Background
Avian influenza is a major zoonotic viral disease that causes significant adverse impacts on poultry production and the global trade [
1]. Previous outbreaks have caused the loss of hundreds of millions of birds, and total economic losses are estimated to be far in excess of US $10 billion [
2,
3]. Vaccination of poultry was implemented in many of the affected countries, especially in those where H5N1 viruses have become enzootic in poultry and wild birds [
3‐
6]. Mandatory vaccinations of chickens with inactivated or recombined H5N1 viruses are considered to prevent disease and mortality in chickens, reduce human cases and help to maintain rural livelihoods and food security [
7]. Moreover, vaccination is thought to be the most effective method to prevent influenza infection [
8].
One of the most interesting approaches to influenza immunization is the use of DNA vaccines. Among the many advantages of this technique, it is worth highlighting that DNA vaccines are fast to produce and modify; foreign antigenic protein produced within the host cells can induce humoral and cellular immune responses and DNA vaccination leads to immunization with an antigen likely to be folded in its native conformation, correctly glycosylated and having normal post-translational modifications [
9,
10]. This last feature is important, because the glycosylation of HA modulates among other host immune response [
11]. Moreover, they are safe due to the absence of infective agents and the possibility of using a single selected antigen, which allows for the differentiation of infection in vaccinated animals [
10,
12].
Many DNA vaccine candidates for protecting chickens from avian influenza have been described [
10,
13,
14], and one such vaccine has been licensed for use in the U.S. [
15]. Protective DNA vaccine candidates were also designed and tested by our team [
16‐
20].
Monitoring the transcriptome of vaccinated animals can allow for the discovery of vaccine-induced correlates of protection [
21,
22]. To our knowledge, there are only three articles describing the transcriptomic response of chickens vaccinated against avian influenza virus. Two of them showed the response of birds vaccinated with inactivated low-pathogenic H9N2 virus (A/Chicken/United Arab Emirates/99), with or without an adjuvant, and subsequently infected with homologous virus [
23,
24]. The third one, by our group, analysed the spleen transcriptome of broilers (Ross 308) vaccinated with two doses of protein (protein/protein), two doses of DNA (DNA/DNA), or the combined prime/boost (DNA/protein) vaccine against H5 avian influenza [
25].
Herein, we compare the previously reported changes in the transcriptomic profiles of the Ross 308 line vaccinated twice with DNA vaccine with the changes in the transcriptomic profiles of laying chickens of two lines, White Leghorn maintained under specific pathogen-free (SPF) conditions and Rosa 1 maintained under standard bedding conditions. Additionally, the transcriptomic profiles of Rosa 1 after one dose of DNA vaccine are presented and discussed.
Discussion
The effects of immunization were first verified at the humoral level. All groups of birds immunized with two doses of the vaccine responded by producing specific anti-HA antibodies. Then, we focused on changes in the spleen transcriptome, evaluated in each immunized group in relation to their respective controls. Immunization of broiler chickens (the Ross [2x] group) resulted in the largest number of DEGs, while vaccination of White Leghorn SPF chickens (the WL [2x] group) revealed the smallest number of DEGs (Fig.
2, Fig.
4). Interestingly, despite a low level of anti-H5 HA antibodies in sera, the Rosa [1x] group, immunized with one dose, showed more DEGs and higher fold-changes than the Rosa [2x] group, vaccinated twice (Fig.
1, Fig.
2, Fig.
4). All chickens vaccinated twice showed similar level of anti-H5 HA antibodies in sera (Fig.
1), however White Leghorn SPF chickens displayed significantly higher HI titre than Ross chickens (and HI titre of Rosa chickens was not determined). HI titre was negatively correlated with the number and Fold-Change range of DEGs. We believe that strong and quick secondary response initiated by memory cells can cause lower changes in gene expression in spleen at day 7 post vaccination.
Moreover, the microarray chips used in this work (Affymetrics Chicken Gene 1.1 ST Array) were built, according to the Affymetrix DataSheet, on the galGal3 genome founded on the Red Jungle fowl (
Gallus gallus), one of the main ancestors of domestic chickens (
Gallus domesticus). Some studies imply that broiler chickens are more closely related to Red Jungle fowl than layers [
27], which could be the reason for the higher number of DEGs detected in broilers (the Ross [2x] group) than in the layers represented by all remaining groups (Fig.
2). On the other hand, broilers may also show more DEGs because of breed selection. For example, the transcriptional profile of the breast muscle in heat-stressed layers was similar to that of broiler chickens kept at the control temperature, while heat stress amplified changes in broilers [
28].
The chicken genome was the first completed genome of a breeding animal [
29]; now, the fourth version (Gallus_gallus-5.0), corrected among others by annotation of 2768 noncoding genes and many CHIR loci, has been released [
30]. The number of chicken mRNAs seems to be lower than in humans [
31]; however, the chicken transcriptome seems to show a similar level of complexity [
32]. That explains the relatively high representation of RNA-encoding DEGs in our results. Moreover, mammalian long non-coding RNAs (lncRNAs) were reported to play critical roles in the immune response to influenza A virus infection [
33], and some lncRNAs were identified as being related to the immune response to influenza A virus in ducks [
34]; however, the roles of many non-coding RNAs remain to be discovered. Interestingly, the proportion of RNA-encoding DEGs of all DEGs is far less in broilers (the Ross [2x] group) than in all remaining groups (Fig.
3). In fact, the number of RNA-encoding DEGs in both groups of Rosa 1 chickens (Rosa [2x] and Rosa [1x]) is higher than that in broilers (the Ross [2x] group), despite a significantly smaller overall number of DEGs (Fig.
2, Fig.
4, Table S
6). The RNA-encoding DEGs reported in this study belong mainly to the miRNA and snoRNA classes (Table S
6). Deep sequencing of the transcriptomes of skeletal muscles from broiler and layer chickens showed that they share a few millions common miRNAs; however, tens of thousands miRNAs were still specific to either broiler or layer skeletal muscle [
35]. Interestingly, during that study the sequence tag annotations demonstrated that known chicken miRNAs and metazoan miRNA homologs accounted for about 50% of all sequence reads in the broiler and layer libraries, whereas snoRNAs were only slightly represented, although sequencing was in that case performed by fractionating total RNA using polyacrylamide gel electrophoresis to enrich for molecules in the range of 16–30 nt [
35]. Differential expression of many miRNAs between the lungs of broilers infected with H5N3 and those of non-infected animals was also reported previously [
36]. Some snoRNAs were differentially expressed between the various immunized groups (Table S
5). Despite the variety in chicken breeds and differences in maintenance conditions, some of these differences may be caused by the recognition of snoRNA derivatives by the Transcript Clusters optimized to detect snoRNA molecules [
37,
38]. Moreover, Transcript Clusters optimized for detection of certain mRNA can detect also regulatory RNA made from pseudogenes [
39]. This fact, together with different time points, can explain differences between expression level of some DEGs (e.g. ASS1, EFCAB4B and UTS2R) in various studied group.
Our study reports differential regulation of many ImmDEGs in spleens of chickens vaccinated with the experimental DNA vaccine (Table S
7). Among them one can find numerous DEGs encoding cytokines (CCLi9, IL5 and IL17D) and their receptors (CCR7, CCR8, CCR8L, CX3CR1, CXCR4, IL1R2, IL5RA, IL12RB2, IL17REL and IL28RA). To our knowledge, at least 10 of the ImmDEGs identified in this study (TLR2–1, IRG1, MX1, OASL, IFIT5, CXCR4, DDX60, NFKBIZ, IFNA and IL12RB2) were reported as differentially expressed during experimental infection of chickens or chicken cells with H5 influenza viruses. TLR2–1 was overexpressed in the lungs of chickens infected with the highly pathogenic AIV H5N1 strain A/Chicken/Jiangsu/k0402/2010 [
40] and was found by us as a DEG in the WL [2x] group. IRG1 was downregulated in CEF cells infected with H5N2 virus [
41] and was identified by us as a DEG in the Rosa [1x] group. The regulation of these two genes observed in this study seems to have proinflammatory consequences (Table S
7).
MX1, OASL and IFIT5 genes encode proteins with known functions in influenza defense [
41]. MX1 was the most overexpressed gene in the Rosa 1 [2x] group (Table
2), and its overexpression was also reported in CEF cells and chickens infected with H5N1 or H5N2 viruses [
40,
42] and in the lungs of chickens infected with H5N1 virus from 24 h post infection [
43]. The OASL gene was downregulated in the Ross [2x] group; however, it was upregulated in CEF cells infected with H5 viruses at 4 h post infection [
42]. Regulation of the expression of this gene in the lungs of chickens infected with highly pathogenic H5N1 virus depended on the time point [
40]. IFIT5 was overexpressed in CEF cells infected with H5 viruses at 12 h post infection [
40] and downregulated in the Ross [2x] group. Differences in expression of the OASL and IFIT5 genes may result from differences in the experimental setup. Regulation of CXCR4, DDX60 and NFKBIZ varied at different time points after infection of chicken lungs [
40]. IFNA was up- and IL12RB2 was downregulated in lungs of chickens infected with H5N1 virus [
43], whereas they were upregulated in the spleens of chickens vaccinated in this study (the WL [2x] and Ross [2x] group, respectively).
In summary, broiler chickens (Ross 308) showed a higher number and wider range of fold-changes in the transcriptional response than laying hens (White Leghorn or Rosa 1). Interestingly, White Leghorn SPF chickens had a lower number and lower range of fold-changes than the Rosa 1 breed. Moreover, the number and range of gene expression changes was higher in the Rosa 1 group that received one dose than in the Rosa 1 group that was boosted. In all groups many RNA-encoding DEGs and DEGs connected to the neuroendocrine-immune system were identified. Their representation was higher in laying chicken breeds than in broilers. Some genes (detected in this study) functionally connected to the immune response were also reported as differentially expressed during experimental influenza infection of chickens or chicken cells.
Publisher’s Note
Springer Nature remains neutral with regard to jurisdictional claims in published maps and institutional affiliations.