The neonatal Fc receptor is a cellular receptor for human astrovirus

Abstract

Human astroviruses (HAstV) are major causes of gastroenteritis, especially in children, and there are no vaccines or antivirals currently available. Little is known about host factors required for their cellular entry. Here we utilized complementary CRISPR-Cas9-based knockout and activation screens to identify neonatal Fc receptor (FcRn) and dipeptidyl-peptidase IV (DPP4) as entry factors for HAstV infection in vitro. Disruption of FcRn or DPP4 reduced HAstV infection in permissive cells and, reciprocally, overexpression of these factors in non-permissive cells was sufficient to promote infection. We observed direct binding of FcRn, but not DPP4, with HAstV virions and the purified spike protein. This suggests that FcRn is a receptor for HAstVs while DPP4 is a cofactor for entry. Inhibitors for DPP4 and FcRn currently in clinical use prevented HAstV infection in cell lines and human enteroids. Our results reveal mechanisms of HAstV entry as well as druggable targets to limit HAstV infection.

Extended Data Fig. 1.. Disruption of FcRn or DPP4 in Caco2 cells limits HAstV infection.

(A) Enrichment scores of positively-selected sgRNAs in genome-wide CRISPR-Cas9 Caco2 cell library infected with HAstV1 at 6dpi, calculated by MAGeCK. (B) Overview of the pooled CRISPR screen. The genome-scale Brunello library was introduced into Cas9 expressing Caco2 cells, followed by selection of transduced cells. After 10 days, pooled Caco2 cells were infected with HAstV1 for 24h, followed by staining using anti-HAstV capsid antibody. Uninfected cells were sorted to determine the sgRNA counts by next-generation sequencing. (C) FcRn protein levels in Cas9-Caco2 cells disrupted for FCGRT or DPP4 using independent sgRNAs per gene or targeted with a control anti-GFP sgRNA. Two independent replicates are shown. (D) Representative histogram for surface expression of DPP4 in naïve Caco2 cells stained with anti-DPP4 antibody or isotype control antibody. (E) Representative histogram showing DPP4 expression in fixed, permeabilized Caco2 cells disrupted for FCGRT or DPP4 or B2M using gene specific sgRNAs or targeted with a control anti-GFP sgRNA. (F) Percentage of DPP4 expressing Caco2 cells disrupted for DPP4 using sgRNAs or targeted with a control anti-GFP sgRNA. (n=3). (G) FcRn protein levels in Caco2 cells expressing low and high levels of DPP4. A representative blot of 2 replicates is shown. (H) Percentage of anti-HAstV capsid antibody stained Caco2 cells grown in tissue culture plates (T.C. plate) for 72h and transwell inserts (Transwell) for 14 days prior to HAstV1 infection. (I, J) Percentage and mean fluorescent intensity (MFI) of DPP4 in Caco2 cells grown in T.C. plates for 72h or transwells for 14 days. (K) Cas9-Caco2 cells disrupted for DPP4 (n=6) or FCGRT (n=6) cells or targeted with a control anti-GFP sgRNA (n=6) grown in transwell inserts for 14 days were infected with HAstV1, then stained with anti-HAstV capsid antibody at 24hpi. (L) Percentage of anti-HAstV capsid antibody stained Caco2 cells treated with PBS (n=5) or isotype antibody (n=5) or various concentrations of anti-DPP4 antibody (n=6) prior to HAstV1 infection. (M) Percentage of anti-HAstV capsid antibody-stained control (n=6), FCGRT (n=6) or DPP4 (n=6) knockout Caco2 cells transfected with HAstV1 RNA at 48h post-transfection. Results were analyzed using Mann-Whitney test (H-J) or Kruskal-Wallis test with Dunn’s post-test (K-M) from two to three independent experiments. ***P=0.0043 (H), ***P<0.0001 (I), **P=0.03 (J), *P=0.07 and **P=0.001 (K), ***P<0.0001 (L). ns=not significant. Bars indicate mean of all data points.

Extended Data Fig. 2.. Enhanced expression of DPP4 or FCGRT promotes HAstV infection.

(A) Schematic of the CRISPR activation surfaceome screen. The human surfaceome library was introduced into dCas9 expressing Caco2 cells, followed by selection of transduced cells. After 10 days, pooled Caco2 cells were infected with HAstV1 or HAstV8 for 24h, followed by staining using anti-HAstV capsid antibody. We sorted top 3% of the HAstV capsid positive cells to determine the sgRNA counts by next-generation sequencing. (B) Representative histogram showing expression of DPP4 in dCas9-Caco2 cells transduced with sgRNAs for DPP4 overexpression. (C) FcRn protein levels in dCas9-Caco2 transduced with sgRNA for overexpressing FcRn. Two independent replicates are shown. (D) FcRn protein levels in 293T and 293T-FCGRT cells. Two independent replicates are shown. (E) Representative histogram showing surface expression of DPP4 in 293T cells stained with anti-DPP4 antibody or isotype control antibody. (F) Abundance of DPP4 and HAstV capsid positive 293T and 293T-DPP4 cells infected with HAstV1. (G, H) HEK293T or HEK293T-DPP4 cells were transfected with FCGRT (n=5), DPP4 (n=5) and/or B2M (n=5) plasmids then 48h later were infected and stained with anti-HAstV capsid antibody at 24 hpi. Results were analyzed using Kruskal-Wallis test with Dunn’s post-test (G and H) from three independent experiments. **P=0.009 and ***P<0.0001 (G), ***P<0.0001 (H). ns=not significant. Bars indicate mean of all data points.

Extended Data Fig. 3.. Purified DPP4 does not directly bind VP25 or VP34.

(A-C) Gel filtration and SDS-PAGE profiles of HAstV1 VP25 (A) HAstV8 VP25 (B) and HAstV1 VP34 (C). HAstV1 and HAstV8 VP25 were eluted from a HiLoad 16/600 Superdex 200 column and HAstV1 VP34 was eluted from a Superdex 75 Increase 10/300 column. (D) Soluble hDPP4 was expressed via transfection in Expi293 cells and eluted via gel filtration on Superdex 200 Increase 10/300 column (left) and further analyzed using MALS (right). The MALS curve (black) is plotted with the derived molecular weight (red) of 191,3000 Dalton ± 0.848%. (E-G) Binding experiments testing full-length spike protein from MERS-CoV, (E) VP25 from HAstV1 and HAstV8 (F, left and right, respectively), and VP34 from HAstV1 (G) against DPP4. In all cases, each protein was immobilized via biosensor and tested against 2-fold dilutions of DPP4 from either 1μM to 62.5nM (E, G) or 1μM to 15.625nM (F). Results were analyzed from three independent experiments.

Extended Data Fig. 4.. Despite similar inhibition of DPP4 activity, vildagliptin and teneligliptin do not prevent HAstV infection.

(A) Percentage of anti-HAstV capsid antibody-positive Caco2 cells at 24hpi treated with PBS (n=6) or various concentrations of nipocalimab (n=6) prior to HAstV1 infection. (B) Percentage of anti-HAstV capsid antibody-positive Caco2 cells treated with PBS (n=6) or vildagliptin (n=6) or teneligliptin (n=6) prior to HAstV1 or HAstV8 infection. (C) Fluorescent 7-Amino-4-Methyl Coumarin (AMC) released by DPP4 enzymatic activity in Caco2 cells treated with PBS (Control; n=4) or DMSO or DPP4 inhibitors (n=4). (D, E) Percentage of FcRn and DPP4 in 293T (n=5), 293T-FCGRT (n=6) and 293T-DPP4 (n=6) cells treated with anti-DPP4 antibody (ab) or nipocalimab (Nipo) prior to HAstV1 infection. *P=0.04, **P=0.001 and ***P<0.0001 (A), *P=0.01 (C). ns=not significant. Bars indicate mean of all data points.

Extended Data Fig. 5.. Flow cytometry gating strategy for HAstV infection.

Cells were gated on FSC/SSC to select the entire population of cells, then gated on FSC-A/FSC-H to select for single cells. Flow plots shown reflect the entire cell population without subgating. To assess for HAstV infection, gating for positive versus negative cells was done using uninfected versus infected control cells and subsequently applied uniformly to experimental cells to avoid bias.

Fig. 1.. FcRn and DPP4 are required for HAstV entry and infection.

(A) Enrichment [-log10 Robust Rank Aggregation (RRA)] scores of positively-selected sgRNAs in HAstV1-negative cells sorted 24hpi of a genome-wide CRISPR-Cas9 Caco2 cell library compared to HAstV1-positive cells, calculated by MAGeCK. (B,C) Cas9-Caco2 cells disrupted for FCGRT (n=6–9) or B2M (n=9) using two independent sgRNAs per gene or targeted with a control anti-GFP sgRNA (n=9) were infected with HAstV1or HAstV8, then stained with anti-HAstV capsid antibody at 24hpi. (D) Mean fluorescent intensity (MFI) of DPP4 in negative and positive cell populations of Caco2 cells infected with HAstV1 (n=6) or HAstV8 (n=6) infected at 24hpi. (E) Cas9-Caco2 cells were disrupted for DPP4 (n=9) using two independent sgRNAs per gene or targeted with a control anti-GFP sgRNA (n=9), then assessed for infection using anti-HAstV capsid antibody 24hpi with HAstV1 or HAstV8. (F) Percentage of anti-HAstV capsid antibody-stained Caco2 (WT) (n=7), DPP4^−/− Caco2 (n=7) and DPP4^−/− Caco2 cells complemented with a DPP4-expressing plasmid (n=7) at 24hpi with HAstV1. (G) Percentage of anti-HAstV capsid antibody-stained Caco2 cells treated with PBS (n=4–5) or 10μg/ml isotype control (n=4–6) or 10μg/ml anti-DPP4 polyclonal antibody (n=6–7) for 12h hours prior to infection with HAstV1 or HAstV8. (H) Cas9-Caco2 cells disrupted for FCGRT (n=8) or DPP4 (n=8) or targeted with a control anti-GFP sgRNA (n=8) were infected with VA1, then stained with anti-VA1 antibody at 48hpi. Results from three independent experiments were analyzed using the Kruskal-Wallis test with Dunn’s post-test (B to H). P<0.0001 (B), P=0.04 and 0.02 (C), P=0.002 (D), P<0.0001 (E), P<0.0001 (F), P<0.0001 (G), P=0.003 (H) ns=not significant. Bars indicate mean of all data points.

Fig. 2.. FcRn and DPP4 promote HAstV infection of non-permissive cells.

(A,B) Enrichment scores of positively-selected sgRNAs in HAstV1- or HAstV8-positive cells sorted 24hpi from a surfaceome CRISPRa Caco2 cell library compared to unsorted cells, calculated by MAGeCK. (C) dCas9-Caco2 cells were targeted for DPP4 (n=8) using two independent sgRNAs per gene or a control anti-GFP sgRNA (n=8), then assessed for infection using anti-HAstV capsid antibody 24hpi with HAstV1 or HAstV8. (D, E) HAstV1 levels in HEK293T (293T) (n=6) or HEK293T stably expressing FCGRT (293T-FCGRT) (n=8) or DPP4 (293T-DPP4) (n=6) at 24hpi. (F) 293T, 293T-DPP4 and 293T-FCGRT cells infected with HAstV1 stained with anti-HAstV capsid antibody (green) at 24hpi. Cell nuclei were stained with DAPI (blue). Representative images from two independent experiments (G) HAstV1 levels in H293T-DPP4 cells treated with 10μg/ml of isotype control (n=7) or 10μg/ml of anti-DPP4 antibody (n=7) prior to HAstV infection at 24hpi. Results from three independent experiments were analyzed using Kruskal-Wallis test with Dunn’s post-test (C) or Mann-Whitney test (D to E and G). **P=0.0034, ***P<0.0001 and *P=0.0157 (C), ***P<0.0001 and **P=0.0043 (D and E), ***P<0.0001 (G). Bars indicate mean of all data points. Scale bar = 100μm

Fig. 3.. FcRn interacts with HAstV spike.

(A) HAstV1 was tested against immobilized FcRn and mAb 8E7 via surface plasmon resonance. (B) Maturation process of HAstV VP90 structural protein. (C) VP25 protein from HAstV1 (left) and HAstV8 (right) was immobilized and tested against 2-fold dilutions of FcRn ranging from 8μM to 62.5nM via biolayer interferometry. (D, E) HAstV1 levels at 24hpi in Caco2 cells treated with PBS (Mock) (n=3–6), non-specific control protein (n=6–7), or soluble FcRN (s-FCRN) (n=7) or soluble DPP4 (s-DPP4) (n=6) prior to infection. Results from 2–3 independent experiments were analyzed using Kruskal-Wallis test with Dunn’s post-test (F and G). ***P<0.0001 (E). ns=not significant. Bars indicate mean of all data points.

Fig. 4.. FcRn and DPP4 inhibitors are antiviral against HAstV.

(A, B) HAstV1 and HAstV8 infection at 24hpi in Caco2 and 293T-FCGRT cells treated with PBS (n=6) or 50μg/ml of nipocalimab (n=6) prior to infection. (C, D) HAstV1 and HAstV8 infection at 24hpi in Caco2 or 293T-DPP4 cells treated with PBS (n=6–8) or 500μM sitagliptin (n=6–8) prior to infection. (E, F) HAstV1 infection at 24hpi in 293T (n=5), 293T-DPP4 (n=6) and 293T-FCGRT (n=6) cells treated with 10μg/ml of anti-DPP4 antibody (ab) or 50μg/ml nipocalimab (Nipo) prior to infection. (G) Percentage of anti-HAstV capsid antibody-positive Caco2 (WT) cells or DPP4^−/− or DPP4^−/− cells treated with nipocalimab 12h prior to infection at 24hpi. Results were analyzed using Kruskal-Wallis test with Dunn’s post-test (A-E) or Mann-Whitney (F) from two to three independent experiments. (H) HAstV1 levels at 24hpi in HIEs differentiated in monolayers on transwell inserts treated with PBS (n=7), 50μg/ml of nipocalimab (n=9), 10μg/ml anti-DPP4 antibody (n=6) and 500μM sitagliptin (n=6) prior to infection. Results from three independent experiments were analyzed using Mann-Whitney test (A to F) or Kruskal-Wallis test with Dunn’s post-test (G). ***P<0.0001 (A to D), **P=0.002 (F), **P=0.002 (G) **P=0.007 and ***P<0.0001 (H). ns=not significant. Bars indicate mean of all data points.