Application of metagenomics for diagnosis of broilers displaying neurological symptoms

Background Thirty-two-day-old broiler chickens at a farm located in northwestern South Korea displayed adverse neurological symptoms including limping, lying down, and head shaking. Approximately 2.1% of chickens died or were culled due to severe symptoms. Five carcasses were submitted to the Avian Disease Division of the Animal and Plant Quarantine Agency (APQA) for disease diagnosis. Results Broilers displayed severe pericarditis and perihepatitis associated with gross lesions. Broilers also displayed microscopic lesions in the cerebrum and in the granular layer of the cerebellum, which were associated with multifocal perivascular cuffing and purulent necrosis in the cerebrum, and severe meningitis with heterophil and lymphocyte infiltration. Staphylococcus spp. were identified in the liver and heart using bacteriological culture. PCR/RT-PCR assays revealed that broilers were negative for avian Clostridium botulinum, Newcastle disease virus, and avian encephalomyelitis virus. Bacterial and viral metagenomic analysis of brain sample further revealed the presence of Pseudomonas spp. and Marek’s disease virus, which are known etiological agents of chicken meningoencephalitis. Conclusions This study reports a diagnostic analysis of gross and histopathological lesions from 32-day-old broilers displaying unique neurological symptoms that revealed the presence of the several neurological diseases including meningoencephalitis. The causative agents associated with meningoencephalitis of broilers that had not been identified by routine diagnostic methods could be diagnosed by metagenomics, which proves the usefulness of metagenomics as a diagnostic tool for unknown neurological diseases in broilers. Supplementary Information The online version contains supplementary material available at 10.1186/s12917-023-03732-y.


Background
Genomic identification and characterization of pathogenic microorganisms (bacteria, viruses, fungi, and parasites) isolated from various hosts and the environment play an important role in the diagnosis, monitoring, and control of infectious diseases [1,2].Though identification of causative agents by isolation and culture is recognized as the gold standard, these methods are labor-intensive, and require between 1 week to more than 6 weeks to provide a diagnosis [3].Furthermore, many pathogens are difficult or impossible to culture using readily available technologies, hampering effective diagnostics [4].
PCR-based techniques are a highly sensitive, specific, and reproducible method to amplify target sequences to detect pathogens more quickly than classical culture methods [2].However, these methods can target only a limited number of pathogens using specific primers or probes; this limits their application to detect newly emerging pathogens, which continually accumulate point mutations, genomic rearrangements, or recombination events.To overcome these limitations, metagenomics is becoming a widespread alternative method to conventional diagnostic techniques [5].
Early use of metagenomics for diagnostics includes the discovery of severe acute respiratory syndrome coronavirus (SARS-CoV), and genetic profiling of cancer mutations in the early 2000s [6][7][8].Since 2005, Next Generation Sequencing (NGS) technologies have been increasingly applied to diagnostics in both humans and livestock, due to decreased sequencing cost and technological advancements.[9,10].In particular, the emergence of zoonotic diseases transmitted between humans and animals, such as SARS-CoV, Middle East Respiratory Syndrome coronavirus, or Coronavirus disease 2019, has led to widespread application of metagenomics in the field [11,12].
Neurologic symptoms in broiler flocks between 4 and 5 weeks of age are uncommon.In chickens, neurological symptoms are usually caused by infectious agents such as highly pathogenic avian influenza virus (HPAIV), Newcastle disease virus (NDV), avian encephalitis virus (AEV), and Clostridium botulinum (C.botulinum).Infection of chickens with bacteria such as Escherichia coli, although uncommon, can cause septicemia or otitis media that may eventually present with neurological symptoms [13].Furthermore, neurological symptoms can have non-microbial etiologies including use of veterinary pharmaceuticals such as anticoccidial drugs.This study presents a histopathological and metagenomic analysis of 32-day-old broiler chickens displaying neurological symptoms.

Gross findings and analysis of histopathological lesions
Gross observation of symptomatic broilers revealed airsacculitis, pericarditis, and perihepatitis in three of five chickens.Histopathological analysis demonstrated presence of acute hepatitis with multifocal necrosis, moderate serositis, myocarditis and severe fibrinopurulent pericarditis in the liver and heart of chickens with gross lesions (Fig. 1A and B C). Furthermore, we observed suppurative meningitis associated with lymphocyte and heterophil infiltration, and meningeal proliferation in the meningeal pia mater of the cerebrum and cerebellum in the absence of gross lesions (Fig. 1D).We also observed encephalitis with lymphocytosis in the cerebral cortex (Fig. 1E).Interestingly, we identified moderate multifocal mononuclear cellular infiltration, consisting mainly of lymphocytes and monocytes, in the perivascular areas of cerebral cortex, which are usually observed in the case of viral encephalitis (Fig. 1F).

Presumptive diagnosis
Following a traditional diagnostics protocol [14,15], we isolated Staphylococcus chromogense and Staphylococcus cohnii from liver and heart samples, respectively, which also contained histopathological lesions.However, these pathogens are not suggested as causative agents of neurological symptoms.All samples were negative for the common etiological agents of neurological symptoms, HPAIV, NDV, AEV, and C. botulinum, by PCR and RT-PCR (Additional file 1: Table S1).Histopathological analysis of the brain sample from this flock identified signatures of bacterial meningoencephalitis and viral encephalitis, but the causative agent was unclear.We therefore applied metagenomics to identify the etiological agent of meningoencephalitis.
The control sample generated 12,044,235 reads and 94.1% had a Q score ≥ 30.After trimming and QC, 11,872,300 reads were assembled into 20,881 contigs.We removed the host genome from the final consensus, and the remaining 9,658 contigs and 441,333 mapped reads were classified into three biological groups: no hits (640,787 reads, 98.80%), bacteria (7,165 reads, 1.10%), and eukaryote (591 reads, 0.09%, except for Gallus gallus).No viral sequences were detected in the control sample (Fig. 2A).

Phylogenetic analysis
Reads for the common chicken pathogens Marek's disease virus (MDV) and Fowl Adenovirus (FadV) were detected using viral metagenomic analysis.As the yield of reads was insufficient to confirm their genetic characteristics, we performed Sanger sequencing of target genes from both viruses.
In order to discriminate whether the obtained MDV sequences are from field or vaccine strains (CVI988, AF493551), partial sequences (short meq : 858 nt, long meq : 1035 nt) of the meq gene were aligned with 23 reference strains, and a phylogenetic tree was generated consisting of three groups.The meq gene of Group I and II was 177 bp shorter than that of group III, which consisted of vaccinal strains.Group I contained two branches, divided into the European virulent (ATE, C12/130) and Chinese virulent strains (LMS, YA, WS03, GX070060, G2).MDV strains identified in the 21AD109 sample were closely related to Kgs-c1 and Nr-c1 strains with 100% nucleotide identity in group I, known as the Japanese strain (Fig. 3A).
We also aligned a partial sequence (494 nt) of the hexon gene from FadV with 15 reference strains.Phylogenetic analysis revealed that the FadV in the 21AD109 sample was closely related to the TR59 strain with 100% sequence identity to serotype 8a(E) (Fig. 3B).

Discussion
Suppurative meningitis in the cerebrum and cerebellum is caused by bacterial infection of the central nervous system (CNS), accessed through the blood stream.We identified P. granadensis from the brain sample of broilers with suppurative meningitis by metagenomic sequencing of 16-23 s rRNA, suggesting that it may be the causative agent of suppurative meningitis of the broilers.P. granadensis was previously isolated from soil in Spain in 2015 and belongs to the P. fluorescens lineage [16].Pseudomonas are a ubiquitous genus of bacteria, commonly found in a wide variety of environments, and are also opportunistic pathogens that cause infections in immunocompromised individuals [17,18].Therefore, it is suggested that P. granadensis caused an opportunistic infection due to immunosuppression by MDV.
To infect the CNS, bacteria should first pass through the blood-brain barrier (BBB).Pathogens can cross the BBB transcellularly, paracellularly, and/or by Trojan-horse mechanisms [19].Studies on these infection modalities have been conducted using several strains in both in vivo and in vitro models [20][21][22][23].Pathogens can also cause BBB dysfunction by interfering with the expression of host factors such as chemokines or cytokines that affect cell adhesion, or by inducing cytotoxicity and apoptosis.These observations imply that all bacteria capable of crossing the BBB by these mechanisms have the potential to cause meningitis.
Our histopathological analysis revealed cuffing in the perivascular area of the brain, which is a typical lesion caused by viruses such as viral encephalitis [24].Viral metagenomic analysis identified MDV and FadV, and phylogenetic analysis confirmed the virulence profile and serotype of strains in clinical samples.
Marek's disease (MD) is a highly oncogenic, lymphoproliferative, immunosuppression, and neuropathic disease [25].Transient paralysis (TP) is acutely induced by MDV within 24-48 h and can lead to either mortality or recovery from early paralytic symptoms.Conversely, in our unique case, mortality occurred continuously for six days in a broiler flock.Histopathological signs of infection in the brain in conjunction with neurological symptoms such as paralysis after MDV infection have only been observed experimentally using in vivo models, and not in the field [26,27].According to the author's knowledge, our study may be the first recorded case of meningoencephalitis in broiler using metagenomics in the field.
FadV is separated into five genotypes (A-E) and 12 serotypes (1-12), some of which can cause inclusion body hepatitis (IBH), hepatitis hydropericardium syndrome (HHS), and gizzard erosion (GE) [28].Since we observed no lesions related to FadV in this study, this virus is an unlikely causative agent of neurological symptoms.We Fig. 3 Phylogenetic tree of the two avian viruses identified in this study.(A) A phylogenetic tree constructed through neighbor joining methods with 1,000 bootstrap replicates using the meq gene of MDV (◆: tested case).(B) A phylogenetic tree generated through maximum likelihood methods with 1,000 bootstrap replicates using the Loop-1(L1) region of the hexon gene of FadV (◆: tested case) may have detected FadV because it is commonly found in poultry farms.
Since the advent of NGS technologies, sequencing has become faster, cheaper, and more accurate.The development of these technologies has enabled comprehensive analysis using metagenomics and can now be applied to clinical samples.Since conventional methods such as PCR or RT-PCR can diagnose only a limited number of known pathogens, metagenomics will be useful for detecting unknown and emerging pathogens that are not routinely tested in diagnostic laboratories.As diseases such as encephalitis that originate in the CNS have a wide range of causative agents, the application of metagenomics rather than standard methods will be more effective for accurate diagnosis.

Conclusions
We performed a standard diagnosis based on gross and histopathological lesions in addition to a metagenomics to diagnose a unique case of mortality associated with neurological symptoms in 32-day-old chickens.We isolated S. chromogense and S. cohnii, the causes of pericarditis and perihepatitis, from the liver and heart using traditional culture methods, and identified causative agents of meningoencephalitis using metagenomics.
Our study may be the first field report demonstrating MDV and bacterial co-infection associated with clinical symptoms and histopathological lesions based on the author's knowledge.The combined use of metagenomics and traditional culture methods represent a paradigm shift in diagnostics.Our flowchart of diagnostic process may help people understand better to combine with metagenomics and traditional methods (Fig. 4).

Samples
Five carcasses from a farm located in northwestern South Korea were submitted to the Avian Disease Division of the Animal and Plant Quarantine Agency (APQA) for diagnosis in 2021.Neurological symptoms including limping, lying down and head shaking were observed in a flock of 32-day-old broiler chickens, resulting in 2.1% of this flock either dying or culled across a 6-day periods.As a results, gross lesions were characterized, and we performed necropsy, bacteriological culture, and virus isolation using specific-pathogen-free chicken embryonated eggs, following APQA diagnostic protocols [14].The trachea, cecal tonsils, kidneys, livers, and brains were collected from 5 broilers, and pooled by each organ.After necropsy, samples were named 21AD109.Specific-pathogen-free chicken brains were used as a control.The collected organs were fixed for 24 h in 10% neutral-buffered formalin for histopathological analysis.The remaining organs were homogenized in 10% phosphate-buffered saline and stored at -70℃ until further processing.

Bacterial examination
The surface of livers and hearts were seared with a hot spatula, and samples were obtained using a sterile swab.Swap samples were subsequently cultured on blood agar and MacConkey agar plates.After incubation overnight at 37℃, single colony was selected, and species identity was determined using the Vitek 2 system (bioMérieux, France) according to the manufacturer's instructions.

Genome samples preparation and PCR amplification
Total bacterial and viral genomic nucleic acid was isolated from homogenized samples using a Maelstrom 4000 DNA/RNA auto-extraction machine (TAN bead, Taiwan).PCR and RT-PCR assays were conducted using the specific primer pairs for C. botulinum, HPAIV, AEV and NDV as reported in previous studies [29][30][31].PCR products were analyzed under UV light after separation by 1.5% agarose gel electrophoresis.

Histopathological examination
Fixed tissue samples were trimmed to an appropriate size for the fixative container and washed in running water for 30 min to remove excess formalin.Trimmed tissues were dehydrated using a STP120 Spin Tissue Processor (Thermo-Fisher Scientific, USA) and embedded in molten paraffin using an embedding cassette (HYUNIL Lab-Mate, Korea).Section 5 μm thick were rehydrated with xylene before staining with hematoxylin and eosin (H&E) in a ST5010 Autostainer XL (Leica Biosystems, Germany).Finally, stained tissues were mounted on glass slides and were analyzed for the presence of histopathological lesions using a microscope.

Bacterial and viral metagenomic sequencing
Homogenized brain sample was centrifuged at 800×g for 5 min and 13,000×g for 10 min, respectively.Supernatants for viral metagenomic sequencing were passed through a 0.45 μm polyethersulfone membrane filter (TPP, Switzerland) to remove residual bacteria and debris.Virus in the residual supernatant was precipitated with 8% polyethylene glycol 6000 and 0.5 M NaCl, as previously described [32,33].Genomic DNA/RNA was extracted using a Maelstrom 4000 DNA/RNA autoextraction machine (TANbead, Taiwan).The extracted DNA/RNA sample was divided into two parts and treated with DNase I and RNase A for viral DNA/RNA enrichment, respectively, according to the manufacturer's protocols (QIAGEN, Germany).First strand cDNA synthesis was performed using PrimeScript™ 1st strand cDNA Synthesis Kit (Takara, Japan) according to the manufacturer's instructions.Viral DNA and cDNA were amplified using a random primer (5′-GCC GGA GCT CTG CAG AAT TCN NNN NN-3′) with a LongAmp® Taq 2X Master Mix (NEB, USA) [34].Illumina library preparation was performed using a Nextera XT DNA Library Preparation kit (Illumina, USA) according to the manufacturer's recommendations.Libraries were quantified using a Qubit fluorometer (Invitrogen, USA) and 4150 TapeStation system (Agilent Technologies, USA).Libraries were denatured and sequenced on an Illumina MiniSeq with a MiniSeq Reagent kit (Illumina, USA).
Homogenized brain sample for 16-23 s rRNA sequencing were amplified with previously modified primers 27 F (5′-AGA GTT TGA TCC TGG CTC AG-3′) and 2490R (5′-GAC ATC GAG GTG CCA AAC-3′) using the Maxime PCR PreMix Kit (Intron, Korea) with the following conditions: 5 min at 94 °C, followed by 30 cycles of 30 s at 94 °C, 30 s at 57 °C, and 1 min at 72 °C, and a final Fig. 4 Schematic of diagnostic process using traditional and metagenomics methods extension step at 72 °C for 10 min [35].The amplified products were purified using a QIAquick PCR purification Kit (QIAGEN, Germany) and an Oxford nanopore library was prepared using SQK-LSK109, following the manufacturer's protocol (Genomic DNA by Ligation, ONT).Sequencing was performed with a flowcell (R9.4) on the MinION MkIB for 12 h.

Phylogenetic analysis
To validate the viral metagenomic data, extracted DNA/ RNA was subject to PCR amplification targeting the meq gene of MDV and hexon gene of FadV, in line with previous studies [45,46].PCR amplification of MDV was performed in a 20 µL total reaction volume with a 2×Black PCR premix (Ventech Science, South Korea), 2.0 µL of primers, and 2.0 µL of template DNA.Conditions for amplification of MDV were as follows: initial denaturation at 94 °C for 5 min, 38 cycles of denaturation at 94 °C for 45 s, annealing at 55 °C for 45 s, and extension at 72 °C for 45 s, and a final extension at 72 °C for 5 min.The amplicon was purified and sequenced by Macrogen (Daejeon, South Korea).Sequences were aligned using Clustal W and phylogenetic trees were constructed using a maximum likelihood method with 1,000 bootstrap replicates in Mega X (V.10.2.6) [47].

Fig. 2
Fig. 2 Results of viral metagenomic and 16-23 s rRNA sequencing.(A) Metagenomic data from the Miniseq system revealed four biological groups.(B) Pseudomonadales was the most frequent taxon in the bacterial group.(C) The viral group consisted of avian associated viruses and bacteriophages.The control sample included no viral reads.(D) Results of 16-23 s rRNA sequencing demonstrated the presence of Pseudomonas spp.including Pseudomonas sp., Pseudomonas granadensis, and Pseudomonas qingdaonensis