Emergence of novel reassortant H3N3 avian influenza viruses with increased pathogenicity in chickens in 2023

Dear editor,

Avian influenza viruses (AIVs) belong to the Alphainfluenzavirus genus (http://ictv.global) of Orthomyxoviridae family [1]. These AIVs are subtyped based on their combination of hemagglutinin (HA) and neuraminidase (NA) surface glycoproteins and there are 16 and 9 different subtypes, respectively [2]. These AIVs circulate in wild aquatic birds, which are considered the natural reservoir [2]. According to the classification of AIV pathogenicity by the World Organization for Animal Health, avian influenza is divided into highly pathogenic avian influenza (HPAI) and low pathogenic avian influenza (LPAI) [3]. With cross-over from wild birds into poultry, viruses of the H5 and H7 subtypes can evolve to become highly pathogenic, causing severe disease, devastating outbreaks and up to 100% mortality in chickens [4]. Most AIVs are classified as LPAI, causing minimal clinical signs or asymptomatic infection in birds, while some subtypes, such as H9N2, can cause obvious clinical symptoms such as respiratory disease and reduced egg production, which can lead to relatively large economic losses [5].

Influenza viruses of the H3 subtype can infect a variety of hosts, such as humans, pigs, horses, dogs, cats, seals, monkeys and birds including poultry and wild birds [6]. There have been three human cases of H3N8 subtype AIV infection in China since 2022, resulting in one death [7,8]. These cases, although sporadic, demonstrate that this subtype of AIV can be transmitted to mammalian hosts through an interspecific barrier and have raised concerns whether H3N8 AIVs will become a future major public health threat [9].

Since 2023, several egg farms in Eastern China have experienced cases of drops in food and water intake, with affected chickens showing respiratory signs including swelling of the sinuses and discharge from the eyes, nares, mouth, severe dyspnea and reduced egg production, with lower mortality rates. The trachea, lungs, liver, and intestines of dead chickens were collected and homogenized to extract DNA and RNA. The egg drop syndrome virus, AIV, Newcastle disease virus, infectious bronchitis virus, infectious bursitis virus and infectious laryngotracheitis virus were tested by real-time polymerase chain reaction methods (national or industry standards), with only AIV being positive. The H3N3 subtype was further identified using HA and NA specific primers [10] and sequencing. The viruses were isolated by inoculation into 10-day-old specific-pathogen-free (SPF) embryonated chicken eggs with homogenate and named A/chicken/China/NT322/2023(H3N3) (NT322/H3N3) and A/chicken/China/NT308/2023(H3N3) (NT308/H3N3), respectively. Throat and cloacal swabs of chickens from a nearby live poultry market were also sampled and one strain A/chicken/China/J1247/2023(H3N3) was isolated.

To investigate the origins of these H3N3 isolates, their genomes were sequenced using Sanger sequencing and phylogenetic analysis was performed. The phylogenetic trees for each gene segment were generated by using the neighbor-joining method in the MEGA 11 package. The bootstrap value was calculated with 1000 replicates. The eight gene segment sequences of the three strains shared 99.5 to 100% nucleotide identity among them, suggesting they were highly homologous. The HA genes of these viruses were highly homologous to those of H3N8 that caused the three human cases and shared 97.5 to 99.2% nucleotide identity between them (Figure 1(A)). Their NA genes were genetically close to those from H10N3 circulating in poultry in China and including the human case in 2021 in China (Figure 1(B)). The internal genes were genetically associated with the H9N2 circulating in chickens in China. The internal gene constellation was more similar to that of the human H10N3 isolate than that of the human H3N8 isolates, sharing 95.9 to 98.7% nucleotide homology with A/Jiangsu/428/2021(H10N3), except PA and M genes which were highly homologous with human H3N8 isolates at 98.8% and 99.5%, respectively.

Figure 1 . Phylogenetic trees of hemagglutinin (A) and neuraminidase (B) genes of the novel three H3N3 AIVs isolated from chickens in China, 2023. The trees were generated by neighbor-joining method with the MAGE 11 software. The viruses from this study were labeled in black circles, the H3N8 and H10N3 AIVs that caused human infections in China were marked in red. Scale bars indicate branch length based on number of nucleotide substitutions per site. (C) Replication of H3N3 and H9N2 viruses in chickens. The SPF chickens were inoculated with the representative viruses; organ samples were collected, and the viruses were titrated in eggs at 3 dpi. The dashed line in each panel indicates the lower limit of detection. (D) Shedding of H3N3 and H9N2 viruses in chickens. The SPF chickens were inoculated with the representative viruses; Oropharyngeal and cloacal swabs were collected from chickens at the indicated times, and viruses were titrated in eggs. OP: oropharyngeal swab; CL: cloacal swab. The dashed line in each panel indicates the lower limit of detection. (E) Histological lesions in the liver (a), kidney (b), spleen (c), thymus (d), lung (e), and bursa of Fabricius (f) of H3N3 infected chickens. Chickens were intranasally inoculated with 10⁶ EID₅₀ virus. Pathogenicity of the H3N3 virus in mice. Five-week-old BALB/c mice were infected intranasally with 10⁶ EID₅₀ virus. Percentage of bodyweight change of mice infected with NT322/H3N3 virus (F). The viral titers of the lungs and nasal turbinates of the infected mice collected at 3 dpi were measured in eggs (G).

Analysis of the HA amino acid sequence revealed the three H3N3 AIVs shared the same amino acid motif PEKQTR/GIF at the cleavage site, suggesting they were of low pathogenicity in chickens. The presence of the receptor binding sites Q226 and G228 (H3 numbering) in HA of all three H3N3 viruses, indicated that they would preferentially bind to SAα-2,3Gal avian-like receptors [11]. Several mutations that would increase polymerase activity and replication ability and increased virulence in mammalian cell lines or mice were identified, including I292V, I504V and A588V in PB2 and K356R in PA. Several mutations that would especially increase pathogenicity in chickens were identified, including K627E in PB2, I127V, I550L and F672L in PA, P64S and l69P in M2 and A42S and F103L in NS1 [12]. There was also a mutation of S31N in M2 which could increase the resistance to amantadine and rimantadine [13].

To investigate the replication ability and pathogenicity of these H3N3 viruses in chickens, five-week-old SPF chickens were inoculated intranasally with 10^6.0 median egg infective doses (EID₅₀) per 0.1 mL of NT322/H3N3. The A/chicken/Yunnan/KM224/2023(H9N2) (KM224/H9N2) virus was used as a control which was isolated from clinically infected chicken farms. All the chickens were monitored daily for clinical signs. Oropharyngeal and cloacal swabs were taken from the live chickens at 3, 5, 7, and 9 days post-infection (dpi). At 3 dpi, three chickens from each group were euthanized and the lungs, trachea, spleen, kidneys, brain, pancreas, and thymus were collected for viral titration in eggs for pathological assessment by staining with hematoxylin and eosin.

During the 14-day experiment period, although none of the chickens died, the NT322/H3N3-infected chickens showed depression and diarrhea and their feed and water intake decreased by about 25 percent. The KM224/H9N2-infected chickens exhibited no obvious clinical symptoms. Both NT322/H3N3 and KM224/H9N2 were detected in the oropharyngeal swabs at 3 and 5 dpi with peak titers at 5 dpi. However, NT322/H3N3 was not detected while KM224 was detected in the cloacal swabs at 3 and 5 dpi (Figure 1(C)). At 3 dpi, NT322/H3N3 was detected in all organs including the lungs, trachea, cecal tonsil, spleen, kidneys, pancreas, brain, bursa of fabricius, heart, liver and thymus, while KM224/H9N2 was not detected in the pancreas, bursa of fabricius, heart, and liver (Figure 1(D)). The results indicated that the H3N3 virus had stronger pathogenicity and replication ability in chickens than the H9N2 virus.

Chickens infected with NT322/H3N3 showed histopathological damage, including severe inflammatory cell infiltration and a small number of nucleated erythrocytes in the hepatic sinus space, a large number of red blood cells in the renal interstitium with glomerular atrophy and renal tubular epithelial cell degeneration, diffuse nucleated erythrocytes and inflammatory cell in the red marrow, a necrotic lesion caused by interstitial cell necrosis in the thymus, an increase in fat vacuoles between alveoli and mass erythrocytes in the alveolar interstitium and disintegration and necrosis of the bursal cortical cells (Figure 1(E)).

To evaluate the pathogenicity and virus replication in a mammalian host, five-week-old female Balb/c mice were intranasally inoculated with 50 μl of 10⁶ EID₅₀ of NT322/H3N3. At 3 dpi, three mice inoculated with NT322/H3N3 were euthanized. Their nasal turbinates, lungs, spleens, kidneys and brains were collected and titrated for virus infectivity in eggs. The weight and mortality of the mice were monitored. Over a 14-day observation period, the infected mice developed no obvious signs of illness and none died. The NT322/H3N3-infected mice experienced only slight weight loss (Figure 1(F)). At 3 dpi, virus replication was found in the turbinates and lungs at low titers, but not in the, spleen, kidney, and brain (Figure 1(G)).

Discussion

This study presents a preliminary analysis of the isolation and characterization of novel reassortant H3N3 viruses from chickens, which have undergone frequent reassortment. Their HA and NA genes are genetically close to the human infected isolates, H3N8 and H10N3, respectively. Their internal genes are derived from the H9N2 circulating in chickens in China and are more similar to that of the human H10N3 isolate. Our surveillance suggested that H9N2 accounts for about 72% of the AIVs identified in that region in 2023 and about 12.3% of birds in LBMs carry H9N2 viruses (unpublished data). Although these H3N3 viruses preferentially bind to avian-like receptors and were not pathogenic in mice, they were likely to possess molecular markers indicative of potential pathogenicity. Multiple human-infecting AIVs, such as H7N9, H10N8, and the recently emerging H3N8 and H10N3, have been proved to contain the internal genes of H9N2 viruses. Poultry that carry H9N2 AIVs provide a favourable environment for different subtypes of viruses to exchange their gene segments with H9N2 AIVs. Lesions caused by LPAIV, such as H9N2, are difficult to reproduce under experimental settings [14], but the H3N3 viruses isolated in this study exhibited similar clinical signs in SPF chickens as those in domestic production, indicating these H3N3 viruses have increased virulence, unlike the H3 virus in wild birds [15].

Although the H3 subtype is a low-pathogenic AIV subtype prevalent in domestic poultry, the emerging avian H3N3 virus from chicken flocks in China appears to exhibit an increasing pathogenicity. Close and continuous surveillance of the H3N3 subtype of AIVs circulating in poultry is required, their pathogenesis should be continuously investigated and their impact on the poultry industry in the future should be evaluated. Effective control measures including vaccination should also be evaluated to reduce the prevalence and decrease the economic impact of the virus infection in poultry, similar to prevention measures for H9N2 infection.

Ethics statement

The animal study was performed in accordance with the institutional animal guidelines and approved by the Animal Care Committee at China Animal Health and Epidemiology Center.

Disclosure statement

No potential conflict of interest was reported by the author(s).

Additional information

Funding

This work was supported by the National Key Research and Development Program of China [grant number 2021YFD1800201] and Shandong Provincial Key Research and Development Program [grant number 2022CXGC010606].