% by vitronectin pretreatment whereas HTN infectivity (Fig. 1A) was reduced by just 60% following pretreatment with 40 μg of vitronectin per ml. HTN was partially inhibited (∼20%) by fibronectin pretreatment, SEO infectivity was reduced slightly (∼10%) by fibrinogen, and PUU infectivity was slightly reduced (20%) by osteopontin. In contrast, PH infectivity (Fig. 1D) was inhibited by >70% by fibronectin but was not blocked by vitronectin pretreatment (21). Preadsorption of other tested ligands (50 μg/ml) (see Materials and Methods) did not block HTN, SEO, PUU, or PH infectivity.

FIG. 1.

Ligand-specific inhibition of hantavirus infectivity. Potentially competitive ligands (5 to 40 μg/ml, 50 μl) were preadsorbed to Vero E6 cells for 1 h prior to viral adsorption. Approximately 400 FFU of HTN, SEO, PUU, or PH was adsorbed to duplicate wells of a 96-well plate. Following adsorption, the inocula were removed and the cells were washed and further incubated for 24 to 36 h at 37°C in 5% CO₂ prior to methanol fixation. Hantavirus-infected cells were immunoperoxidase stained, as previously described (53), with polyclonal rabbit anti-nucleocapsid sera made to bacterially expressed and nickel affinity-purified NY-1 N-protein. Infected cells were quantitated and compared to cells with control infections without competitor proteins. The results were reproduced in at least three separate experiments and are presented as the percent inhibition with respect to control infections. Pretreatment with 1 to 500 μg of BSA per ml did not affect HTN, SEO, PUU, or PHV infectivity (results not shown).

Vitronectin and fibronectin are ligands for specific cell surface integrins. To further analyze the involvement of cellular integrins in the entry of HFRS-causing hantaviruses, antibodies to specific integrins were preadsorbed to cells before they were infected with HTN, SEO, or PUU. Pretreatment of Vero E6 cells with antibodies to β₃ integrin subunits inhibited HTN, SEO, and PUU infectivity by >70% but had no effect on PH infectivity (<5%) (Fig. 2). In addition, antibodies to α_v integrin subunits blocked PUU by >70% but reduced the infectivity of SEO and PUU by only 50% and had no effect on PH. In contrast, α₅- and β₁-specific antibodies inhibited PH infectivity by ≥60% (Fig. 2).

FIG. 2.

Infectivity of HTN, SEO, PUU, and PH inhibited by integrin-specific antibodies. Duplicate wells of Vero E6 cells were pretreated for 1 h at 37°C with 20 μg of antibodies to specific integrins per ml, and after being washed with PBS they were similarly incubated with a 1/2,000 dilution of anti-rabbit or anti-mouse sera. The monolayers were washed, and HTN, SEO, PUU, or PH was adsorbed (Fig. 1). Infected cells were quantitated as in the experiment in Fig. 1. The number of FFU observed 36 h postinfection is expressed as a percentage of control infections for each viral inoculum. Polyclonal rabbit sera to β₁, β₃, β₄, α₁, α₂, α₅, α_V, and α₅β₁ (blocking), as well as blocking monoclonal antibodies to β₂ (MAB1962) and α_vβ₃ (MAB1976), were from Chemicon.

In Fig. 3 polyclonal antibodies to α_vβ₃ or α₅β₁ integrins were bound to Vero E6 cells prior to hantavirus adsorption. Similar to integrin subunit-specific sera, α_vβ₃-specific sera inhibited HTN, SEO, and PUU infectivity by approximately 70 to 80% while failing to block PH infectivity. Reciprocally, antisera to α₅β₁ blocked PH infectivity (by >80%) but failed to inhibit HTN, SEO, or PUU infections. Vero E6 cells contain β₃ integrin subunits as demonstrated by Western blotting of cell lysates (data not shown). These results suggest that β₃-specific integrins mediate the cellular entry of HFRS-causing hantaviruses. Slight reductions in HFRS infectivity were also effected by some additional integrin ligands or α_v-specific sera with different hantaviruses (Fig. 3). This suggests that hantavirus interactions with cellular integrins are not identical even though they have common β₃ integrin involvement. Preadsorption of a mixture of sera (50 μg/ml each) which recognize α₅β₁, α_vβ₃, β₂, and β₄ integrins did not further reduce the infectivity of any hantavirus (not shown).

FIG. 3.

HFRS hantavirus infectivity is inhibited by α_vβ₃-specific antibodies. Duplicate wells of Vero E6 cells were pretreated with 20 ng to 20 μg of α_vβ₃ or α_vβ₃ polyclonal rabbit sera per ml for 1 h at 37°C (results for 20 μg/ml are shown in the figure). The monolayers were washed three times with PBS, and a 1/2,000 dilution of anti-rabbit sera was incubated with cells for 1 h at 37°C. The monolayers were washed three times with PBS and infected with approximately 400 FFU of HTN, SEO, PUU, or PH for 1 h at 37°C. They were then washed and incubated for 24 h prior to immunoperoxidase staining of the hantavirus N-protein present in infected cells (see Materials and Methods) (Fig. 1). Infected cells were quantitated as in Fig. 1, and the results are presented as the percent inhibition with respect to control infections.

α_vβ₃ integrins are present on the surface of endothelial cells infected by hantaviruses. To determine if α_vβ₃ integrins mediate the entry of HTN, SEO, PUU, or PHV into endothelial cells, we treated HUVECs with antibodies to integrin subunits prior to infection (Fig. 4). Both α_v- and β₃-specific sera reduced PUU infectivity by >70%, whereas HTN and SEO were inhibited >70% by only β₃-specific sera and approximately 40% by α_v-specific sera. In contrast, PH infectivity was inhibited by only β₁- and α₅-specific sera. This demonstrates that the infection of human endothelial cells by HFRS-causing hantaviruses (HTN, SEO, PUU, and PH) is also mediated by specific integrin interactions.

FIG. 4.

Infection of HUVECs by HTN, SEO, PUU, and PH is inhibited by integrin-specific antibodies. HUVECs were pretreated with 20 ng to 20 μg of integrin-specific antibodies per ml for 1 h as described in the legend to Fig. 2. Following primary-antibody addition (for 1 h at 37°C) and PBS washing, a 1/2,000 dilution of anti-rabbit or anti-mouse sera, respectively, was incubated with the cells for 1 h at 37°C). HUVECs were subsequently infected with approximately 400 FFU of HTN, SEO, PUU, and PH for 1 h at 37°C. The monolayers were washed and incubated for 24 h prior to immunoperoxidase staining of the hantavirus N-protein present in infected cells (53). Infected cells were quantitated as in Fig. 1, and the results are presented as the percent inhibition with respect to control infections.

To demonstrate that α_vβ₃ integrins are required for the cellular entry of HTN, SEO, and PUU, we studied the ability of recombinant integrins to mediate hantavirus infection of β₃ integrin-deficient CHO cells. CHO cell lines containing recombinant α_vβ₃ (CHO-VNRC) or α_IIbβ₃ (CHO-A5) (10, 35) integrins were infected with constant amounts of HTN, SEO, PUU, or PH. CHO-A5 or CHO-VNRC cells containing β₃-specific integrins dramatically enhanced the infectivity of HTN (Fig. 5), SEO, and PUU (results not shown) viruses but did not facilitate PH infections. Pretreatment of CHO cell lines with antisera to β₃ integrins specifically reduced the number of cells infected by HTN, SEO, or PUU by >90% (Fig. 5). A small number of infected CHO cells are observed following infection by these and other hantaviruses (21). These findings demonstrate that the introduction of β₃-specific integrins into cells facilitates HTN, SEO, and PUU infection but also suggests that additional cell surface interactions may contribute to hantavirus infectivity.

FIG. 5.

Specificity of HTN and PH infection of CHO, CHO-A5, and CHO-VNRC cells. The number of infected cells following HTN or PH inoculation of CHO, CHO-A5, and CHO-VNRC cell lines is presented (10, 35). Inoculation with or without prior treatment of cells with 20 μg of rabbit anti-β3 polyclonal sera per ml was performed as in the experiment in Fig. 2. Duplicate wells of Vero E6 cells were pretreated for 1 h at 37°C with 20 μg of antibodies to specific integrins per ml and, after being washed with PBS, similarly incubated with a 1/2,000 dilution of anti-rabbit or anti-mouse sera. Monolayers were washed, and HTN or PH were subsequently adsorbed and quantitated 24 h postinfection as in Fig. 1.

Cellular β₃ integrins are present in combination with the 50-kDa IAP, which is a cell surface receptor for thrombospondin (18, 19, 56). To determine if IAP-hantavirus interactions facilitate viral entry, we determined whether antibodies to IAPs were capable of inhibiting hantavirus infections. Prebinding either of two IAP-specific blocking monoclonal antibodies to cells had no effect on HTN infectivity (Fig. 6) and similarly did not alter SEO or PUU infectivity. In contrast, a β₃-specific monoclonal antibody inhibited HTN virus infectivity by >90%. Similarly, when both IAP-specific and β₃-specific antibodies were prebound to cells, no additional inhibition was observed over that due to β₃ antibody addition alone. These findings suggest that IAPs do not participate in additional hantavirus interactions with cells.

FIG. 6.

HTN infectivity is not inhibited by antibodies to integrin-associated protein. Duplicate wells of Vero E6 cells were pretreated with 20 ng to 20 μg of monoclonal antibodies (2D3 and B6H12) to the 50-kDa IAP per ml for 1 h at 37°C or with LB556, which is specific for β₃ integrin subunits (18, 19, 56). The monolayers were washed three times with PBS, and a 1/2,000 dilution of anti-mouse sera was incubated with cells for 1 h at 37°C. Monolayers were washed three times with PBS and infected with approximately 400 FFU of HTN, SEO, PUU, or PH for 1 h at 37°C (results for HTN are shown). The monolayers were washed and incubated for 24 h at 37°C prior to immunoperoxidase staining of the hantavirus N-protein present in infected cells (53). Infected cells were quantitated as in Fig. 1, and the results are presented as the percent inhibition with respect to control infections. mAb, monoclonal antibody.

Integrin-ligand interactions are often mediated by RGD tripeptides. To determine if hantavirus-integrin interactions are RGD independent, we determined whether RGD- or RGE-containing peptides blocked HTN, SEO, or PUU infectivity. Preadsorbtion of RGD- or RGE-containing peptides to cells and coincubation of RGD or RGE peptides competitively during hantavirus adsorption had no effect on infectivity, even at concentrations as high as 500 μg/ml (Fig. 7). However, treatment of cells with RGD- but not RGE-containing peptides (20 μg/ml) prior to addition of vitronectin (10 μg/ml) completely abolished the ability of vitronectin to inhibit HTN, SEO, or PUU infectivity. This is consistent with our finding that additional RGD-containing ligands failed to inhibit hantavirus infectivity. As a result, vitronectin is likely to block hantavirus infectivity through steric interference and not through a competitive RGD binding blockade. Similar to PH and HPS-associated hantaviruses (SN and NY-1), interactions of HTN, SEO, and PUU with integrins are RGD independent.

FIG. 7.

HFRS integrin interactions are RGD independent. Vero E6 cells were pretreated with potentially competitive GRGDSP or GRGESP control peptides (≤500 μg/ml), vitronectin (20 μg/ml), or β₃-specific monoclonal antibody 15 (50 μg/ml) in 50 μl for 1 h at 37°C prior to viral adsorption. GRGDSP was added to the monolayers 60 min prior to addition of vitronectin or fibronectin and subsequent virus adsorption in GRGDSP-vitronectin or GRGDSP-fibronectin experiments. Following RGD pretreatments, approximately 400 FFU of HTN, SEO, PUU, or PH was adsorbed to duplicate wells of a 96-well plate (results for PUU are shown). Following adsorption, the inocula were removed and the cells were washed and further incubated for 24 to 36 h prior to methanol fixation. Hantavirus-infected cells were immunoperoxidase stained, as described previously (53), with polyclonal rabbit anti-nucleocapsid sera. Infected cells were quantitated and compared to control infections without competitor proteins. The results are presented as the percent inhibition with respect to control infections. BSA at all concentrations tested (1 to 100 μg/ml) did not affect HTN, SEO, PUU, or PHV infectivity (data not shown).

c7E3 (ReoPro) is a commercially available α_vβ₃-specific humanized Fab fragment which is used therapeutically to inhibit thrombus formation in and around vascular stents (8, 9, 45). We tested the ability of c7E3 to inhibit the HTN, SEO, and PUU HFRS-causing strains as well as the NY-1 and SN HPS-associated strains and the nonpathogenic PH. Figure 8 demonstrates that c7E3 prebinding inhibited the infection of all hantaviruses which enter cells via interactions with α_vβ₃ but did not inhibit PH and further suggests a therapeutic potential for c7E3 or other α_vβ₃-specific antibody interventions for hantavirus disease.

FIG. 8.

Fab c7E3 (ReoPro) inhibits the infectivity of HPS- and HFRS-causing hantaviruses. Duplicate wells of Vero E6 cells were pretreated with 20 ng to 40 μg of mouse-human hybrid Fab, c7E3, to α_vβ₃ per ml for 1 h at 37°C. The monolayers were washed three times with PBS, and a 1/2,000 dilution of anti-mouse sera was incubated with the cells for 1 h at 37°C. The monolayers were washed three times with PBS and infected with approximately 400 FFU of HTN, SEO, PUU, NY-1, SN, and PH for 1 h at 37°C. They were then washed and incubated for 24 h prior to immunoperoxidase staining of the hantavirus N-protein present in infected cells (see Materials and Methods) (Fig. 1). Infected cells were quantitated as in Fig. 1, and the results are presented as the percent inhibition with respect to control infections.

DISCUSSION

HFRS was first described over 50 years ago, and the first hantavirus was isolated in 1978 (33). A growing number of hantaviruses which cause human disease have recently been discovered throughout the world. DOB, Tula, Khabarovsk, and others are hantaviruses which have recently been identified in Eurasia (49). HPS-associated hantaviruses were discovered in North America in 1993 but have now been identified throughout the Americas (17, 25, 34, 37, 54). Interestingly, the first person-to-person transmission of a hantavirus (Andes) was recently reported and suggests that some hantaviruses have adopted new means of their transmission (12). These findings have increased interest in the interactions of hantaviruses with cells which mediate viral entry and provoke pathogenic responses.

In this report, we demonstrate that HFRS-causing hantaviruses, HTN, SEO, and PUU, gain cellular entry via specific α_vβ₃ or α_IIbβ₃ integrins. We previously demonstrated that the cellular entry of HPS-associated hantaviruses, SN and NY-1, also occurs through interactions with β₃ integrins (21). In contrast, entry of PH is not mediated by β₃ integrins and PH is not associated with any human disease (60, 62). Although the cellular entry of the few additional nonpathogenic hantavirus strains has yet to be studied, the ability of pathogenic HPS- and HFRS-causing hantaviruses to enter cells, via the same integrins, provides a compelling rationale for common pathogenic aspects of HFRS and HPS. Changes in hantavirus-integrin usage or the use of additional cell surface receptors for viral entry could also participate in the transmission of hantaviruses among animal hosts and to humans.

Although it is possible that adapting hantaviruses to growth in Vero E6 cells selects for hantaviruses which gain entry via β₃ integrins, there are a number of reasons why this is very unlikely. We have now demonstrated that five separate hantavirus isolates, HTN, SEO, PUU, SN, and NY-1, with up to 60% divergent amino acid sequences in their surface glycoproteins enter cells via β₃ integrins and are effectively blocked by only one integrin ligand, vitronectin. Further, when the hantavirus genome was sequenced in its entireity from a patient, a small-mammal host, and Vero E6 cells after passage five times, no amino acid sequence differences were observed in any viral protein (6).

Our findings indicate that antibodies to integrins do not completely block hantavirus entry. However, anti-integrin antibodies do not completely inhibit adenovirus, papillomavirus, foot-and-mouth disease virus (FMDV), or coxsackievirus cellular entry (13, 26, 40, 46, 47, 58). Some of the residual binding (10%) reported for other viruses was observed to be nonsaturable and likely to be nonspecific (13). Uninhibitable hantavirus infectivity could similarly be due to some additional nonspecific interactions of hantaviruses with the cell surface or to additional specific hantavirus interactions which we have not yet identified. There is some indication of differential inhibitory effects on individual HFRS-causing hantaviruses by additional integrin-specific sera and ligands (Fig. 1 and 2). However, additive effects of α and β integrin antibodies or antibodies to α_vβ₃ integrins did not reduce the number of residually infected cells. In addition, IAP, which is a cell surface receptor present in tight conjunction with β₃ integrins (18, 19, 56), does not appear to mediate hantavirus entry, since antibodies to IAP had no effect on hantavirus infectivity in the presence or absence of anti-integrin antibodies. These findings do not alter our findings that the introduction of α_vβ₃-specific integrins into cells dramatically enhances cellular infection by HTN, SEO, and PUU and as a result that β₃ integrins confer the cellular entry of these HFRS-causing viruses.

Viruses which enter cells via α_vβ₃ integrins include adenoviruses, FMDV, and coxsackievirus A9, although adenoviruses and FMDV are also reported to use additional cell surface receptors (26, 40, 46, 47, 58). However, α_vβ₃ integrin usage by these viruses is dependent on integrin recognition of virally encoded RGD motifs at the ligand binding site (1, 26, 40, 46, 47, 58). RGD motifs are absent from all hantavirus proteins, and RGD peptides or additional RGD-containing integrin ligands fail to block infection by the hantaviruses which we have tested. Our findings demonstrate that HTN, SEO, and PUU (HFRS-causing hantaviruses), in addition to previously reported HPS-associated hantaviruses, associate with α_vβ₃ and α_IIbβ₃ integrins through unique RGD-independent interactions (21). This further suggests that vitronectin sterically blocks hantavirus-integrin interactions and that hantaviruses interact with unique integrin regions or require more complex cell receptor interactions for cellular entry.

Hantaviruses predominantly infect endothelial cells and macrophages and impact platelet functions. β₃ integrins are present on macrophages and are prominent endothelial-cell and platelet receptors. Integrins bind extracellular matrices; specify cell-cell adherence of platelets, macrophages, and endothelial cells; and mediate endothelial-cell migration (29, 55). β3 integrins also activate specific intracellular signaling pathways within platelets and endothelial cells, which further regulate a variety of adherence functions as well as intracellular calcium currents (2, 7, 32, 52, 59).

β₃ integrins play key roles in maintaining vascular integrity through platelet and endothelial cell barrier functions. Additionally, α_vβ₃ and α₅β₁ integrins on smooth muscle cells mediate vasodilation and vasoconstriction of arterioles, respectively, through intracellular signaling pathways linked to L-type Ca²⁺ channels (2, 32, 38, 39, 59). Thus, α_vβ₃- and α₅β₁-specific integrins differentially regulate blood flow and vascular tone (59). Interestingly, nonpathogenic PH and pathogenic hantaviruses are linked to the use of integrins which effect opposite calcium currents in endothelial and smooth muscle cells (2, 32, 38, 39, 59). The differential linkage of α_vβ₃ and α₅β₁ integrins to calcium currents in endothelial and smooth muscle cells provides a further rationale for the correlation of integrin usage and hantavirus pathogenesis.

Since hantaviruses alter vascular permeability and cause thrombocytopenia, hantavirus interactions with β₃ integrins could participate in common pathogenic HPS and HFRS responses. It remains to be addressed whether hantavirus interactions with β₃ integrins cause thrombocytopenia or alter endothelial-cell permeability. However, the regulatory functions of β₃ integrins in both platelet activation and endothelial-cell barrier function are clear and provide plausible means for hantavirus-protein interactions to mediate endothelial-cell- and platelet-specific vascular changes. The presence of large differences in the glycoprotein composition of HFRS- and HPS-causing hantaviruses (approximately 60 and 40% unique residues in the G1 and G2 proteins, respectively) further suggests how differential pathogenic responses could be effected by individual viruses despite common β₃ integrin usage.

Hantavirus interactions with α_vβ₃ integrins provide a clear means of altering vascular permeability during infection and also provide a potential point for therapeutic intervention during hantavirus infections. ReoPro is a human-mouse hybrid Fab fragment, c7E3, which recognizes both α_IIbβ₃ and α_vβ₃ integrins and is used therapeutically to inhibit platelet aggregation (8, 9, 45). We tested the ability of c7E3 to block the infectivity of both HFRS- and HPS-associated hantaviruses and demonstrated that it is capable of inhibiting these hantaviruses (Fig. 8). These findings suggest that c7E3 should be considered for the therapeutic treatment of hantavirus diseases. However, blocking platelet aggregation during hantavirus infection could also have negative effects on thrombocytopenic patients. Antibodies which block hantavirus-integrin interactions without altering platelet activation may provide less complicated and more specific intervention for hantavirus infection.

The divergence of hantavirus surface glycoproteins and common β₃ integrin usage provides further insight into the interaction of hantaviruses with cells. Differences in the respective G1 and G2 glycoproteins of hantaviruses are substantial, and this variation is likely to be associated with additional interactions which contribute to unique pathogenic responses to individual viruses. β₃ integrin usage also suggests that common elements exist on G1 or the more highly conserved G2 surface glycoprotein which mediate viral attachment to integrins. Although there is currently no data defining the virion attachment protein, the development of antibodies which recognize the hantavirus attachment protein and block integrin interactions are of interest since they are likely to provide an additional point for therapeutic intervention and vaccine development.