Avian Influenza a H9N2 Viruses in Morocco, 2018–2019

Low pathogenic H9N2 avian influenza (LPAI H9N2) is considered one of the most important diseases found in poultry (broiler, laying hens, breeding chickens, and turkeys). This infection causes considerable economic losses. The objective of this work was to monitor and assess the presence of avian influenza virus (AIV) H9N2 in eight different regions of Morocco using real-time RT-PCR, and to assess the phylogenetic and molecular evolution of the H9N2 viruses between 2016 and 2019. Field samples were collected from 108 farms suspected of being infected with LPAI H9N2 virus. Samples were analyzed using H9N2-specific real-time RT-PCR. Highly positive samples were subjected to virus isolation and seven isolates were fully sequenced. Low pathogenic H9N2 avian influenza virus was introduced in Morocco in 2016. We show that in 2018–2019, the virus was still present irrespective of vaccination status. Phylogenetic and molecular analyses showed mutations related to virulence, although our viruses were related to 2016 Moroccan viruses and grouped in the G1 lineage. Specific amino acid substitutions were identified in Moroccan H9N2 viruses that are believed to lead to increased resistance to antiviral drugs.


Introduction
Low pathogenic H9N2 avian influenza is an emerging disease that causes important economic losses in the poultry sector and is considered a threat to both poultry farms and public health.
Being a member of the genus Alphainfluenzavirus, and of the family Orthomyxoviridae, avian influenza viruses are enveloped RNA virus, with a genome composed of eight negative-sense RNA segments [1]. They are classified as low or highly pathogenic, on the basis of their virulence and hemagglutinin (HA) and neuraminidase (NA) sequences [2]. There are currently18 HA and 11 NA identified, including the bat-specific H17-H18 and N10-N11 [3]. Avian influenza virus subtype H9N2 is pathotyped as a low pathogenic virus (LPAI). However, co-infections with other pathogens can lead to severe outbreaks with high mortality rates and severe economic losses [4,5].
H9N2 LPAIV had first been described in 1966 in a turkey farm in the USA [6]. Since then, it has been reported in numerous countries around the world. Between 1992 and 1994, an H9N2 outbreak occurred in Guangdong Province, China, and affected chicken farms with a mortality rate of 10% to 40%, with a reduction in egg-laying rate of 14% to 75% [7].
In 1996, H9N2 LPAIV was reported in South Korea [8]. In 1998, it was isolated from most provinces in China and, as a result, it was considered to be one of the most widespread avian influenza virus in China [9]. Other countries in the Middle East and North Africa have been affected by this virus including Iran [10], Saudi Arabia [11], Jordan [12], the United Arab Emirates [13], Tunisia [14], Egypt [15], Sultanate of Oman [16], and Libya [17].
Phylogenetic analysis of the genome of LPAI H9N2 strains allowed to classify this virus into two distinct lineages: Eurasian and American. Though the Eurasian lineage contains several clades, most of the strains detected were classified in two clades (G1 and Y280) [18]. The G1 clade is represented by the A/Quail/Hong Kong/G1/1997 prototype virus, which mainly circulates in South China, Central Asia and the Middle East, while the Y280 clade viruses circulate throughout China and are represented by the A/Duck/Hong Kong/Y280/1997 prototype.
The main sources of LPAI H9N2 infections are the domestic and wild avian species. Wild birds are considered one of the reservoirs of the virus, and can transmit it over long distances. Transmission of the H9N2 virus can occur through direct contact with infected animals, and the infection can spread between farms through the movement of infected birds, vehicles, contaminated equipment or people with contaminated shoes or clothing [19]. In Pakistan, sparrows were shown to play a very important role in the transmission of the virus between farms [20]. In general, the sensitivity and receptivity of H9N2 is strongly dependent on the avian species (chicken and turkey). However, other species raised for consumption and/or hunting, such as guinea fowl, quail, pheasant, partridge, duck, goose, and ostrich are also considered sensitive. LPAI H9N2 virus has also been reported to be transmissible to mammals including dogs and cats [21] as well as humans [22][23][24][25].
The LPAIV H9N2 strain (SF1, GenBank accession number SCA48100) introduced in Morocco in January 2016, belongs to the G1 lineage, and is closely related to viruses circulating in the Middle East. As a response, the competent authorities authorized vaccination of any type of chicken as the best choice to limit the rapid spread of this disease [26]. However, in order to guide veterinarians to a rational choice of vaccines, it is important to determine and to phylogenetically analyze the circulating strains.
The aim of this study was to monitor the presence of LPAI H9N2 viruses in farms where animals with respiratory signs are reported using real-time reverse transcription PCR. Sequencing of isolates was performed in order to detect potential mutations that might affect the efficacy of commercial vaccines.

Specimen Collection
In collaboration with private veterinarians, a total of 151 samples, which included organs (trachea, lungs) and tracheal swabs, were collected from 108 commercial broilers farms (vaccinated and non-vaccinated) in eight regions of Morocco. Our sampling was based on chickens suspected of being infected with LPAI H9N2 virus, and presenting respiratory signs (rales, sneezing), associated with a decrease in food consumption and a drop in production. The specimens were collected in a period of 11 months, from 28/06/2018 to 31/05/2019.

RNA Extraction and Real Time RT-PCR
RNA extraction was performed using the NucleoSpin ® RNA Virus Kit (Macherey-Nagel, Düren, Germany, No. 740956.250), following the manufacturer's instructions. In order to detect the H9N2 virus, the extracted RNA was amplified on the 7500 Fast Real-Time PCR System thermal cycler (Applied Biosystems, Foster City, CA, USA), using the primers and probe for generic detection of H9 subtypes described by [27], which target a conserved region in the HA2 subunit of the HA gene sequence.

Virus Isolation
In order to obtain a maximum viral load detectable by conventional RT-PCR and for full genome sequencing purposes, 17 samples among those with the highest Ct in RT-qPCR, from 17 different farms, were grown on 10-day-old, specific pathogen free (SPF) embryonated eggs. Briefly, the eggs were mirrored and the air chamber was delimited. The viral inoculums were prepared by mixing 0.2 mL of the viral suspension, 0.6 mL of sterile PBS and 0.2 mL of antibiotic OXY-Kel 20 L.A (oxytetracycline) and injected via allantoic cavity route using a sterile needle into the air chamber of the embryonated eggs. After viral inoculation, the eggs were incubated at 37 • C and examined daily for five days to assess the viability of the embryos. After the death of the embryo, eggs were refrigerated at 4 • C for 4 h. Then, the lesions on the embryos were observed and the allantoic fluids were collected, clarified, and stored at −80 • C until use.

Full Genome Amplification and Sequencing of H9N2 Moroccan Isolates
Viral RNA was extracted from allantoic fluids harvested from the 7 SPF embryonated eggs with the highest viral load, using the Macherey Nagel kit (Duren, Germany, No. 740956.250). Whole genome sequencing of 2018-2019 Moroccan isolates was performed with an Illumina MiSeq System (Illumina, San Diego, CA, USA) as previously described [28]. The preparation of libraries was performed using an Illumina Nextera XT library prep kit (FC-131-1096) (Illumina, San Diego, CA, USA) following the manufacturer's instructions. A tape station was used to verify the library quantity and quality. CLC genomic workbench was used for genomes assembly. The nucleotides sequences of all characterized strains in this study are submitted in the GenBank database under accession numbers summarized in Table 1.

Sequences and Phylogenetic Analyses
Bioedit 7.2.5 software [29] and ClustalW (version 1.83) [30] were used to compare and align nucleotide sequences of the complete genomes of seven Moroccan H9N2 isolates.
The phylogenetic tree was constructed by the maximum likelihood method, using the Mega 6.06 software [31]. The Blast [32] and Bioedit programs [30] were used to determine the sequence identity and compare the Moroccan strains with those selected from Genbank.

Statistical Analysis
Statistics describing the correlation between H9N2 positivity and the different factors: regions and vaccination status were calculated for each variable, including the mean and percentage distribution of frequencies. A non-parametric test (chi-squared test) was used to calculate the correlation between the H9N2 frequency in farms and their vaccination status.

Case History and H9N2 Detection
One hundred and fifty-one samples from respiratory tissues and tracheal swabs were collected between June 2018 and May 2019 from different areas of Morocco: Fes-Meknes, Rabat-Sale-Kenitra, Casablanca-Settat, Draa-Tafilalet, Benimellal-Khenifra, Souss-Massa, Marrakech-Safi, and the eastern region. The samples were tested by real time RT-PCR to detect the presence of influenza virus. A total of 83 were positive for AIV with cycle threshold (Ct) values varying from 12 to 39 (Table A1), of which 40%, 56%, and 4% of the samples had a Ct below or equal to 25, between 25 and 35, and above 35, respectively. The epidemiological survey resulted in a positivity rate of the disease of 58% (63 positive farms out of 108 sampled farms) ( Table 2).

Vaccination Status
The positivity rate (relative prevalence) of LPAI H9N2 positive farms was estimated to be 50% in unvaccinated farms (32 positive farms out of 64 chicken unvaccinated farms tested), while it was 70% in vaccinated farms (31 positive farms out of 44 chicken vaccinated farms tested). The overall vaccination rate reached 41% (44 out of 108 farms tested against H9N2) ( Table 2).
The presence of LPAIV H9N2 was detected differently between the groups of vaccinated and unvaccinated farms, but the difference was not statistically significant (95% CI, p value: 0.9).

Molecular Characterization and Phylogenetic Analysis of the Eight Viral Segments
Viruses from highly positive samples were isolated. The genome of 7 of them was fully sequenced with an IlluminaMiSeq System [28]. Phylogenetic analysis showed that our Moroccan H9N2 viruses isolated from chickens were in the same cluster as the other Moroccan viruses detected in 2016, and grouped into G1 lineage. They were compared with relevant virus sequences available in GenBank.
Based on HA and NA phylogenetic trees, our Moroccan viruses were closely related to viruses previously isolated in the Emirates  The sequence analyses of the seven Moroccan isolates showed several substitutions in both HA and NA sequences when compared to 2016 strain SF1 (Tables A2 and A3).
All seven Moroccan isolates had the RSSR*GLF motif at the HA cleavage site, which is a characteristic and signature of the low pathogenic H9N2 viruses.
Among mutations associated to resistance to antiviral molecules, no changes from SF1 strain were identified for NA H274Y substitution [37] or M2 S31N mutation [38] (Table 3).

Discussion
Our analysis included 151 field samples from 108 poultry farms suspected of being infected by the LPAI H9N2 virus and presenting mainly respiratory signs as well as poor zootechnical performances (decrease in production, decrease in feed consumption and mortalities), as reported previously [39].
The results revealed that 58% of the samples were positive for LPAIV H9N2. However, we cannot extrapolate these results on the epidemiological profile of the LPAI H9N2 virus in Morocco since our sampling was not representative enough of the national territory and only 8 regions out of 12 (Fes-Meknes, Rabat-Sale-Kenitra, Casablanca-Settat, Draa-Tafilalet, Béni Mellal-Khenifra, Souss-Massa, Marrakech-Safi and the Oriental) were sampled.
The spread of LPAI H9N2 virus in Morocco can be explained mainly by the movement of farmers, workers, and feed suppliers without compliance with biosecurity rules, to which can be added the transport of live chickens [13,15,26]. It should be noted that the application of sanitary biosecurity measures in broiler farms has been shown to be insufficient to prevent the entry of the virus [26]. In addition, mutations associated with resistance to antiviral molecules are still present in our LPAI H9N2 strains. The M2 S31N mutation is known to increase resistance to antiviral molecules, especially amantadine and rimantadine [38]. Likewise, other studies have shown that the absence of the H274Y substitution in the NA protein can confer to the virus a sensitivity to neuraminidase inhibitors such as oseltamivir [37].
The positivity rate of AI H9N2 positive farms was estimated to be 50% in unvaccinated chicken farms (32 positive farms out of 64 chicken-unvaccinated farms tested), while it reached 70% in vaccinated chicken farms (31 positive farms out of 44 chicken-vaccinated farms tested). A recent study evaluating the efficacy of four different commercial vaccines on H9N2 LPAIV SF1 strain has shown that they conferred a very limited protection against the infection [40]. Park and collaborators indicated that vaccination against H9N2 virus coupled with continuous infection of vaccinated flocks is an advantage for mutant viruses selection [41], whereas other studies report that vaccination decreases viral pressure in the field by reducing the level and duration of viral shedding [42]. Other explanations include the quality of the vaccine (either that it was not of the same strain as the virus currently circulating, or that it had a low antigen concentration [43,44]). We were not able to gather sufficient information relative to the vaccines used and their composition to be able to conclude on this point.
In addition, the vaccination rate against H9N2 was low (41%), which can be explained by the high cost of vaccination or by the fact that some farmers consider the vaccine is ineffective.
For unvaccinated specimens that tested negative, despite respiratory signs, other respiratory diseases, including BI or NDV, might be the cause of the observed clinical signs.
In this study, we demonstrated a relationship between our viruses, isolated in 2018-2019, and those isolated in Morocco in 2016, which all belong to the G1 lineage [26]. In order to evaluate the evolution of the Moroccan H9N2 virus over time (i.e., after its first introduction into Morocco), phylogenetic and genetic analyses were carried out.
On the HA and NA phylogeny, the 2018-2019 Moroccan viruses were close to those from Emirates (2015) This similarity can be explained by the common border between Morocco and Algeria, and by the history of commercial exchanges within western Africa countries. The evolution of the influenza virus directly depends on its genomic properties, which leads us to follow and verify the presence of possible mutations over time (especially on the HA and NA genes, which are the main proteins targeted by antibodies). Moroccan viruses harbor several mutations in HA and NA. Some have already been characterized, such as HA Q226L, which is known to enhance binding to mammalian-like receptors [45]. For other mutations, further studies are necessary in order to determine whether they could affect the virulence of the virus in poultry, or increase transmissibility to human. Potential glycosylation sites were identified in our Moroccan strains. As compared to Moroccan strains isolated in 2016, there was one amino acid change within a glycosylation site (297 NISK→NVSK, H3 numbering throughout) in four of the seven sequenced isolates. Changes in glycosylation sites may affect the host range and virulence of influenza viruses [46], though we do not know if it is the case here.

Conclusions
The low pathogenic avian influenza virus H9N2 is endemic within the country despite vaccination. Biosecurity issues in farm management, combined with high mutation potential are likely to cause dynamic changes in LPAI H9N2 strains. This prompts us to propose appropriate surveillance and adaptation of vaccines to circulating strains in order to better understand and fight public health risks.         I  T  T  Y  S  A  I  I  I  R  I  T  S  S  V  S  I  K  V  N  S  K  T  S  N  I  AS13  I  T  T  Y  S  A  I  I  I  R  I  T  S  S  V  S  I  R  V  D  S  K  N  A  S  I  AS14  I  T  T  Y  S  T  T  T  T  R  I  P  L  S  V  N  V  K  V  N  S  T  T  S  N  M  AS29  I  T  T  H  P  A  I  I  I  K  I  T  S  S  V  S  I  R  V  N  S  K  T  A  S  I  AS32  I  I  T  Y  S  A  I  I  I  R  T  T  S  S  I  S  I  K  V  N  S  K  T  S  S  I  AS71  M  I  T  Y  S  A  I  I  I  R  T  T  S  A  V  S  I  K  A  N  F  K  T  S  S  I  AS76  M  I  T  Y  S  A  I  I  I  R  T  T  S  A  V  S  I  K  V  N  F  K  T  S  S  I  AS77  I  I  M  Y  S  A  I  I  I  R  T  T  S  S  I  S  I  K  V  N  S  K  T  S  S  I