HPAIV H5N1 genotypes of clade 2.3.4.4b differ in virulence for ducklings
Phylogenetic analyses on current genotypes in combination with the epidemiological information reveal, that despite the diversity of simultaneously circulating genotypes, the current epizootic as a whole was dominated by single, individual genotypes (Fig. 1A) 20. For the phenotypical characterization of co-circulating HPAIV H5N1 clade 2.3.4.4b genotypes we focused on the last week of December 2021 (week 52), a time period when seven different genotypes of HPAIV H5N1 were present in Germany and compared the virulence to frequency of detection. Representative virus isolates of six out of these seven genotypes present in December could be obtained from samples obtained in 2021–2022 by cultivation on LMH-cells. Conformity with original samples was demonstrated by full- genome sequencing, verifying their distinct genotypes (Fig. 1B, genotypes #3 - #8). For comparison, two older H5N8 clade 2.3.4.4b viruses representing two further distinct genotypes (#1, #2) were included in the pathotyping approach. Both H5N8 genotypes circulated in Germany until April 2021, but have since than been replaced by H5N119.
In seven-day old pekin ducklings, both H5N8 viruses of genotypes #1 and #2 induced 100% mortality within three to six days pi, resulting in an IMPI score of 2.52 and 2.69, respectively (Fig. 1B; sFig 1), However, their pathogenicity was exceeded by genotype #3 (IMPI 2.92), a genotype emerged in October of 2021 and dominating the German HPAI epidemic with a total of 60 (43.8%) of the sequenced HPAIV H5N1 viruses up to the end of 2021. In frequency of detection genotype #4 was the second highest disseminated genotype, detected in 43 (31.4%) of HPAIV H5N1 viruses sequenced in 2021 (Fig. 1A; sTab 1). Compared to both precursor H5N8 viruses, IMPI scores of three genotype #4 virus isolates were significantly higher with scores of 3.00, 2.95 and 2.99, respectively (p 0.0002–0.0254, see sFig1) and inducing mortality within 2 days pi. These later three HPAI H5N1 2.3.4.4b viruses of the same genotype #4 have been recovered in February 2022 (#4(1)), April 2022 (#4(2) and July 2022 (#4(3)) and were included to evaluate consistency of IMPI scores. (Table 1). The results indicate, that the virulence of this genotype was stable and even slightly higher than genotype #3: Whereas for genotype #3, 7 of the 10 inoculated ducklings were alive on day one pi, two genotype #4 isolates induced death in all (#4(1)) and nine ducklings within one day, resulting in significant higher IMPI scores (p = 0.0031 and p = 0.0174). With regards to frequency of detection the order of the genotypes changed in the following year 2022, with 133 of 222 sequenced viruses belonging to genotype #4 (59.9%), in contrast to only 60 (27.0%) of the sequenced viruses from 2022 belonging to genotype #3. Nevertheless, during 2021 and 2022 genotypes #3 and #4, highly virulent for ducklings by IMPI, together made up 75 and almost 90% of all sequenced HPAI viruses in Germany.
Of indistinguishably high duck virulence was an isolate representing genotype #7, inducing 100% mortality within 24 hours (IMPI of 2.96). This particular virus was recovered from a wood pigeon (Columba palumbus) that was euthanized during an outbreak in a wild bird sanctuary in January 202241. However, this genotype remained a minor population, with one detection up to the end of 2021 and four in 2022, constituting 1.8% of sequenced viruses in the period 2021 and 2022.
The remaining three genotypes that co-circulated in December 2021 (#6, #8 and #9) remained rare detections, representing 10 (2.8%), 3 (0.8%) and 6 (1.7%) of the total number of viruses sequenced in Germany for the period up to the end of 2022. Obtained virus isolates from 2022 from genotypes #6, #8 and #9 were of moderate duck virulence with significant lower IMPI scores compared to dominate genotype #3 (#8 p = < 0.0001, #9 p = < 0.0001) and genotype #4 (p = < 0.0001–0.018) (sFig1): besides birds that succumbed to infection two to six days after inoculation, single birds showed only moderate or no clinical signs at all up to the end of the observation period, resulting in IMPI scores between 1.22 and 1.52. For example, in the group with the lowest IMPI, i.e. the genotype #8 virus, only four out of nine birds died or had to be euthanized because of severe signs of disease dominated by CNS disorders, while five birds had recovered from slight apathy at day 7 to 9 after infection (dpi) (Fig. 2; sFig 1).
HPAIV H5N1 clade 2.3.4.4b genotypes co-circulating in December 2021 (week 52) in Germany were characterized with respect to their duckling virulence by IMPI and compared to HPAIV H5N8, the subtype dominating HP epidemics until April 2021. Frequency of detection of genotypes with high (red star; IMPI > 2) and moderate virulence (light blue star, IMPI < 2) is given in relation to all different genotypes detected in Germany during 2021 and 2022 (A). Genetic composition is given by color code as previously described 20 with IMPI scores given above the virus cartoons (B).
Outcome of infection with moderate HPAIV H5N1 genotypes
Regardless of the group, all ducklings that survived until the end of the observation period, i.e. day 10 pi, had seroconverted by ELISA (sTable 4) proving they all had become infected. For ducklings of the genotype #8-infected group - showing the lowest IMPI value - also organ samples taken at the end of the experiment tested virus positive and viral RNA was detected in the brain of all 5 ducklings surviving until 10 dpi (9.1x103- 2.8x105 VE/ml). Likewise, ducklings inoculated with genotypes #6 or #9, that were sacrificed at the end of the ten-day observation period, showing no (n = 4) or only mild signs of depression (n = 4), harbored viral RNA in the brain (sTable 3). However, attempts to isolate infectious virus from late brain samples of genotype #8-infected ducklings failed. To a lesser extent, residual viral RNA was detected in the heart (4 of 5) the lungs (3 of 5) and the duodenum (2 of 5). The exemplary histological evaluation of 3 ducks, sacrificed 10 days after i.m. inoculation with genotype #8, revealed that virus antigen was still detectable multifocally in neurons and oligo focally in glia cells of the CNS in 2 out of 3 animals. Although clinical signs were lacking at 10 dpi, all ducks showed a moderate to severe, subacute, necrotizing meningoencephalitis predominantly in the cerebrum and brainstem, but hardly in the cerebellum (Fig. 3). No virus antigen was detected in the remaining examined tissues, however, 2 out of 3 ducks exhibited mild, chronic interstitial pneumonia, partially with cellular debris in the air sac and one animal showed minimal, chronic myocarditis. No lesions were recorded in the kidneys (data not shown). For details on virus antigen detection, see supplementary table (sTab10).
These results revealed striking differences in duckling pathogenicity between different genotypes of HPAIV H5N1 of clade 2.3.4.4b, indicating that epidemiologically dominant genotypes were more virulent for ducks. Interestingly, ducklings surviving infection up to day 10 dpi with genotypes with moderate virulence, although clinically inconspicuous, showed meningoencephalitis and virus antigen-positive neurons and glial cells in the brain, confirming a general neurotropism of the different HPAIV H5 genotypes.
Genotype-related distinct virulence for ducklings is also verified by the occulo-nasal inoculation route
To investigate the influence of the inoculation route on duckling virulence, one isolate of high (genotype #4 (3); IMPI: 2.99) and moderate duckling virulence (genotype #8; IMPI: 1.22) were tested in ducklings applying occulo-nasally inoculation route, mimicking a natural route of infection. Like after the i.m. inoculation, the virulence of the two genotypes differed dramatically in ducklings: Whereas genotype #4 induced mortality in all ten inoculated and both sentinel ducks, six out of ten ducklings inoculated with genotype #8 and one sentinel animal survived infection and showed no clinical signs after day 8 p.i. All seven surviving ducklings from genotype #8 group were subsequently observed for 21 dpi and remained clinically unremarkable and seroconverted (sTab 4;4.1). However, compared to the i.m. inoculation, the course of disease after o.n. inoculation was delayed resulting in lower clinical scores (CS) for both genotypes: CS of 2.34 vs. IMPI of 2.99 (p < 0.0001 *) for genotype #4 and CS of 0.65 vs. IMPI of 1.22 (p 0.0383 *) for genotype #8 (Fig. 2A). Concerning viral replication, the higher virulence of genotype #4 virus was associated with a faster increase of viral shedding in o.n. inoculated ducklings (Fig. 2B). Starting at 1 dpi already 9 out of 10 pharyngeal (1.75x103 to 3.46x107 VE/ml) and 10 out of 10 cloacal swabs (1.85x105 to 1.05x1010 VE/ml ) yielded virus.The highest viral load was detected in pharyngeal swabs from both sentinal animals with 1,02x1010 and 2,38x1010 VE/mL on 2 and 5 dpi respectively. Pharyngeal viral shedding remained on that high level up to the death of the animals.
In comparison, in the group of genotype #8 inoculated ducklings, 8 out of 10 pharyngeal swabs were virus positive on 1 dpi, but at a considerable lower level (5.13x103 to 2.87x105 VE/ml) than in the genotype #4 group. During the following days, pharyngeal viral shedding increased, with 10 out of 10 swab samples testing positive on days 3 and 4 pi, with the peak of viral shedding observed on day 4 (5,49x103- 2,15x107 VE/mL) remaining up to 28 times lower (sTab 2) compared to genotype #4 inoculated ducklings. Remarkably, cloacal shedding remained at very low levels in this group: within the first 7 dpi, only single cloacal swabs tested positive at low loads with significant differences in shedding between groups on day 3 pi (p 0.0143 *). Nevertheless, both sentinel animals of the genotype #8 group tested positive in pharyngeal swab samples from day 3 or day 6 pi, demonstrating successful virus transmission. Blood samples taken on the last day of the observation period, i.e. 21 dpi, confirmed seroconversion in all six surving inoculated and one sentinel animal of the genotype #8 group by ELISA (sTab 4) and also by HI. Testing with different antigens, sera showed preferential binding to homologous antigen with HI-titers between 6–8 [log2], compared to HI-titers between 4–6 [log2] to heterologous antigens from previous years (2014 and 2016) (sTab 5).
By occulo-nasal inoculation mimicking a natural route of infection the dualism of highly and moderately virulent genotypes of HPAIV H5N1 clade 2.3.4.4b was confirmed.
Clinical manifestations after occulo-nasal inoculation reveal dominance of neurological signs
Diseased animals inoculated with highly virulent genotype #4 had to be euthanized on day 3 because of progredient CNS disorders. Virological examination of these birds revealed presence of influenza A virus RNA in almost all tested organs (sTab 2), with the highest viral load detected in the kidneys, with virus loads ranging from 2.09x103 VE/ml to 1.86x106 VE/ml. Histopathological analysis of organs of three acutely deceased birds revealed an acute, necrotizing meningoencephalomyelitis in all birds with intralesional virus antigen detection particularly in the forebrain (score 4) and brain stem (score 1–3), and to a lesser extent in the cerebellum (score 0–3) (Fig. 3) and spinal cord with dorsal root ganglia (score 1–3 ) (sFig 2). Target cells comprised neurons, glia cells, ventricle and central canal epithelium. Furthermore, a severe, acute necrotizing rhinitis, sinusitis, air sacculitis and pneumonia was recorded, with intralesional virus antigen (score 2–3) detected in the olfactory and respiratory epithelium of the conchae, in the respiratory sinus and air sac epithelium as well as in the pulmonary bronchial, parabronchial and capillary epithelium (sFig 3; 3.1).
Likewise, organ samples from four ducklings that died in the moderately virulent genotype #8 group between day 5 and 6 p.i., had particularly high viral loads in the brain (2.53x105 to 4.77x108 VE/ml) and infectious virus was recovered from the euthanized birds. Also a sentinel animal to this group, that was euthanized on day 6 p.i., while suffering from central nervous disorder, had a particular high viral load in the brain ( 2.21x109 VE/ml) (sTab 2). Lower viral loads were also detected in the lungs, gastrointestinal tract, kidneys, heart, and especially liver, indicating systemic viral spread. Histological examination of three ducklings that succumbed during acute infection, identified a moderate, acute to subacute, necrotizing meningoencephalitis with oligofocal to diffuse antigen detection (score 1–4) showing the same distribution pattern and target cells as described after 3 dpi for genotype #4, except for the spinal cord, that appeared unaffected. Oligofocally, individual ducks tested positive for virus matrix protein in the submucosal glands of the beak, the sinus and air sac as well as in the myocardium associated with necrosis (sFig 3; 4). Although lacking virus antigen, the regenerative lesions in the olfactory and sinus epithelium (sFig 3) and mild interstitial pneumonia (not shown) indicate past infection. No findings were recorded for the kidneys at 5–6 dpi.
Remarkably, at the end of the experiment on day 21 p.i. residual amounts of viral RNA were still detected by RT-qPCR in the brain of four out of six inoculated surviving animals; but not in other organs tested (sTab 2). The evaluation of the H&E and IHC slides did not yield any pathologic findings or virus antigen in the remaining tissues (sFig 2–4).Similar to the IMPI group of the same genotype, at this late stage, no infectious virus was recoverable. Histological examination did not find any specific lesions in the brain (Fig. 3), but identified an irregular architecture of the olfactory mucosa and squamous metaplasia in the respiratory epithelium of the conchae in 2 out of 3 ducks, lacking antigen labeling (sFig 3). Again, this could be an indication of a past infection with regeneration of the affected mucosal areas.
Genotype-related molecular determinants distinguishing highly from moderately virulent genotypes in ducklings
Molecular comparison of the various HPAIV H5N1 genotypes, as depicted in Fig. 1B, reveals that only 3 genes of genotype #3, the virus that dominated the HPAIV H5N1 epidemic in December 2021, are present in all other viruses, i.e. besides the HA gene only NA- and M-gene of all investigated HPAIV H5N1 viruses are belonging to the same phylogenetic branch. However, the number of reassorted genes varied considerably for the different genotypes studied: For the other two HPAIV H5N1 genotypes with IMPI score of > 2, either a single gene (genotype #7: NP) or three genes coding for proteins that are part of the polymerase complex (genotype #4: PB2, PA and NP) differ compared to genotype #3. Likewise, in the group of viruses with moderate duckling pathogenicity (IMPI < 2), an exchange of up to 4 segments could be observed. In addition to genes encoding for proteins of the polymerase complex (PB2, PB1, PA and NP), the NS gene (genotype #9) is derived from genetically divergent branches as genotype #3. The single exception within the group of viruses with moderate duckling virulence, is genotype #8: Compared to highly virulent genotype #7, only the PB1 gene is exchanged. Even though genotypes #7 and #8 share a related NP-gene that differs from genotype #3, this exchange alone obviously did not affect the IMPI. This points to PB1 as an important determinant for the decrease of duckling virulence of genotype #8.
Phylogeny of the PB1 clearly distinguishes genotype #8 from the other HPAIV H5N1 viruses studied (Fig. 4A). This specific branch of genotype #8 (Fig. 4B) encloses HPAIV H5N3 from the same species than the studied H5N1 genotype #8 virus, i.e. red knots, circulating in 2020 and 2021 in Germany and France. Besides, closely related PB1 genes were present in a variety of subtypes of low pathogenicity like H1, H3, H4 or H10 that were detected in samples from wild birds, with the earliest detection in 2015 (A/duck/Mongolia/154/2015 (H1N2), grey dot). At protein level, genotype #8 virus accumulated altogether nine mutations in PB1 compared to genotype #7 (Fig. 4C). These mutations within the PB1 stretch all over the protein, but did not include the catalytic residues 446 or 447 (Fig. 4C). However, at position 180, located in the nuclear localization motif 42, glutamic acid in the highly virulent genotype #7 was replaced by a Glycine (E180G) in the moderately virulent genotype #8 (Fig. 4). Besides, aa substitutions in PB1-F2, a protein translated from an alternative open reading frame (ORF) of segment 2 were present. Altogether, deduced PB1_F2 proteins of genotype #7 and #8 differed in 10 of the 90 aa (11,1%), which is remarkable compared to only 9 different aa in the PB1 ORF (aa 1,2%; 755aa). In contrast, proteins of the other five gene segments had no (NP) or only single point mutations compared to genotype #7 (sTab 8).
This genetic comparison of natural occurring reassortants highlights that multiple genes may affect duckling virulence. Natural phylogenetic analysis in particular of highly and moderate duckling virulent genotype #7 and #8 respectively, indicate that point mutations within PB1 and PB1-F2 may attribute to this phenomenon of shift of duckling virulence.