Background

Influenza is one of the commonest respiratory infectious diseases of humans. Influenza A virus (IAV) is a negative-sense, single-stranded, enveloped RNA virus that belongs to the influenza A genus of the Orthomyxoviridae family. The threat of IAV to human health is significant and continual, which could be in the form of seasonal epidemics and occasional pandemics [1]. It has been estimated that more than 50 million people globally have died from influenza [2]. The highly pathogenic avian influenza (HPAI) H5N1 has infected thousands of people and spread globally in transported poultry or via indirect contact [2).

Table 2 Summary of RNA-Seq data

Global changes in expression in response to host genes

RNA isolates were individually prepared from both H5N1 virus and mock infected A549 or 293 T cells. To identify their functions during viral infection, RNA deep sequencing was performed to analyze the total RNA profiles in uninfected or H5N1 infected A549 and 293 T cells. As shown in Fig. 2a, a total of 1299 promoted and 1422 suppressed differentially expressed genes (DEGs) were identified in A549 cells infected with A/QH virus. In contrast, 293 T cells showed relatively few DEGs, with 144 promoted and 9 suppressed genes (Additional file 1: Table S1).

Fig. 2
figure 2

Identification and characterization of IAV infection. a A549 or 293 T cells were infected with A/QH viruses at an MOI of 0.01 for 8 h. Total RNA was extracted and used for RNA-seq analysis. The expression values shown in shades of yellow and green indicate gene level above and below the median expression value across all the samples (log scale 2, from − 2 to + 2), respectively. b The 20 highest fold-changes of the RPKM value of gene expression in A549-QH and 293 T-QH was sorted (log scale 2, from A549-QH to 293 T-QH). c qRT-PCR analysis of the expression of selected genes in A/QH virus-infected A549 or 293 T cells relative to uninfected controls. The fold-difference was measured by the 2-△△Ct method; RNA levels were normalized to corresponding β-actin. Error bars represent standard deviation. d Venn diagram showing the distribution of shared differentially expressed genes in 293 T and A549 cells infected with A/QH virus

To compare gene expression profiles between A549 and 293 T cells, we filtered out genes with low expression (FPKM = 0.001), calculated the fold-changes in RPKM value of each gene from the two samples, we used the statistical enriched and listed the top 20 fold-changes by the absolute value of log scale 2 obtained from both cell lines (Fig. 2b). The Venn diagram shown in Fig. 2c illustrates the overlap of A549 and 293 T cells responding to A/QH virus infection induced DEGs, which may help to investigate the various influences induced by virus infection. Up-regulated genes in A529 and 293 T cells included 2721 genes and 153 genes, respectively, with 79 genes common to both cell types. These results indicated a difference in human cells upon virus infection, with some shared alterations in expression among different cell types.

Confirmation of DEGs by quantitative RT-PCR

By performing quantitative RT-PCR (qRT-PCR) on selected DEGs, the differential expression profiles of genes obtained by RNA-seq analysis were validated. Although minor differences were observed between these two types of analysis due to their intrinsic differences, the results of these two analyses demonstrated the same relative regulation of DEGs. Consistent with virus propagation over time, most of the DEGs displayed a higher level of mRNA in A549 cells than that in 293 T cells (Fig. 2d).

pro-inflammatory cytokines and chemokines production induced by A/QH infections have been considered to be linked to elevated morbidity and mortality in humans. As shown in Fig. 2d, the expression of tumor necrosis factor-alpha was not changed in infected A549 or 293 T cells. However, the levels of AKR1C3, TXN, and the interferon response-related genes IFIT2, IFIT1, as well as ISG15 were markedly elevated in A/QH-infected A549 cells in comparison with infected 293 T cells. The expression of RIN1 and ELOVL3 in 293 T cells was higher than in A549 cells. These findings suggested a remarkable initiation of the response of the host gene in A/QH-infected A549 cells.

GO and KEGG pathway enrichment analyses based on DEGs

By importing datasets representing genes with changed expression profiles obtained from RNA-seq analyses for the analyses of GO and KEGG pathway enrichment, we examined possible biological interactions of DEGs and determined important functional networks by A/QH infection in human cells. The results of GO analysis of the five most common the subclasses of cellular processes are presented in Tables 3 and 4. The most common cellular process was biological regulation in infected A549 cells and chronic inflammatory response to antigenic stimulus in infected 293 T cells. The cellular process, which was in the biological process category, was most significantly regulated by A/QH infection.

Table 3 Gene ontology (GO) enrichment for differentially expressed genes (DEGs) upon A/QH virus infection in 293 T
Table 4 Gene ontology (GO) enrichment for differentially expressed genes (DEGs) upon A/QH virus infection in A549

Furthermore, DEGs were mapped into the KEGG pathway database to further explain the individual function analysis. Hundreds of signaling pathways was found to be enriched in cells, among which 32 pathways in 293 T cells and 60 pathways in A549 cells were found to be significantly altered (Q-value< 0.05 and p < 0.01). The top 20 enriched pathways in A/QH-infected A549 or 293 T cells are summarized in Fig. 3. Metabolic pathways and purine metabolism signaling pathway are high associated with the responses to A/QH infection in 293 T cells (Fig. 3a). In A549 cells, many receptor-interaction signaling pathways were enriched, including the cytokine-cytokine receptor interaction, the Toll-like receptor-downstream pathway, and the Jak-STAT signaling cascade (Fig. 3b).

Fig. 3
figure 3

Top 15 enriched pathways based on DEGs in A/QH-infected cells. Pathway analysis allowed the construction of a scatter plot of KEGG pathway enrichment statistics for DEGs following A/QH infection of 293 T cells (a) and A549 cells (b). Rich factor is the ratio of the number of differentially expressed genes noted in the pathway terms to all gene numbers noted in this pathway term. A greater Rich Factor indicates higher intensiveness. Q-value is the corrected p-value ranging from 0 to 1 (blue). A lower Q-value indicates higher intensiveness

GO analyses have also been employed to reveal the possible functions of the identified unique gene transcripts in cell samples. With the examinations of the differential unique genes in the A/QH group relative to cellular function and metabolic processes in A549 cells, we found that the most enriched biological processes mainly include the oxidation-reduction process, defensive response to viral invasion, and the signaling pathway mediated by type I interferon (Fig. 4), with alcohol dehydrogenase (NADP+) activity and dehydrogenase activity as the most significant molecular functions (Additional file 2: Table S2 and Additional file 3: Table S3).

Fig. 4
figure 4

GO functional enrichment of unique differentially expressed genes in A549 cells. GO enrichment of unique differential expression genes indicated that biological processes and molecular function were most enriched, judging by their p-values

Discussion

The highly pathogenic avian IAV strain, H5N1, can cause severe disease and death, especially in older individuals in poor health with other underlying diseases [16]. Host barriers have hindered the efficient growth of influenza virus from one species in other species [17]. The receptor specificity of hemagglutinin and the internal genes of influenza virus are critical in host-range restriction. In addition to the virus itself, the host restriction protein has an irreplaceable role during virus replication [18, 19]. Airway epithelial cells are major target cells for H5N1 virus infection. Human A549 cells (wildly used cell lines as pulmonary epithelial cells) serving as virus-susceptible cells and 293 T cells (used as human kidney cells) serving as virus-insusceptible cells were used in the present study. H5N1 A/QH virus was used to infect A549 and 293 T cells. The virus is characterized to spread rapidly via migratory birds and is likely the major agents for several human infections throughout Europe and China in the past few years [20, 21]. Interestingly, virus propagation was higher in A549 cells than in 293 T cells. The reduced viral replication of virus in 293 T cells provided evidence to further reveal potential host factors and immunological pathways that are involved in the resistance against the infection of influenza virus, especially the H5N1 influenza virus.

Previous studies have already extensively studied the influences of influenza virus proteins on immune escape. However, the exact mechanism underlying the regulation of the balance of host responses to virus infection has not been fully elucidated [22,23,24]. The utilization of genome-wide profiling techniques (e.g., microarray analysis) has identified novel host factors involved in IAV infection [25]. Herein, transcriptome analysis using the RNA-seq technology was applied to identify DEGs of two host cell types during infection. Numerous genes displayed significantly altered expression levels and some common DEGs were revealed in the H5N1 virus-infected human cell lines. These results have discovered and verified a series of cellular targets that may be associated with the cellular responses to the infection of influenza virus.

Previous study suggest that the transcriptional activity of interferon-stimulated genes (ISGs) are regulated by the interferon-mediated immune response, which collectively play a significant role in innate antiviral defense [26]. The IFIT family of genes is a type of ISG, with important antiviral and immune regulation roles. We detected the up-regulation of interferon inducible proteins, including IFIT1, IFIT2, ISG15, and tumor necrosis factor-alpha (TNF-α), which are coordinately expressed in A549 cells in response to H5N1 virus. The expression levels of these ISGs showed no significant differences in 293 T cells infected with H5N1 virus. ISGs can target almost any step in influenza virus life cycle [27, 28]. Some of the most potent antiviral effectors reinforce the system by further stimulation of ISGs, inducing a “cytokine storm” in these cells due to continuous secretion of cytokines [29]. Our results demonstrate that IAV infection can cause strong innate responses in human lung cell lines sensitive to IAV infection and reveal a series of ISGs that possess relationship with IAV infection. The attenuated induction of the interferon-mediated immune mediators may limit H5N1-induced virus propagation in 293 T cells.

Metabolism, as a biological process, contribute to a series of chemical reactions that modify a specific molecule for storage. It has been shown that host cellular metabolic networks are changed by viral infections [30]. Our study demonstrate that metabolic pathways were prominent in the response of A549 cells, suggesting that some genes involved in metabolic pathways, such as AKR1C3 and TXN, play key regulatory roles. The regulatory function of the AKR1C3 gene in influenza virus infection remains to be further studied. A previous report indicated that AKR1C isoenzymes 2 and 3 may be related to the progression of hepatitis C virus-related liver disease in males. AKR1C gene expression has been implicated in the tumorigenesis of other malignancies via the regulation of metabolism [31, 32]. ELOVL3 is involved in metabolic regulation in humans [33]. The present data of pathways in 293 T cells demonstrated the significant enrichment of metabolic processes. Whether these regulations are related to the low replication capacity of the virus in 293 T cells will be our focus in future studies.

A recent clinical study revealed a significant connection between early cytokine responses, immune cell recruitment, and inferior prognoses during H5N1 infection [34]. Extensive transcriptome studies have shown that cytokine storm and uncontrolled inflammatory responses are common features of influenza virus lethality [35]. The present findings are consistent with the previous data. Reducing excessive host response appears to be a rational strategy of treating severe influenza, and drugs aimed to perform this strategy have made some progress in treating influenza. A deeper exploration of the transcriptome data of influenza virus will certainly help us discover more specific therapeutic targets or pathways leading to host overproduction of pathological transcriptional responses to influenza virus and then develop therapeutic drugs and strategies to combat lethal influenza infections.

Conclusion

Our results highlight a reliable approach to investigate the interactions of influenza virus with the host cells. The transcriptome analysis results demonstrated sensitive genes and pathways of two human cell types infected with H5N1 influenza virus. We believe that this data will inform the development of therapeutic drugs and strategies to lessen lethal influenza infections.