Creating resistance to avian influenza infection through genome editing of the ANP32 gene family

Abstract

Chickens genetically resistant to avian influenza could prevent future outbreaks. In chickens, influenza A virus (IAV) relies on host protein ANP32A. Here we use CRISPR/Cas9 to generate homozygous gene edited (GE) chickens containing two ANP32A amino acid substitutions that prevent viral polymerase interaction. After IAV challenge, 9/10 edited chickens remain uninfected. Challenge with a higher dose, however, led to breakthrough infections. Breakthrough IAV virus contained IAV polymerase gene mutations that conferred adaptation to the edited chicken ANP32A. Unexpectedly, this virus also replicated in chicken embryos edited to remove the entire ANP32A gene and instead co-opted alternative ANP32 protein family members, chicken ANP32B and ANP32E. Additional genome editing for removal of ANP32B and ANP32E eliminated all viral growth in chicken cells. Our data illustrate a first proof of concept step to generate IAV-resistant chickens and show that multiple genetic modifications will be required to curtail viral escape.

Fig. 1. Breeding strategy for homozygous ANP32A^N129I-D130N chicken.

a ANP32A editing strategy: two nucleotide changes (red letters) introduce asparagine (N) position 129 (N129I) and aspartic acid (D) position 130 (D130N) missense mutations. The third nucleotide change (green letters) is a synonymous mutation in the gRNA PAM and serves as a marker control for allelic contribution from the male and female surrogate hosts. b Male and female PGC cultures were derived from the blood of individual chick embryos. The PGCs were edited, and clonal lines of GE PGCs were propagated and analysed. GE PGCs were differentiated into fibroblast-like cells for IAV polymerase assays. To generate GE chicks, GE PGCs were mixed with B/B dimerisation compound (to induce cell death of host embryo germ cells) and injected into iCaspase9 host embryos, which were incubated to hatch. After hatching, the surrogate hosts were raised to sexual maturity and directly mated. All offspring from eggs laid by the surrogate hosts were biallelic for the edit and contained the parent-specific PAM nucleotide change. c the activity of reconstituted IAV polymerase was assessed in fibroblast-like cells derived from ANP32A^knockout (Knockout), ANP32A^N129I-D130N (N129I-D130N) or wild-type (WT) PGCs. Cells were transfected with avian IAV polymerase (PB2/627E - black bars) or human-adapted isoforms (PB2/627K - grey bars), Firefly minigenome reporter and Renilla reporter control plasmids and then incubated at 37 °C for 48 h. Wild-type chicken ANP32A (chA) cDNA was co-expressed with minigenome plasmids to rescue polymerase activity in ANP32A^N129I-D130N cells. Data shown are Firefly activity normalised to Renilla plotted as mean ± SEM derived from (n = 3) three independent experiments each consisting of three technical replicates. Error bars represent standard error of mean (SEM). One-way ANOVA and Dunnett’s multiple comparison test were used to compare polymerase activity in the GE cells with polymerase activity in WT cells. Unpaired two-tailed t-test was used to compare ANP32A^N129I-D130N and ANP32A^N129I-D130N +chA data. Statistical annotations are defined as *P ≤ 0.05, ****P ≤ 0.0001. d Image: wild-type (WT) hen (left) and homozygous ANP32A^N129I-D130NGE hen (right, blue ring on right shank).

Fig. 2. Assessment of low-dose IAV infection in ANP32A^N129I-D130N chickens.

a Schematic of low-dose in vivo challenge of 2-week-old chickens with H9N2-UDL influenza A virus (A/chicken/Pakistan/UDL01/08). Chickens were housed in negative pressure poultry isolators. Prior to challenge all birds were bled from the wing vein to obtain pre-infection sera. Groups of ten WT (black) chickens or ten ANP32A^N129I-D130N (white) chickens were intranasally inoculated with 1 × 10³ PFU of H9N2-UDL virus per bird. Uninoculated sentinel chickens were introduced into the isolators 24 h post infection to assess for transmission from the directly inoculated birds. Oropharyngeal cavities of each bird were swabbed daily from the day of inoculation (D0) until day 7 (D7) post-inoculation. Infectious virus titre in swabs was measured by plaque assay on MDCK cells (b, c). c Bird ID number for directly inoculated ANP32A^N129I-D130N birds above the detection limit is indicated. DL detection limit of 10 PFU/ml for plaque assay.

Fig. 3. Assessment of high-dose IAV infection in ANP32A^N129I-D130N chickens.

a Schematic of high-dose in vivo challenge of 2-week-old chickens with H9N2-UDL influenza A virus (A/chicken/Pakistan/UDL01/08). Chickens were housed in negative pressure poultry isolators. Prior to challenge all birds were bled from the wing vein to obtain pre-infection sera. Groups of ten WT (black) chickens or ten ANP32A^N129I-D130N (white) chickens were intranasally inoculated with 1 × 10⁶ PFU of H9N2-UDL virus per bird. Uninoculated naive sentinel chickens were introduced into the isolators 24 h post challenge (day 1 pi) to assess for transmission from the directly inoculated birds. Oropharyngeal cavities of each bird were swabbed daily from the day of inoculation (day 0) until day 7 post inoculation. Infectious virus titre in swabs was measured by plaque assay (b, c). c Bird ID number for plaque-positive directly inoculated ANP32A^N129I-D130N birds are indicated. Bird #941 was one of four directly inoculated ANP32A^N129I-D130N birds culled on day 3 pi for post-mortem examination. DL detection limit of 10 PFU/ml for plaque assay.

Fig. 4. Assessment of mutations identified in polymerase genes of viruses isolated from infected ANP32A^N129I-D130N chickens.

a, b Influenza A virus (H9N2-UDL) polymerase harbouring single or combinations of PA and PB2 mutations detected in virus isolated from ANP32A^N129I-D130N chickens was reconstituted together with NP by plasmid transfection in eHAP1 human cells that lack ANP32 expression and complemented with chicken (a) or human (b) ANP32-FLAG proteins. Polymerase activity was measured at 24 h post-transfection by Firefly luciferase signal generated from a minireplicon and a Renilla luciferase transfection control. Data shown are Firefly activity normalised to Renilla plotted as mean ± SEM derived from three (n = 3) independent experiments, each consisting of three technical replicates. Error bars represent SEM. Data statistically analysed by one-way ANOVA and Dunnett’s multiple comparison test to determine polymerase constellations whose activity varied from wild-type H9N2-UDL polymerase. Statistical annotations are defined as *P ≤ 0.05, **P ≤ 0.01, ***P ≤ 0.001, ****P ≤ 0.0001. c WT embryonated eggs were inoculated with 100 PFU of wild-type H9N2-UDL (H9N2-UDL^WT) virus or the double mutant variant (H9N2-UDL^{PA-E349K PB2-M631L}) containing the PA-E349K and PB2-M631L mutations. The inoculated eggs were incubated at 37.5 °C. Allantoic fluids were collected at the indicated timepoints. Data are PFU/ml in allantoic fluids measured by plaque assay and statistically analysed by multiple unpaired two-sample T-test. Statistical annotations are defined as ns not significant. d Human airway epithelial cells were infected with human-adapted H1N1 virus (A/England/195/2009) (England/195) or H9N2-UDL^WT virus or H9N2-UDL^{PA-E349K PB2-M631L} and incubated at 37.0 °C. Cell culture supernatants were harvested at the indicated timepoints and titrated by plaque assays. Data was statistically analysed by multiple unpaired two-sample T-test to compare growth of England/195 virus with growth of H9N2-UDL^WT virus or H9N2-UDL^{PA-E349K PB2-M631L} at each timepoint. Data was analysed by one-way ANOVA and Dunnett’s multiple comparison test. Statistical annotations are defined as *P ≤ 0.05, **P ≤ 0.01. e WT embryonated eggs were inoculated with 100 PFU of a mixture of H9N2-UDL^WT and H9N2-UDL^{PA-E349K PB2-M631L} virus containing <10% of the mutant virus. The inoculated eggs were incubated at 37.5 °C. Allantoic fluids were collected at the indicated timepoints and followed by viral RNA purification. Next generation sequencing was performed on purified viral RNA to determine variant frequency in each egg. f Human airway epithelial cells were infected with a mixture of H9N2-UDL^WT virus and H9N2-UDL^{PA-E349K PB2-M631L} virus containing <20% of the mutant virus and incubated at 37.0 °C. Cell culture supernatants were harvested at the indicated timepoints and followed by viral RNA purification. Next generation sequencing was performed on purified viral RNA to determine variant frequency at each timepoint. g, h Location of amino acids mutated in virus isolated from ANP32A^N129I-D130N chickens in the asymmetric polymerase dimer in combination with chANP32A (PDB:6XZP) (g) or the symmetric polymerase dimer (PDB:6QXB) were plotted using ChimeraX (h). g Influenza virus asymmetric polymerase dimer (PDB: 6XZP) showing ANP32A in light green, amino acids 129 and 130 highlighted in purple, PB2-627K in red, PB2-M631L in orange, PA-E349K in yellow, PA-Q556R in dark green and PA-T691I in blue. h Influenza virus symmetric polymerase (PDB: 6QX8) showing PB2-627K in red, PB2-M631L in orange, PA-E349K in yellow, PA-Q556R in dark green and PA-T691I in blue.

Fig. 5. Deletion of ANP32 A, B, and E eliminates viral polymerase activity and viral proliferation.

a ANP32A deletion strategy: two gRNAs were used to generate a 15-kb deletion in ANP32A in male and female PGCs. Clonal lines of ANP32A-knockout (AKO) PGCs were isolated, propagated and injected into iCaspase9 host embryos which were incubated to hatch. After hatching, the surrogate hosts were raised to sexual maturity and directly mated. All offspring from eggs laid by the surrogate hosts were biallelic for ANP32A deletion (see Supplementary Fig. 14). b, c WT or ANP32A^N129I-D130N or AKO 11-day-old embryonated eggs were inoculated with wild-type H9N2-UDL (H9N2-UDL^WT) virus or the double mutant variant (H9N2-UDL^{PA-E349K PB2-M631L}) containing the PA-E349K and PB2-M631L mutations. The inoculated eggs were incubated at 37.5 °C. Allantoic fluids were collected 48 h later and PFU/ml measured by plaque assay. DL detection limit of plaque assay (10 PFU/ml). Data were statistically analysed by unpaired two-tailed T-test of transformed data (Y = (Log(Y)). Statistical annotations are defined as ns not significant, **P ≤ 0.01, ****P ≤ 0.0001. d PGCs were edited to delete ANP32A, ANP32B, or ANP32E or combinations of the deletion (see Supplementary Fig. 17). PGCs were subsequently differentiated into fibroblast –like cells and used to assay polymerase activity and viral replication. e Wildtype (WT) H9N2-UDL polymerase or the mutant isoform harbouring PA-E349K and PB2-M631L mutations was reconstituted together with NP by plasmid transfection into chicken PGC-derived fibroblast-like cells. Polymerase activity was measured at 48 h post-transfection by detection of Firefly luciferase signal generated from a minireplicon normalised to a Renilla luciferase transfection control. Data shown are Firefly activity normalised to Renilla plotted as mean ± SEM derived from three (n = 3) independent experiments, each consisting of three technical replicates. Data was statistically analysed by one-way ANOVA, and Dunnett’s multiple comparison test to compare polymerase activity in wild-type cells with activity in other cell lines. Error bars are SEM. ns= not significant, *P ≤ 0.05, ****P ≤ 0.0001. Statistical annotations are defined as ns not significant, *P ≤ 0.05, ****P ≤ 0.0001. N129I-D130N cells refer to cells with the homozygous ANP32A^N129I-D130N genotype. f, g PGC-derived fibroblast-like cells were infected with a recombinant virus harbouring the HA and NA genes of the H1N1 PR8 virus, and the polymerase and other internal genes from the highly pathogenic H5N1 avian influenza virus A/turkey/Turkey/2005 (Tky05) (f) or a highly pathogenic H5N1 clade 2.3.4.4b virus (A/chicken/Scotland/054477/2021) (g). Cell culture supernatants were harvested at the indicated timepoints and titrated by plaque assays. Data was statistically analysed by one-way ANOVA and Dunnett’s multiple comparison test to compare virus growth in wild-type cells with virus growth in other cell types at each timepoint. Statistical annotations are defined as ns=not significant, *P ≤ 0.05, **P ≤ 0.01, ***P ≤ 0.001, ****P ≤ 0.0001.