The neonatal Fc receptor is a pan-echovirus receptor

Significance

Echoviruses are associated with aseptic meningitis and induce severe and sometimes fatal disease in neonates and young infants. Here, we identify the neonatal Fc receptor (FcRn) as a pan-echovirus receptor. FcRn is expressed on the surface of the human placenta, and throughout life on intestinal enterocytes, liver hepatocytes, and in the microvascular endothelial cells that line the blood–brain barrier. This pattern of expression is consistent with the organ sites targeted by echoviruses in humans, as the primary entry site of infection is the intestinal and secondary sites of infection include the liver and brain. These findings provide important insights into echovirus pathogenesis and may explain the enhanced susceptibility of neonates to echovirus-induced disease.

Abstract

Echoviruses are amongst the most common causative agents of aseptic meningitis worldwide and are particularly devastating in the neonatal population, where they are associated with severe hepatitis, neurological disease, including meningitis and encephalitis, and even death. Here, we identify the neonatal Fc receptor (FcRn) as a pan-echovirus receptor. We show that loss of expression of FcRn or its binding partner beta 2 microglobulin (β2M) renders cells resistant to infection by a panel of echoviruses at the stage of virus attachment, and that a blocking antibody to β2M inhibits echovirus infection in cell lines and in primary human intestinal epithelial cells. We also show that expression of human, but not mouse, FcRn renders nonpermissive human and mouse cells sensitive to echovirus infection and that the extracellular domain of human FcRn directly binds echovirus particles and neutralizes infection. Lastly, we show that neonatal mice expressing human FcRn are more susceptible to echovirus infection by the enteral route. Our findings thus identify FcRn as a pan-echovirus receptor, which may explain the enhanced susceptibility of neonates to echovirus infections.

Echoviruses are small (∼30 nm) single-stranded RNA viruses belonging to the Picornaviridae family. These viruses make up the largest subgroup of the Enterovirus genus and consist of ∼30 serotypes. Enteroviruses are the main causative agents of aseptic meningitis worldwide, with echovirus 9 (E9) and echovirus 30 (E30) among the most commonly circulating serotypes (1). The neonatal and infant populations are at greatest risk for developing severe echovirus-induced disease, and infection within the first few weeks of life can be fatal (2, 3). In neonates, vertical transmission may occur before or at the time of delivery following a maternal infection (4). Echovirus infections in utero, both at late and earlier stages of pregnancy, have also been associated with fetal death (5–9).

Echoviruses are primarily transmitted through the fecal–oral route where they target the gastrointestinal (GI) epithelium. In primary human fetal-derived enteroids, echoviruses exhibit a cell type specificity of infection and preferentially infect enterocytes (10). The basis for this cell type-specific tropism is unclear. Decay accelerating factor (DAF/CD55) functions as an attachment factor for some echoviruses (11), but DAF expression does not sensitize nonpermissive cells to infection (12), suggesting that another cell surface molecule functions as the primary receptor. While integrin VLA-2 (α₂β₁) is a primary receptor for E1 (13), it does not serve as a receptor for other echoviruses. Other work has implicated a role for MHC class I receptors in echovirus infections due to inhibition of viral binding, entry, or infection by monoclonal antibodies to MHC class I and/or beta 2 microglobulin (β2M) (12, 14, 15), which is required for efficient cell surface trafficking of MHC class I receptors. However, the primary receptor for most echoviruses is unknown.

Here, we identify the human neonatal Fc receptor (FcRn) as a primary echovirus receptor. We show that human cells deficient in FcRn expression are resistant to echovirus infection and infection is restored by FcRn expression. Concomitantly, expression of human FcRn renders murine-derived cell lines and primary cells permissive to echovirus infection. In contrast, expression of the murine homolog of FcRn has little effect on viral infection in either human or mouse cells, suggesting a species-specific role for FcRn in echovirus infection. In addition, we show that a monoclonal antibody recognizing β2M, which noncovalently associates with FcRn and is required for FcRn cell surface expression, significantly reduces echovirus infection in primary intestinal epithelial cells and that recombinant FcRn in complex with β2M neutralizes echovirus infection and directly interacts with viral particles. Lastly, we show that neonatal mice expressing human FcRn are more susceptible to echovirus infection by the enteral route. Our data thus identify FcRn as a primary receptor for echoviruses, which has important implications for echovirus pathogenesis.

Results

Human Cells Deficient in FcRn Are Nonpermissive to Echovirus Infection.

We screened a panel of cell lines for their susceptibility to echovirus infection and found that human placental choriocarcinoma JEG-3 cells were resistant to infection by seven echoviruses (E5–E7, E9, E11, E13, and E30) but were highly permissive to the related enterovirus coxsackievirus B3 (CVB) (Fig. 1A). Levels of echovirus infection in JEG-3 cells were comparable to those observed in mouse embryonic fibroblasts (MEFs), which are resistant to echovirus infection, and were significantly less than those observed in permissive cell types, including human intestinal Caco-2, HeLa, human brain microvascular endothelial cells (HBMECs), and human osteosarcoma U2OS cells (Fig. 1A and SI Appendix, Fig. S1 A–C). The resistance of JEG-3 cells to echovirus infection occurred at the level of viral binding or entry, as infection was restored when cells were transfected with infectious viral RNA (vRNA) (SI Appendix, Fig. S1D).

Fig. 1.

JEG-3 cells are resistant to echovirus infection due to low FcRn expression. (A) JEG-3 cells (blue bars) or HeLa cells (gray bars) were infected with the indicated echovirus (1 pfu/cell) for ∼24 h. Viral titers (log₁₀ TCID50/mL) from the indicated cell types are shown as mean ± SD. Significance was determined using a standard t test (*P < 0.05). (B) Venn diagram from differential expression analysis using the DeSeq2 package in R between JEG-3 cells and either primary human fetal-derived enteroids (blue), HBMEC (green), Caco-2 cells (yellow), or JAR cells (red). There were 118 shared genes differentially down-regulated between JEG-3 cells and these cell types (red square). (C) Heatmap of 118 genes differentially down-regulated in JEG-3 cells and the indicated cell type (at Bottom) based on log₂ reads per kilobase of transcript, per million mapped reads (RPKM) values. Transcripts with no reads are shown in gray. (D) Heatmap of FcRn, B2M, HLA-A, -B, or -C expression in the indicated cell type (based on log₂ RPKM values). (E) JEG-3 cells were transfected with vector control (pcDNA), human HLA-A, or FcRn for 24 h, and then infected with the indicated echovirus for 24 h. Viral titers (log₁₀ TCID50/mL) are shown as mean ± SD with significance determined with a one-way ANOVA with Dunnett’s test for multiple comparisons (***P < 0.001). The relative expression of HLA-A and FcRn is shown in SI Appendix, Fig. S1H. Data in A and E are shown as mean ± SD.

We performed RNAseq-based transcriptomics analyses between nonpermissive JEG-3 cells and permissive cell types, including Caco-2 cells, HBMECs, and primary human enteroids harvested from fetal small intestines, which are highly sensitive to echovirus infection (10), to identify cell surface receptors differentially down-regulated in JEG-3 cells. Because JEG-3 cells arise from choriocarcinomas and express many placental-specific transcripts, we also included JAR cells in our analyses, another human choriocarcinoma line that is more permissive to echovirus infection than JEG-3 cells (SI Appendix, Fig. S1E). Using this approach, we identified 118 transcripts differentially down-regulated in JEG-3 cells (P < 0.001, log₂ z score less than −2, Fig. 1 B and C). Of these 118 transcripts, the neonatal Fc receptor (FCGRT, referred to here as FcRn), was the most significantly down-regulated cell surface receptor in JEG-3 cells (P < 0.001, log₂ z score less than −2), (Fig. 1D and SI Appendix, Fig. S1F). We confirmed the significantly lower levels of expression of FcRn in JEG-3 cells relative to permissive cell lines (HBMEC, HeLa, and JAR) and primary human fetal enteroids by RT-qPCR (SI Appendix, Fig. S1G). In contrast, there were no differences in expression of β2M, which is required to traffic FcRn to the cell surface (Fig. 1D and SI Appendix, Fig. S1G).

To determine if the lack of FcRn expression was directly responsible for the low levels of echovirus infection in JEG-3 cells, we ectopically expressed human FcRn (hFcRn). Expression of hFcRn in JEG-3 cells significantly increased their susceptibility to infection by E5, E11, and E30 (∼10,000-fold, Fig. 1E and SI Appendix, Fig. S1H). In contrast, expression of the related MHC class I or MHC class I-like molecules HLA-A and HLA-C and hemochromatosis protein (HFE), which also require β2M for cell surface expression, had no effect on infection (Fig. 1E and SI Appendix, Fig. S1 H and I). These data show that expression of hFcRn restores echovirus infection in nonpermissive human cells.

Expression of Human FcRn Restores Echovirus Infection in Mouse Cells.

Echoviruses do not infect mouse cells efficiently (SI Appendix, Fig. S1 A and B). Since ectopic expression of hFcRn in human cells in which endogenous levels were low restored their susceptibility to infection, we next determined whether the murine homolog of FcRn (mFcRn) was also sufficient to promote infection. Whereas expression of hFcRn in JEG-3 cells restored infection of a panel of echoviruses (E5, E7, E11, E13, and E30) by ∼10,000-fold, expression of mFcRn had no significant effect (Fig. 2A and SI Appendix, Fig. S2 A and B). Similarly, we found that expression of hFcRn, but not mFcRn, rendered MEFs and Chinese hamster ovary (CHO) cells highly susceptible to echovirus infection (Fig. 2B and SI Appendix, Fig. S2D). Collectively, these data show that expression of human, but not mouse, FcRn is sufficient to confer cellular susceptibility to echovirus infection, indicating a species-specific role for FcRn in echovirus infections.

Fig. 2.

Loss of FcRn expression reduces echovirus infection. (A) JEG-3 cells were transfected with vector control (pcDNA), human HLA-A, human FcRn (hFcRn), or mouse FcRn (mFcRn) for 24 h and then infected with the indicated echovirus for 24 h (E11^G Gregory strain and E11^S Silva strain). Viral titers (log₁₀ TCID50/mL) are shown as mean ± SD with significance determined with a Kruskal–Wallis test with a Dunn’s test for multiple comparisons (***P < 0.001). The relative expression of HLA-A, hFcRn, and mFcRn is shown in SI Appendix, Fig. S2B. (B) MEFs were transfected with pcDNA, human HLA-A, or either hFcRn or mFcRn, respectively, for 24 h and then infected with the indicated echovirus for 24 h. Viral titers (log₁₀ TCID50/mL) are shown as mean ± SD with significance determined with a Kruskal–Wallis test with Dunn’s test for multiple comparisons (**P < 0.01; ***P < 0.001). The relative expression of HLA-A, hFcRn, and mFcRn is shown in SI Appendix, Fig. S2C. (C) HBMECs were transfected with an siRNA against β2M (orange bar) or two independent siRNAs against FcRn (FcRn-1 and FcRn-2) alone or in combination (FcRn 1+2) (blue bars), or scrambled control siRNA (CONsi, gray bars) for 48 h and then infected with CVB or the indicated echovirus for an additional 16 h. Shown are viral titers (log₁₀ TCID50/mL) as mean ± SD with significance determined with a Kruskal–Wallis test with Dunn’s test for multiple comparisons (*P < 0.05; **P < 0.01; ***P < 0.001; ns, not significant). (D) At Top, control (WT) U2OS cells or two clones of U2OS cells depleted of FcRn expression by CRISPR/Cas9-mediated gene editing (FcRn^KO-1, FcRn^KO-2) were infected with the indicated echoviruses, or with CVB as a control, for ∼20 h. Levels of infection were assessed by immunostaining for double-stranded viral RNA and are shown as %positive cells over DAPI-stained nuclei. Significance was determined by a one-way ANOVA with a Dunnett’s test for multiple comparisons (***P < 0.001). At Bottom, immunoblotting for FcRn in control (Con) or FcRn^KO-1 and FcRn^KO-2 cells. GAPDH is shown at Bottom as a loading control. (E) Primary human intestinal epithelial cells were incubated with anti-β2M monoclonal antibody (blue bars) or isotype control antibody (gray bars) (2 μg/mL for both) for 30 min before infection with the indicated echovirus in the presence of antibody for an additional 24 h. Shown are viral titers (log₁₀ TCID50/mL) as mean ± SD from three independent HIE preparations with significance determined with a t test (*P < 0.05). Data are shown as mean ± SD.

Loss of FcRn Expression Renders Cells Resistant to Echovirus Infection.

We next determined whether loss of FcRn expression rendered cells expressing FcRn less susceptible to infection. For these studies, we used RNAi-mediated silencing or CRISPR/Cas9-mediated depletion of FcRn. We found that RNAi-mediated silencing of FcRn expression in HBMECs, an immortalized human blood–brain barrier cell line that expresses high levels of FcRn by two independent siRNAs, led to significant (∼1,000- to 10,000-fold) decreases in echovirus infection but had no effect on CVB infection (Fig. 2C and SI Appendix, Fig. S3 A and B). Similar results were obtained in human osteosarcoma U2OS cells (SI Appendix, Fig. S3C). In addition, silencing of β2M expression led to comparable reductions in infection (Fig. 2C and SI Appendix, Fig. S3 A–C). In contrast, RNAi-mediated silencing of other cell surface molecules that require β2M for trafficking, such as HLA-A, HLA-B, HLA-C, and HFE had no significant effect on echovirus infection in HBMECs (SI Appendix, Fig. S3D). Importantly, echovirus replication in β2M- and hFcRn-RNAi transfected cells was restored when cells were transfected with infectious vRNA (SI Appendix, Fig. S3E), supporting that the inhibition occurred at the stage of virus binding or entry. Moreover, we found that infection of many echoviruses (E5–E7, E9, E11, E13, E25, E29, and E30–E32) was reduced by β2M siRNA and FcRn siRNAs using a high content imaging-based screen (SI Appendix, Fig. S3G).

Next, we utilized CRISPR/Cas9-mediated gene editing to knock out FcRn expression. We found that infection of E5, E11, and E30 was significantly reduced in two clones of U2OS cells in which FcRn was knocked out (Fig. 2D and SI Appendix, Fig. S4 A and B). In contrast, infection by CVB was unchanged (Fig. 2D and SI Appendix, Fig. S4 A and B).

Consistent with a role for FcRn in echovirus infection, blocking antibodies to β2M inhibited infection by E5, E7, E9, E11, E13, and E30 in a cell line (U2OS, SI Appendix, Fig. S4C) and significantly reduced E7, E11, and E30 infection in primary intestinal epithelial cell monolayers derived from human fetal small intestines (Fig. 2E), where FcRn localizes to the subapical domain (SI Appendix, Fig. S4D). Collectively, these data show that FcRn expression is required for echovirus infection.

FcRn Facilitates Echovirus Attachment and Directly Interacts with Viral Particles.

We found that echovirus infection in cells depleted of FcRn could be restored by transfection of cells with vRNA, which suggested that this inhibition occurred at the stage of viral binding or entry.

We therefore determined whether down-regulation of FcRn expression would alter echovirus binding. We found that silencing of FcRn expression in HBMECs significantly reduced cell surface binding of E5, E7, E9, E11, and E30 to HBMECs (Fig. 3A). In contrast, this silencing had no effect on CVB binding (Fig. 3A). Residual levels of viral binding in HBMECs may be mediated by cell surface factors such as DAF that facilitate binding of some echoviruses (11). We also found that echovirus binding to primary mouse fibroblasts isolated from transgenic mice expressing human, but not mouse FcRn under the control of the endogenous human promoter [B6.Cg-Fcgrt^tm1Dcr Tg(FCGRT)32Dcr/DcrJ, hereafter referred to as hFcRn^Tg] (SI Appendix, Fig. S5A) (16, 17) was significantly higher than in cells isolated from wild-type (WT) (C56Bl/6) mice (Fig. 3B), indicating that human FcRn facilitates echovirus cell surface attachment.

Fig. 3.

FcRn mediates echovirus binding. (A) HBMECs were transfected with an siRNA against FcRn (FcRn-1, blue bars) or scrambled control siRNA (CONsi, gray bars) for 48 h and then the extent of viral binding of CVB or the indicated echovirus (50 pfu/cell) as assessed by a RT-qPCR–based binding assay. The extent of binding is shown as a fold from CONsi control (mean ± SD). Significance was determined using a t test (*P < 0.05; **P < 0.01; ns, not significant). (B) The extent of viral binding of CVB or the indicated echovirus (50 pfu/cell) was assessed in primary fibroblasts isolated from WT C57BL/6 mice (gray bars) or from hFcRn^Tg mice using a RT-qPCR–based binding assay. Shown is the extent of binding in cells isolated from four mice of each type, which is shown as mean ± SD. Significance was determined using a t test (*P < 0.05; **P < 0.01; ns, not significant). (C) E11 or E30 (10⁶ particles) were incubated with recombinant β2M (rβ2M, 2.5 μg/mL) or the extracellular domain of FcRn in complex with β2M (rFcRn-β2M, 2.5 μg/mL) for 1 h at 4 °C, preadsorbed to HBMECs for 1 h at 4 °C, washed, and then cells infected for 16 h. Shown are representative immunofluorescence images for double-stranded viral RNA (a replication intermediate, green). DAPI-stained nuclei are shown in blue. (D) E11 or E30 (10⁶ particles) were incubated with rβ2M (2.5 μg/mL), rFcRn-β2M (2.5 μg/mL), HLA-A (2.5 μg/mL), or HLA-C (2.5 μg/mL) for 1 h at 4 °C, preadsorbed to HBMECs for 1 h at 4 °C, washed, and then cells infected for 16 h. The level of infection was assessed by immunostaining for vRNA normalized to DAPI-stained nuclei. Shown is the percent of infection normalized to rB2M controls from experiments performed in triplicate (>1,000 cells total) as mean ± SD. Significance was determined with a Kruskal–Wallis test with Dunn’s test for multiple comparisons (***P < 0.001). (E) E11 or E30 (10⁸ particles) were incubated with rβ2M (5 μg/mL) or 6x His-tagged extracellular domain of FcRn in complex with β2M (rFcRn-β2M, 5 μg/mL) for 1 h at 4 °C, then incubated with Ni-NTA agarose beads for 1 h at 4 °C. Following extensive washing, immunoblots were performed for the viral capsid protein VP1 (Top) and then membranes were stripped and reprobed with an antibody recognizing the extracellular domain of FcRn (Middle). In parallel, level of input virus was immunoblotted with anti-VP1 antibody (Bottom). Data in A, B, and D are shown as mean ± SD.

To determine whether FcRn directly interacts with echovirus particles, we used a recombinant protein approach with a purified heterodimer containing the extracellular domain of FcRn in complex with β2M (rFcRn-β2M). We found that incubation of viral particles with rFcRn-β2M before infection neutralized both E11 and E30 infection (Fig. 3 C and D). In contrast, incubation with purified β2M alone, or recombinant HLA-A or HLA-C had no effect (Fig. 3 C and D). To determine whether there was a direct interaction between FcRn and echoviral particles, we performed in vitro binding assays using rFcRn-β2M. Using this approach, we found that rFcRn-β2M coprecipitated with purified E11 and E30 in in vitro binding assays, (Fig. 3E), demonstrating a direct interaction between FcRn and echovirus particles.

FcRn Promotes Infection in Neonatal Mice by the Enteral Route.

We found that exogenously overexpressed hFcRn rendered murine-derived cells sensitive to echovirus infection (Fig. 2B and SI Appendix, Fig. S2D). To further define the role of FcRn in echovirus infection, we compared echovirus infection in primary fibroblasts isolated from WT C57BL/6 and hFcRn^Tg mice. Primary fibroblasts isolated from WT mice were resistant to echovirus infection, as expected (Fig. 4A and SI Appendix, Fig. S5B). In contrast, cells isolated from hFcRn^Tg mice were highly permissive to echovirus infection and exhibited >10,000-fold enhanced susceptibility to infection (Fig. 4A and SI Appendix, Fig. S5B).

Fig. 4.

Expression of human FcRn enhances E11 infection in vivo. (A) Primary fibroblasts isolated from WT C57BL/6 mice (gray bars) or hFcRn^Tg mice (red bars) were infected with the indicated echovirus, or with CVB as a control for 24 h. Viral titers (log₁₀ TCID50/mL) are shown as mean ± SD from cells isolated from four mice of each type. Nd, not detected. Significance was determined using a standard t test (***P < 0.001). (B) E11 titers in the indicated tissues as determined by TCID50 assays from WT C57BL/6 mice (10 total) of hFcRn^Tg mice (9 total) infected for 7 d by oral gavage with 10⁶ E11 particles. Titers were normalized to tissue weight (organs, Left y axis) or volume (blood, Right y axis). LI, large intestine; SI, small intestine. Data are shown as data points from individual mice as a mean ± SD with significance determined by a nonparametric Mann–Whitney U test (*P < 0.05; **P < 0.01; ns, nonsignificant). (C and D) Immunohistochemistry for E11 using an antibody recognizing the VP1 capsid protein from WT or hFcRnTg infected as indicated in B. Shown are H&E (Left) and IHC (Right) from liver (C) or small intestine (D). Scale bars are shown at Bottom Left.

To define the role of FcRn in echovirus pathogenesis, we developed an in vivo model of E11 infection of WT and hFcRn^Tg neonatal mice by the enteral route. FcRn is expressed at high levels throughout the intestinal tract, liver, and endothelium in hFcRn^Tg mice (18, 19). We infected 7-d-old WT or hFcRn^Tg mice with 10⁶ infectious particles of E11 by oral gavage and then collected tissues (brain, liver, small intestine, large intestine, and stomach) at 3 or 7 d postinoculation. We did not detect any infectious virus as determined by TCID50 assays when tissues were collected at 3 d postinoculation (SI Appendix, Fig. S5C). However, at 7 d postinoculation, we detected significantly higher titers of E11 in the blood and livers of hFcRn^Tg mice compared with WT control mice (Fig. 4B). In addition, we also detected infectious E11 in the small intestines, large intestines, and stomachs of some hFcRn^Tg, but not WT, mice (Fig. 4B). There was no detectable E11 in tissue harvested from WT mice (Fig. 4B). We confirmed E11 replication in isolated organs by performing immunohistochemistry using an antibody against the VP1 viral capsid protein. E11 infection was present in the livers of hFcRn^Tg mice, where viral replication was localized to select cell types (Fig. 4C), and in the epithelium lining the small intestine (Fig. 4D). In addition, E11 replication was localized to the muscle of both the colon and stomach, with no detectable replication in the epithelium (SI Appendix, Fig. S6 A and B). These data show that expression of hFcRn in neonatal mice is sufficient to permit E11 infection by the enteral route.

Discussion

Here, we identify FcRn as a primary receptor for echoviruses. We show that expression of FcRn is necessary and sufficient for echovirus infection and that FcRn directly binds echovirus particles and facilitates viral binding. We also show that expression of human, but not mouse, FcRn restores echovirus infection in nonpermissive mouse and human cells and thereby identify a species-specific mechanism of infection. Our data show that a number of clinically relevant echoviruses commonly associated with human disease, including E9, E30, and E11, utilize FcRn as a receptor, suggesting a pan-echovirus role. In contrast, FcRn plays no role in the infection of related enteroviruses, including CVB and poliovirus (PV). Our findings provide important insights into the cellular receptor used by echoviruses to initiate their infections and into echovirus pathogenesis.

FcRn transports and regulates the circulating half-life of IgG throughout life (20). In addition, FcRn is responsible for the development of passive immunity through the transfer of maternal-derived antibodies. In humans, expression of FcRn on the placenta (21) is solely responsible for the establishment of passive immunity in the fetus due to transport of maternal-derived IgG across the placental surface directly into fetal blood (22). This differs in rodents, where passive immunity is established postnatally from maternal-derived IgG in milk/colostrum (23). FcRn is expressed throughout life in a variety of cell types, including the small intestine, the microvasculature of the blood–brain barrier, myeloid cells, and hepatocytes, among others (20). Although echoviruses are primarily transmitted via the fecal–oral route, viral-induced disease is associated with infection of secondary organs, most notably the liver and brain. Our in vivo studies support a role for FcRn at these sites of echovirus infection, as we identified E11 infection in the livers of hFcRn^Tg neonatal mice. However, it is unclear whether the virus preferentially infects hepatocytes or Kupffer cells, the specialized macrophages that line liver sinusoids, which express high levels of FcRn (18). In addition, we found that E11 infection at other sites such as the colon and stomach was largely restricted to the subepithelial muscular layer, which also expresses FcRn (24). The expression of FcRn on cell types present in the intestine, brain microvasculature, and liver may thus explain the tropism of echoviruses for these tissues and the viral mechanism to bypass the barriers presented by the cells comprising these sites.

In addition to IgG, FcRn also binds albumin, which regulates hepatic injury (25, 26). Of note, previous work has shown that albumin inhibits E7 infection at the stage of viral uncoating (27), suggesting that the interaction between echoviruses and FcRn may occur at the interface of albumin–FcRn binding. Although FcRn binds to albumin and IgG at distinct sites (28), both of these interactions occur within the low pH (≤6.5) environment of endosomes, with release occurring in the basic pH (≤7.5) of the bloodstream. In contrast, our findings suggest a direct interaction between echoviruses and FcRn that occurs at the neutral pH of the cell surface before viral entry, although it is unclear whether different echoviruses exhibit differences in affinity for the receptor. Indeed, our in vitro pulldown assays suggest that E11 might exhibit higher affinity for the receptor than E30. However, it is possible that differences in the ratio of infectious to noninfectious particles between echovirus preparations accounts for these differences. Once internalized, it is possible that the interaction between FcRn and echoviruses is altered by the low pH of endosomes, which may facilitate subsequent genome release and/or endosomal escape. We have shown that E11 preferentially infects enterocytes (10), with enhanced infection from the basolateral surface of human intestinal epithelial (HIE) cells (29). This polarity of infection is consistent with the enhanced expression of FcRn in enterocytes in the intestine and its enrichment to the basolateral surface. Following replication, E11 is released bidirectionally from HIE from both the apical and basolateral domains (29). Given that FcRn mediates bidirectional transport (30), this raises the possibility that echoviruses could be transported from either the apical or basolateral domains to cross the intestinal barrier.

Echoviruses are associated with severe disease in neonates, particularly during the first 2 weeks of life and in those born prematurely. The vertical transmission of echoviruses is thought to occur at the time of delivery and be associated with maternal infection in the preceding days or weeks. However, fetal infections in utero have also been associated with disease and/or death (5–9), suggesting that vertical transmission might also occur during pregnancy. FcRn is highly expressed on syncytiotrophoblasts (31, 32), the fetal-derived cells that comprise the outermost cellular barrier of the human placenta and which directly contact maternal blood. These cells are highly resistant to viral infections due to intrinsic antiviral defense pathways (21). However, given that FcRn expressed on the surface of these cells transcytoses maternal-derived IgG directly into the underlying fetal blood, our identification of FcRn as an echovirus receptor raises the possibility that echoviruses might have higher rates of transplacental transfer than has been previously appreciated. In addition, it should be noted that the highest rates of transplacental IgG transfer occur in the third trimester, with the level of maternal-derived IgG greater in the fetus than in the mother (33). Thus, a maternal echovirus infection in the later stages of pregnancy could potentially lead to FcRn-mediated placental infection or transplacental viral transport and expose the fetus to virus before delivery. Further defining the role of FcRn in echovirus infections in utero and postnatally will provide important insights into echovirus-induced fetal and neonatal disease.

Our work presented here identifies FcRn as a pan-echovirus receptor. Given that FcRn-based therapeutics have been developed to target a variety of human diseases (34), our findings also point to FcRn as a possible target for anti-echovirus therapeutics to ameliorate virus-induced disease. Future studies identifying the mechanism by which echoviruses utilize FcRn to enter or bypass barrier tissues such as the GI epithelium, blood–brain barrier, and placenta will provide important insights into a variety of aspects of echovirus pathogenesis.

Materials and Methods

Additional materials and methods are located in SI Appendix, Supplemental Materials and Methods.

Cell Lines.

HBMECs were obtained from Kwang Sik Kim, Johns Hopkins University, Baltimore, MD, described previously (35), and grown in RPMI-1640 supplemented with 10% FBS (Invitrogen), 10% NuSerum (Corning), nonessential amino acids (Invitrogen), sodium pyruvate, MEM vitamin solution (Invitrogen), and penicillin/streptomycin. JEG-3, JAR, U2OS, and Caco-2 (BBE clone) cells were purchased from the American Type Culture Collection (ATCC) and cultured as described previously (36, 37). HeLa cells (clone 7B) were provided by Jeffrey Bergelson, Children’s Hospital of Philadelphia, Philadelphia, PA, and cultured in MEM supplemented with 5% FBS and penicillin/streptomycin. Primary human intestinal epithelial cells were isolated from crypts isolated from human fetal small intestines as described (10). Complete methods can be found in SI Appendix, Supplemental Materials and Methods.

Animals.

All animal experiments were approved by the University of Pittsburgh Animal Care and Use Committee and all methods were performed in accordance with the relevant guidelines and regulations. Primary fibroblasts were generated from 4-wk-old B6.Cg-Fcgrt < tm1Dcr > Tg(CAG-FCGRT) 276Dcr/DcrJ (cat. no. 004919) and control C57BL/6J (cat. no. 000664) mice purchased from The Jackson Laboratory. For collection of primary cells, mice were killed according to institution standards and ears and tail were removed, incubated in 70% ethanol for 5 min, and then rinsed twice in PBS + 50 μg/mL kanamycin for 5 min. Hair was removed and tissue was cut into small pieces and incubated in 9.4 mg/mL collagenase D (11088858001, Roche) and 1.2 mg/mL pronase (1088858001, Roche) in complete DMEM at 37 °C with shaking at 200 rpm for 90 min. The resulting cell suspensions were filtered through 70-μM cell strainers, collected at 580 g, resuspended in complete DMEM containing 10 units penicillin and 10 μg streptomycin/mL and 250 ng/mL amphotericin B and cultured at 37 °C in a humidified 5% CO₂ incubator.

For animal infections, 3- or 7-d-old neonatal mice were infected with 10⁶ pfu of E11 in 50 μL of 1× PBS by oral gavage using a 24-gauge round-tipped needle. Mice were killed at various intervals postinoculation and organs harvested into 500 μL–1 mL of DMEM and stored at −80 °C before TCID50 assay. Samples were thawed at 37 °C and homogenized with a TissueLyser LT (Qiagen) for 8 min, followed by brief centrifugation at 400 × g. Viral titers in organ homogenates were determined by TCID50 assay.

Viruses and Infections.

Experiments were performed with CVB (RD strain), PV (Sabin strain, type 2), echovirus 5 (Noyce strain, E5), echovirus 6 (Burgess strain, E6), echovirus 7 (Wallace strain, E7), echovirus 9 (Hill strain, E9), echovirus 11 (Gregory or Silva strains, E11^G and E11^S), echovirus 13 (Del Carmen strain, E13), echovirus 25 (JV-4, E25), echovirus 29 (JV-10, E29), echovirus 30 (Bastianni strain, E30), echovirus 31 (Caldwell strain, E31), or echovirus 32 (PR-10 strain, E32) that were provided by Jeffrey Bergelson and originally obtained from the ATCC. Viruses were propagated in HeLa cells and purified by ultracentrifugation over a sucrose cushion, as described (38).

Unless otherwise stated, infections were performed with 1 pfu/cell of the indicated virus. In some cases, viruses were preadsorbed to cells for 1 h at 4 °C in serum-free MEM supplemented with 10 mM Hepes followed by extensive washing in 1× PBS or complete media. Infections were then initiated by shifting cells to 37 °C for the times indicated. Viral titers were determined by TCID50 assays in HeLa cells using crystal violet staining.

Binding assays were performed by preadsorbing 50 pfu/cell of the indicated virus to cells for 1 h at 4 °C in serum-free MEM supplemented with 10 mM Hepes followed by extensive washing with 1× PBS. Immediately following washing, RNA was isolated and RT-qPCR performed for viral genome-specific primers, as described below.

For experiments using blocking antibodies, cells were incubated with the indicated antibodies (at 5 μg/mL) for 1 h at 4 °C in serum-free DMEM containing 10 mM Hepes. For anti-DAF IF7 blocking experiments, all incubations were performed in DMEM containing 10% FBS and 10 mM Hepes. Following this incubation, viruses were preadsorbed to cells in the presence of antibodies for an additional 1 h at 4 °C in serum-free or serum-containing medium, washed extensively, and then the cells were infected at 37 °C for the indicated time in the presence of antibodies.

Statistics.

All statistical analysis was performed using GraphPad Prism. Data are presented as mean ± SD. A Student’s t test or one-way ANOVA was used to determine statistical significance, as described in the figure legends. Parametric tests were applied when data were distributed normally based on D’Agostino–Pearson analyses; otherwise nonparametric tests (such as Mann–Whitney U tests) were applied. P values of <0.05 were considered statistically significant, with specific P values noted in the figure legends.