Genome-wide analysis of differentially expressed mRNAs, lncRNAs, and circRNAs in chicken bursae of Fabricius during infection with very virulent infectious bursal disease virus

Background Infectious bursal disease virus (IBDV) causes acute, highly contagious, immunosuppressive, and lethal infectious disease in young chickens and mainly infects the bursa of Fabricius (BF). To investigate interactions between IBDV and its host, RNA sequencing was applied to analyze the responses of the differentially expressed transcriptional profiles of BF infected by very virulent IBDV (vvIBDV). Results In total, 317 upregulated and 94 downregulated mRNAs were found to be significantly differentially expressed in infected chickens, compared to controls. Long non-coding RNA (lncRNA) and circular RNA (circRNA) alterations were identified in IBDV-infected chickens, and significantly different expression was observed in 272 lncRNAs and 143 circRNAs. Gene Ontology and Kyoto Encyclopedia of Genes and Genomes pathway enrichment analyses were performed to assess the functions of significantly dysregulated genes, which showed that the JAK-STAT signaling pathway, the NOD-like receptor signaling pathway, and apoptosis may be activated by IBDV infection. We predicted interactions between differentially expressed genes and produced lncRNA-mRNA and circRNA-miRNA-mRNA regulator network. Conclusions The present study identified the expression profiles of mRNAs, lncRNAs, and circRNAs during vvIBDV infection and provides new insights into the pathogenesis of IBDV and antiviral immunity of the host.


Background
Infectious bursal disease virus (IBDV), a non-enveloped double-stranded RNA virus, is a member of the family Birnaviridae; it can cause acute, highly contagious, and immunosuppressive disease in chickens aged 3-6 weeks, leading to high mortality and considerable economic losses [1,2]. Serotype-I and -II strains of IBDV are recognized, and serotype I, which includes attenuated, classically virulent, and very virulent (vv) variants, can cause different degrees of pathogenicity and mortality in chickens [3]. IBDV predominantly targets the precursors of B lymphocytes, particularly those in the bursa of Fabricius (BF), an important immune organ, the infection of which may lead to B lymphocyte depletion and BF deterioration [4,5]. The severe immunosuppression caused by IBDV increases the susceptibility of chickens to other infectious diseases and reduces immune responses to vaccination [6]. Therefore, elucidating the interactions between IBDV and its host is a matter of urgency to determine effective strategies for preventing and controlling IBDV infections.
High-throughput sequencing technology and proteomic approaches have been used for this purpose, and various types of non-coding RNAs (ncRNAs) are also gaining increasing attention in this regard [7], including long non-coding RNAs (lncRNAs) [8] and circular RNAs (circRNAs) [9]. lncRNAs are longer than 200 nucleotides in length and can regulate gene expression at various levels, including epigenetics and transcriptional and post-transcriptional regulation. lncRNAs typically occur at low abundance and are frequently not conserved between species [10]. Moreover, lncRNAs have been displayed to affect viral replication by regulating the expression of antiviralrelated key genes [11][12][13], indicating that lncRNA might play a crucial role in antiviral responses. Cir-cRNAs form covalently closed continuous loops, which have been observed to be widely expressed in plants [14] and animals [15,16]. More importantly, studies have demonstrated that circRNAs influence viral infection by inhibiting viral factors [17,18]. Therefore, it is important to identify lncRNAs and circRNAs as well as their targets to understand the dynamics of gene regulation and effectively control the occurrence and development of disease.
Although many studies have been conducted to assess the effects of host lncRNAs on IBDV infection and its underlying molecular mechanisms [11,[19][20][21], the genome-wide expression effects and functional roles of lncRNAs and circRNAs during vvIBDV infection have not been examined so far. We investigated the expression profiles of mRNAs, lncRNAs, and circRNAs associated with vvIBDV infection of chicken' BF and constructed lncRNA-mRNA and circRNA-miRNA-mRNA co-expression networks, which may provide valuable information for new therapeutic approaches to control this disease.

Identification of lncRNAs and circRNAs in chicken bursa tissue
In total, 589,776,342 raw reads were obtained from the control and IBDV-infected bursa tissues. After data filtering and quality control, 584,284,990 clean reads of high quality were retained, and the proportion of clean reads ranged from 98.99 to 99.13% (Additional file 1: Table S1). Subsequently, reads mapping to ribosomal RNA (rRNA) were removed, and the proportion of retained reads ranged from 86.69 to 99.54% (Additional file 2: Table S2). After removing the rRNA, clean reads were mapped to the chicken reference genome ( Table 1). The transcripts were screened using Cufflinks software (v. 2.1.1) [22] based on the location of the transcripts on the reference genome, a transcript length of ≥200 bp, and an exon count of ≥2. In total, 4324 known lncRNAs transcripts (Additional file 3) were obtained from chicken bursa tissue, including 1957 (45.3%) intergenic lncRNAs, 706 (16.3%) bidirectional lncRNAs, 1442 (33.3%) sense lncRNAs, and (5.1%) 219 lncRNAs of other types. Nevertheless, intronic lncRNA and antisense lncRNA were not detected in our study. Moreover, 1086 novel lncRNAs were screened based on three protein-coding potential software analyses ( Fig. 1a; Additional file 4). The 1086 novel lncRNAs comprised 610 (56.2%) intergenic lncRNAs, 212 (19.5%) sense lncRNAs, 109 (10.0%) bidirectional lncRNAs, 76 (7.0%) antisense lncRNAs, and 79 (7.3%) lncRNAs of other types (Fig. 1b). No intronic lncRNAs were detected in the current study. Besides, anchor reads were mapped to the chicken reference genome, 7808 novel circRNAs were identified in the study (Additional file 5; Table 2).

Differentially expressed profiles of mRNAs, lncRNAs, and circRNAs
The number of mRNAs, lncRNAs, and circRNAs shared between the infection group and the control group is shown in Fig. 2. Based on the criteria p < 0.05 and fold change > 1.5 or 2, 411 mRNAs (Additional file 6), 272 lncRNAs (Additional file 7), and 143 circRNAs (Additional file 8) were considered differentially expressed (Table 3). Moreover, 317 mRNAs were upregulated and . Differentially expressed mRNAs, lncRNAs, and circRNAs were used for cluster analysis. A heat map indicated that the control and IBDV-infected individuals produced two separate clusters (Fig. 5). The expression patterns of the mRNAs, lncRNAs, and circRNAs thus differed between the two groups, suggesting that the differentially expressed genes (DEGs) in the chicken bursa tissue infected with IBDV were significantly different from those in the noninfected tissue.

Comparison of mRNA and lncRNA characteristics
In total, 44188 mRNAs and 5410 lncRNAs were identified in all samples. To examine the differences in the mRNAs and lncRNAs, genetic structure and sequence conservation was compared, and the distribution of the transcript length of the lncRNAs differed slightly from that of the mRNAs (Fig. 6a). However, there were fewer exons in the lncRNAs than in the mRNAs (Fig. 6b). In addition, the open reading frames of the lncRNAs were shorter than those of the mRNAs (Fig. 6c).

Functional annotation of DEGs
Gene Ontology (GO) and Kyoto Encyclopedia of Genes and Genomes (KEGG) pathway analyses were performed to analyze the 411 differentially expressed mRNAs and the target genes of the differentially expressed lncRNAs and cir-cRNAs to examine the functions of gene dysregulation during IBDV infection. GO enrichment of the biological processes (BPs), cellular components (CCs), and molecular functions (MFs) of the differentially expressed mRNAs, lncRNAs, and circRNAs is shown in Figs. 7 and 8. BPs associated with the enriched mRNAs predominantly included biological regulations, cellular processes, and single-organism processes; enriched mRNA CCs included cells, cell parts, and organelles; and significantly enriched MFs included binding and catalytic activities ( Fig. 7; Additional file 9). Based on the GO analyses of the differentially expressed lncRNA target genes, the most enriched BPs were cellular processes, biological regulation, and single-organism processes; the most enriched CCs were cells, cell parts, and organelles; and the most enriched MFs were binding and catalytic activity (Fig. 8a). GO enrichment analysis was performed for the antisense, cis, and trans roles of the target genes of the lncRNA, showing that target genes were also mainly enriched in cellular processes, cells, and binding (Additional file 10: Figure S1). The circRNA results were consistent with those of the lncRNAs (Fig. 8b).

Fig. 1
Screening and classification of lncRNAs in chicken BF. a The novel lncRNAs were mainly classified as intergenic lncRNAs, bidirectional lncRNAs, and sense lncRNAs. b CNCI, CPC, and the SwissProt database were used to analyze the coding potential of the novel lncRNAs. RNAs identified by all three analytical tools were chosen as candidate lncRNAs. c The novel lncRNAs were mainly classified as intergenic lncRNAs, sense lncRNAs, bidirectional lncRNAs, and antisense lncRNAs KEGG is the main public pathway-related database, and it has been used previously to determine the significantly enriched pathways of dysregulated gene products during IBDV infection [23,24]. The top 20 pathways associated with the mRNAs, lncRNAs, and circRNAs are shown in Fig. 9. The results show that mRNAs differentially expressed owing to IBDV infection were associated with the JAK-STAT signaling pathway, the NOD-like receptor signaling pathway, and the cytokine-cytokine receptor interaction signaling pathway, among others ( Fig.  9a; Additional file 11). According to the KEGG analyses of the target genes of the differentially expressed lncRNAs, spliceosome, JAK-STAT signaling pathway, ribosome, and Toll-like receptor signaling pathway were enriched (Fig. 9b). The target genes of lncRNA antisense, cis, and trans-regulation were mainly enriched in the notch signaling pathway, the glycosphingolipid biosynthesis-lacto complement, and spliceosome, respectively (Additional file 12: Figure S2). In the cir-cRNAs, the MAPK signaling pathway, protein processing in the endoplasmic reticulum, and ubiquitinmediated proteolysis were identified as predominantly enriched KEGG pathways (Fig. 9c).

Validation of differential gene expression by quantitative PCR
To validate the accuracy of the RNA-sequencing results, 10 differentially expressed mRNAs, lncRNAs, and cir-cRNAs were selected and quantified by reversetranscription quantitative PCR (RT-qPCR; Fig. 12); these RT-qPCR results showed trends similar to those of RNA sequencing.

Discussion
lncRNAs and circRNAs has emerged as regulator of gene expression and play important roles in various diseases [29,30]. Many lncRNAs and circRNAs of chicken have been identified and analyzed by high-throughput sequencing technology, and suggesting that aberrantly expressed lncRNAs and circRNAs in cells or tissues play a crucial role in virus-host interactions [31][32][33][34]. Interesting, classical IBDV infection affects lncRNAs expression in DF-1 cell has been testified [11]. Infectious bursal disease caused by IBDV is one of the most important infectious diseases that severely affect the poultry industry. Infection with vvIBDV strains results in high mortality in chickens, and vvIBDV appears to have emerged as the  predominant clinical disease type in nearly all poultryproducing regions of the world. Therefore, the BF of experimentally vvIBDV-infected chickens and control individuals were collected to study the differential expression profiles of mRNAs, lncRNAs, and circRNAs. In this study, 411 mRNAs, 272 lncRNAs, and 143 cir-cRNAs were considered differentially expressed. The results indicated that these DEGs may play a significant role in the vvIBDV infection process, suggesting that they may include potential therapeutic targets for treating IBDV infections. Host cells counteract the invasion of viral particles through a cascade of mechanisms, and IBDV infection elicits the expression of various genes. Previous study reported that the expression of interferon regulatory factor 8 (IRF8), TRIM25, IFIT1, Mx1; of STAT1, STAT4, and of Toll-like receptors, including TLR3 and TLR7, is increased in IBDV-infected chickens or DF-1 or DT40 cells [4,7,11,[35][36][37][38]. In our study, the JAK-STAT signaling pathway, the NOD-like receptor signaling pathway, the cytokine-cytokine receptor interaction signaling pathway, apoptosis, the chemokine signaling pathway, and the Toll-like receptor signaling pathway were significantly enriched according to the KEGG enrichment analyses of the respective differentially expressed mRNAs. IBDV exploits these pathways to induce the expression of STAT1, STAT3, STAT4, TRIM25, IFIT1, and Mx1 in the bursal tissue, and our results were in line with those of previous studies. In STAT1, a member of the STAT family, phosphorylation induces the expression of interferon-stimulating genes through a series of signal transduction steps to elicit antiviral mechanisms [39,Fig. 4 Volcano plots of the differentially expressed mRNAs (a), lncRNAs (b), and circRNAs (c). The red and green dots indicate the significantly upregulated and downregulated genes (p < 0.05), respectively. The black dots indicate the genes that were not significantly differentially expressed (p > 0.05) 40]. Interestingly, STAT1 can be regulated by many differentially expressed lncRNAs, suggesting that STAT1 may be an important regulator during IBDV infection in chickens. Additionally, STAT3, STAT4, TRIM25, IFIT1, and Mx1 may be involved in interactions between vvIBDV and the host.
The potential functions of lncRNAs are commonly predicted according to their target genes. Increasing evidence suggests that lncRNAs have important roles in adaptive or innate immune responses [41][42][43]. In a previous study, we found that loc107051710 suppressed the replication of IBDV by upregulating type I interferon production [11]. In the current study, it is worth noting that LOC107053928, LOC107054815, LOC107053352, and LOC107053557 were identified as regulated various target genes associated with immunomodulation, including STAT1, STAT3, STAT4, TRIM25, and IFIH1. TRIM25 is a member of the tripartite motif family of E3 ubiquitin ligases and has been demonstrated to play an important role in RIG-I antiviral pathways. Reportedly, TRIM25 can promote transcriptional upregulation of type I interferons (IFNs) by binding viral RNA-activated RIG-I [44,45]. It has been well established that type I IFNs plays a crucial role in the antiviral processes, where type I IFNs can activate the STAT1 phosphorylation after binding its ligand to induce interferon-stimulating genes (ISGs) expression, such as Mx1, OAS, and IFIH1, through the JAK-STAT signaling pathway [40,46]. Therefore, this implied that these lncRNAs and their target genes, STAT1, STAT3, STAT4, IFIH1 and TRIM25, might play a vital role in the IBD anti-viral response. Intriguingly, gene targeting studies have shown that STAT1 target genes can promote antagonizing proliferation and inflammation; however, STAT3 was the opposite [47,48]. Therefore, the expression levels of STAT1 and STAT3 may reflect the balanced response of the  organism. We believe that further insight into the roles of LOC107053928, LOC107054815, LOC107053352, and LOC107053557 is very important for understanding the complex regulatory mechanism of gene expression in response to vvIBDV infection in chicken BF.
CircRNAs, a newly discovered class of ncRNAs, can affect the prognosis of diseases, especially tumors [49,50] by regulating the activity of target genes and by participating in the regulation of gene transcription and protein production [51,52]. In the current study, 63 upregulated and 80 downregulated circRNAs were identified, and their expression levels were generally lower than those of the mRNAs and lncRNAs. Most circRNAs in normal and cancer tissues in humans also occur at low abundance and may thus be by-products of pre-mRNA splicing [53,54]. Additionally, circRNA can act as a competing endogenous RNA (ceRNAs) that impair miRNA activity through sequestration, thereby upregulating miRNA target gene expression [55]. In the study, circRNA-miRNA-mRNA network was constructed and showed that the regulatory relationships between cir-cRNAs, miRNAs and target mRNAs were complex. In the network, circRNAs novel_circ_000574 and novel_ circ_001469 were expressed 131,072-fold and 286,862fold higher levels, respectively, during IBDV infection. Their potentially affected target genes involved in  immune-related included STAT1 and IRF7 by binding to miR-1587-x and miR-4507-y respectively, suggesting that these circRNAs also play an important role in IBDV pathogenicity.

Conclusion
Many signaling pathways were found to be involved in IBDV infection, particularly the JAK-STAT signaling pathways; however, further research is needed to assess their effects on the pathogenesis of IBDV infection. Additionally, we constructed lncRNA-mRNA and circRNA-miRNA-mRNA co-expression networks and predicted that LOC107053928, LOC107054815, LOC107053352, and LOC107053557 may affect viral replication by regulating STAT1, STAT3, STAT4, IFIH1 and TRIM25 expression. Our results provide new insights into the pathways and mechanisms that mediate host immune responses to vvIBDV. Further studies are needed to explore the biological functions of LOC107053928, LOC107054815, LOC107053352, and LOC107053557 during vvIBDV infection.

Study animals and viruses
Three-week-old specific-pathogen-free (SPF) White Leghorn chickens were obtained from the Harbin Institute of Veterinary Medicine (Harbin, China). A vvIBDV strain isolated from chicken BF and maintained in our laboratory was used in this study (strain LJ-5).

Sample collection and preparation
Eighteen chickens were randomly assigned to two groups (control group and vvIBDV infection group) with nine individuals per group, and the two groups Fig. 11 CircRNA-miRNA-mRNA regulatory networks. The red triangles, blue diamonds, and purple ellipses indicate the differentially expressed circRNA, miRNAs, and mRNAs, respectively were housed independently. Chickens in the control group were injected with phosphate-buffered saline (PBS; PH = 7.4), and the vvIBDV infection group chickens were inoculated with the LJ-5 strain virus at a dose of 10 3 × 50% embryo lethal dose (ELD) per 0.2 mL via eye-nose drops [56,57]. The chickens were housed at an animal facility under pathogen-free conditions and were provided a standard diet and water. Moreover, these chickens were observed 2-3 times daily. On the third day after inoculation, three chickens from each group were randomly selected and euthanized via T-61 intravenously (0.4 ml/kg) to isolate the BF. The animal remains were treated innocuously according to the requirements of school animal welfare management. The procedures used in this experiment were approved by the Institutional Animal Care and Use Committee of Northeast Agricultural University (2016NEFU-315, 13 April 2017). All sections of this report adhere to the ARRIVE Guidelines for reporting animal research (Additional file 16). The National Centre for the Replacement, Refinement, and Reduction of Animals in Research (NC3Rs) is an independent scientific organization that minimize the use of animals in research and improve animal welfare and help overcome the challenges and limitations of the use of animals in research to the benefit of the whole bioscience community [58]. Samples were placed in liquid nitrogen and stored at − 80°C. All instruments were treated with DEPC before use to remove RNases.

RNA isolation, library preparation, and sequencing
Total RNA was isolated using Trizol reagent (Life Technologies, Carlsbad, USA), and the purity and integrity of the RNA were measured using a Nanodrop micro-spectrophotometer (Thermo Fisher Scientific, Waltham, USA) and an Agilent 2100 Bioanalyzer (Agilent Technologies, Santa Clara, USA), respectively. RNA samples qualified for analyses under the following conditions: concentration ≥ 100 ng/μL, RNA integrity number ≥ 7.0, and RNA 260/280 ratio of 1.

Filtration of raw sequencing reads
Raw reads containing adapters or bases of low quality affect assembly and analysis; therefore, to obtain clean high-quality sequences, raw reads were filtered to remove adapters, low-quality reads, and poly-N reads.
Clean high-quality data was used for further analyses.

RNA sequence mapping and transcriptome assembly
The short reads alignment tool Bowtie 2 (v. 2.2.8) [59] was used to map the sequences using an rRNA database. After removing the rRNA reads, the sequences were mapped to the Gallus gallus GRCg6a reference genome using TopHat2 (v. 2.1.1) [60]; Cufflinks (v. 2.1.1) [22] was used to assemble mapped reads independently with a reference annotation-based transcript assembly  [61] and CPC [62] were used to assess the potential protein-coding functions of novel lncRNAs. Transcripts were mapped using the SwissProt database to assess protein annotation. The intersection of the respective results was chosen as lncRNAs. Phast (v. 1.3) [63], phyloFit [64], and PhastCons [65] were used to evaluate sequence conservation in the transcripts, calculate phylogenetic models among species, and compute the conservation scores of coding genes and lncRNAs, respectively. For circRNA, unmapped reads were extracted from the above results, and the ends of the unmapped reads were intercepted (default 20 bp) to identify the anchor reads, which were then mapped to the reference genome. The results were processed using find_circ software to predict circRNAs.
Prediction of lncRNA and circRNA target genes lncRNAs are involved in many post-transcriptional regulation processes, as are some other small RNAs such as microRNA (miRNA), the regulation of which depends on complementary base pairing. Some antisense lncRNAs may regulate gene silencing, transcription, and mRNA stability. To assess the interactions between lncRNAs and mRNAs, RNAplex software [66] was used to predict the complementary correlation of antisense lncRNAs and mRNAs. This software program includes the Vienna RNA package, and best base pairing predictions were based on a calculation of minimum free energy in the thermodynamic structure. Moreover, the cis and trans target genes of the differentially expressed lncRNAs were predicted. For the cis target genes, the mRNAs and the genomic location of the lncRNAs were mapped. We searched coding genes less than 100 kb upor downstream of each lncRNA and then analyzed their functions. For the trans targets, correlations between the lncRNA and protein-coding gene expression was analyzed using Pearson's correlation coefficient, and protein-coding genes with absolute correlation values > 0.94 were screened. The differentially expressed lncRNAs (fold change > 1.5 and p < 0.05) and predicted target genes (fold change > 2 and p < 0.05) were chosen to construct a co-expression regulatory network, which was visualized using Cytoscape software (v3.6.0) [67]. The binding sites of the miRNAs on the circRNAs and mRNAs were predicted using mirTarBase software [68], and the target relationships of miRNAs-mRNAs and miRNAs-circRNAs were assessed accordingly. Subsequently, we synthesized targeted relationships between  the miRNAs and mRNAs and between the miRNAs and circRNAs to identify the mRNAs associated with the circRNAs.

Functional annotation of DEGs
To understand the functions of the differentially expressed transcripts, including the mRNA and the target genes of the lncRNAs and circRNAs, the GO (http:// www.geneontology.org/) and KEGG (http://www.kegg. jp/kegg/) databases were used to perform GO and KEGG enrichment analyses, respectively. GO terms and KEGG pathways with p values < 0.05 were considered significantly enriched.

RT-qPCR validation of the RNA-sequencing results
To assess the accuracy of the sequencing results, five upregulated and five downregulated mRNAs, lncRNAs, and circRNAs were selected and quantified by RT-qPCR. Primers were designed using oligo6 software (Tables 4, 5 and 6). Total RNA was extracted from vvIBDV-infected and uninfected bursal tissue. RT-qPCR was performed using SYBR Green Master (Roche, Mannheim, Germany) and an ABI 7500 Real-Time PCR system (Applied Biosystems, Waltham, USA). β-Actin was used as an internal control of the mRNA. The 2 −ΔΔCt method was used to calculate the relative expression levels of the target genes. Experiments were conducted using three replicates.

Statistical analyses
GraphPad Prism5 software was used to analyze the RT-qPCR results with one-way ANOVA. Data are shown as the means ± standard error of the mean. Statistical significance was considered at p < 0.05.