TRAF3 deficiency in MDCK cells improved sensitivity to the influenza A virus

Abstract

Tumor necrosis factor receptor-associated factor 3 (TRAF3), an adaptor protein, has significant and varying effects on immunity depending on cell types. The role of TRAF3 in Madin-Darby Canine Kidney Epithelial (MDCK) cell resistance to influenza A virus (IVA) remains elusive. In the present study, CRISPR-Cas9 gene editing technology was used to construct the TRAF3 knockout MDCK cells (MDCK-TRAF3^−/−). Hemagglutination assay, plaque assay, transcriptome, and quantitative real-time PCR were performed after IVA infection. The results showed that after IVA infection, HA titers and virus titers were promoted, interferon I-related pathways were significantly blocked, and transcription of several antiviral-related genes was significantly decreased in MDCK-TRAF3^−/− cells. Thus, our study suggests that TRAF3 gene knockout reduced MDCK cell's resistance to IVA, thereby resulting in a promising way for IVA isolation and vaccine manufacturing.

1. Introduction

Pandemic and seasonal influenza caused by influenza viruses can seriously harm human health [1]. Influenza vaccination has been the most potent approach to prevent infection [2]. Madin-Darby Canine Kidney Epithelial (MDCK) cells from healthy female adult dog kidneys are highly capable of supporting the proliferation of most IVAs and are thus widely used in influenza vaccine production [[3], [4], [5]]. If the sensitivity of MDCK cells to the influenza virus can be further improved, the vaccine yield could be increased, or more influenza virus strains can be isolated. This is of great significance for controlling the spread of influenza virus. It is well known that type I interferon (IFN) has a broad spectrum of anti-viral activity. Therefore, one option is suppressing interferon-related signaling pathways to reduce cellular resistance to influenza viruses. For instance, the knockout of MAVS or IRF7 gene in MDCK cells could promote influenza production [6,7].

Here, we focused on TRAF3. It was identified initially as the intracellular signaling molecule of CD40 and related to cell survival, metabolism, apoptosis, inflammation, and anti-viral immunity [[8], [9], [10], [11], [12], [13]]. As an adaptor protein, TRAF3 is a universal regulator that can directly interact with multiple substances and play a role in different signaling pathways, with a wide range of physiological functions [14]. For example, it was involved in oxidative stress [15], inhibited Noda virus infection [16], had a specific effect on lupus nephritis through modulating the balance of Th17 and Treg cells [11] and was associated with tumors development [17].

Previous studies showed that TRAF3 is involved in various interferon-related signaling pathways, including the NOD-like receptors (NLRs), Toll-like receptors (TLRs), and RIG-I-like receptors (RLRs) signaling pathways [[18], [19], [20]]. It stimulates the production of type I interferons. Knockdown of TRAF3 in A549 cell line reduces the expression of antiviral-related genes, primarily by inhibiting type I IFN signal pathway [21]. Therefore, we expect to reduce type I interferon production by inhibiting multiple type I interferon-related signaling pathways via TRAF3 gene knockout.

However, previous research has shown that TRAF3 functions are highly context-dependent, with receptor and cell-type specificity. In different cell types, TRAF3 may have different or even opposite functions. For instance, TRAF3^−/− DC are known to have a defect in type I IFN production [22], but Xie et al. showed that the type I IFN pathway was elevated in TRAF3^−/− B cells [23]. Previous studies recognized TRAF3 as a negative regulator of TGF-beta-activated kinase 1 (TAK1) in the immune system. The degradation of TRAF3 results in cytoplasmic translocation of the MYD88-associated multiprotein complex and sequential activation of the TAK1 [19,24]. However, Gong et al. found that TRAF3 is a positive regulator of TAK1 in ischemic stroke [25]. Therefore, the antiviral function of TRAF3 in MDCK cells may be different from that in other cells, which needs to be further confirmed.

In the present study, we knocked out the TRAF3 gene in MDCK cells and observed changes in interferon-related signaling pathways to investigate the antiviral function of TRAF3 in MDCK cells. Our results enrich the functional study of TRAF3 and provide theoretical insights for cell modification related to influenza vaccine production.

2. Materials and methods

2.1. Cells and virus

Wild-type (WT) MDCK (CRL-2935™) cells purchased from American Type Culture Collection (ATCC, United States) were cultured in Virus Production Serum Free Medium (VP-SFM) at 37 °C with 5% CO₂. A/Victoria/2570/2019 H1N1 vaccine strain was donated by National Institute for Biological Standards and Control (NIBSC, United Kingdom). Nine-day specific pathogen-free (SPF) eggs were used for virus seed propagation and the virus was stored at −80 °C. After tested virus infection, cells were subject to cultivation in an FBS-free medium that contained TPCK-trypsin (2 μg/mL).

2.2. Construction of Cas9/sgRNA and TRAF3-targeting vectors

The first exon of TRAF3 gene was chosen for the targeting site (http://www.e-crisp.org/E-CRISP/). After efficiency and off-target analysis (Supplementary file 1), an sgRNA sequence (Table 1) was selected, synthesized and constructed into the pX330 vector. The pX330 plasmid, carrying both sgRNA and Cas9 expression unit, was obtained from Addgene [26]. At the same time, sequences of approximately 1 kb in length were selected as 5 ′- and 3' -end homologous arms at positions more than 100 bp away from the sgRNA sequences. Both homologous arms were obtained by PCR using wild-type MDCK genome DNA as a template. The sgRNA and the homologous arm sequences are shown in Table 1. Subsequently, the TRAF3-targeting vector was constructed by inserting the 5′ and 3′ homologous arms into pY75 plasmid. The pY75 plasmid is a derivative of pUC19 [27], which was constructed and generously provided by Dr. Daiguan Yu.

Table 1.

Primers for construction of sgRNA and homologous arms.

Primer name	Primer sequence (5′–3′)
TRAF3-sgRNA-Forward	CACCGGAGACCCCAGGAGATATAGG
TRAR3-sgRNA-Reverse	AAACCCTATATCTCCTGGGGTCTC
5′-TYB-Forward	CGGAATTCCTGGCTATTTCCCACTGTGG
5′-TYB-Reverse	GGACTTCTGCTGTCAGTAGAAGATATCCC
3′-TYB-Forward	ACGCGTCGACTGGTCTCAGGAGTTCAATGG
3′-TYB-Reverse	CCCAAGCTTACAAAACAAAGATTGGAGGG

2.3. Electroporation of vectors into MDCK cells and screening of positive clones with puromycin

The electrotransfection method was used in the study. MDCK cells were digested and counted prior to the electroporation. Then 10⁶ MDCK cells and 8 μg vectors (4 μg pX330 and 4 μg pY75) were added to each cuvette, and electrotransfection was performed at 110v voltage, 1 pulse, and 10 ms per pulse using Gene Pulser Xcell Electroporation System (Bio-Rad, United States) (Supplementary file 2). After electrotransfection, the cells were spread into six-well plates and cultured following standard procedure. The VP medium containing puromycin (8 μg/mL) was added to replace the original one when the cells nearly reached 80% confluently and then changed at 2–3-day intervals. At 7–10 days following the selection with puromycin, limited dilutions of puromycin-resistant cells were performed to obtain monoclonal cells.

2.4. Genotyping with PCR and sequencing

After obtaining monoclonal cells, genomic DNA was extracted for identification, and three paired primers were designed to screen positive monoclonal cells (Table 2). The first set of primers designated as full-length PCR primers were targeted upstream and downstream of the 5′ and 3′ homologous arms, respectively. This could simultaneously identify wild-type, heterozygous and homozygous cell clones. The other two sets of primers designated as located PCR primers were targeted upstream of the 5′ homologous arm sequence and puromycin sequence or downstream of the puromycin sequence and 3′ homologous arm sequence, respectively, to further confirm the homozygous clone cells and exclude the possibility of the interference of plasmid residues in cells. The final PCR amplification products were sent for DNA sequencing.

Table 2.

Primers used in PCR.

Name of primer	Sequence of primer (5′–3′)
F1	ACAGTGTCTGCTGTTGTTGG
R1	AGAGCACCTGCACTTTAAGG
F2	ACAAGGCTCTAGAGAGTGGG
R2	GAAGAACTCCAGCATGAGAGG
F3	GCCAATAGCAGCTTTGCTCC
R3	GCTGGACATGTGTATCCTGGG

2.5. Western-blot (WB) assay

After washing with pre-chilled sterile PBS, cells were subject to lysis through repeated freezing and thawing. After centrifugation at 12,000 rpm for 5 min, a BCA protein assay kit (Beyotime Ins. Biotech, China) was used to measure the protein content. Subsequently, the proteins were separated using 12% SDS-PAGE, followed by transfer onto nitrocellulose filter membrane (NC) membranes. Next, 5% bovine serum albumin (BSA) was added for a 2 h period to block membranes, followed by overnight primary antibody incubation (anti-TRAF3; Abcam; 1:1000，United Kingdom) under 4 °C. After rinsing with PBST, a secondary antibody (Goat Anti-Rabbit IgG H&L (HRP), Abcam; 1:50000，United Kingdom) was added and incubated for a 1 h under ambient temperature. Proteins of the membrane were visualized using enhanced chemiluminescence (ECL) reagent (Vazyme Biotech, China); autoradiograms were scanned and analyzed using Amersham ImageQuant 800 system (Cytiva，Japan). Calnexin was used as a loading protein.

2.6. RNA extraction and quantitative real-time PCR (qRT-PCR)

TaKaRa MiniBEST Universal RNA Extraction Kit (Takara, Shiga, Japan) was used to extract the total cellular RNA. The cDNA was obtained using the PrimeScript™ IV 1st strand cDNA Synthesis Mix kit (Takara, Shiga, Japan) with 2 μg RNA as the template. Premix Ex Taq™ (Probe qPCR) (Takara, Shiga, Japan) was used for qRT-PCR with 100 ng cDNA as the template. The 7500 Fast Real-Time PCR System (ABI Life Technologies, Singapore) was used for recording. The PCR procedure was described in the operating instructions of Premix Ex Taq™ (Probe qPCR), and the primer sequences are listed in Supplementary file 3. Finally, the 2^–ΔΔCT approach [28] was adopted to examine gene expression levels, and 18S rRNA was used for normalization.

2.7. Detection of influenza virus hemagglutination titer

The MDCK-TRAF3^−/− and wild-type MDCK cells were infected with H1N1 influenza virus at MOI 0.01 (multiplicity of infection). After 48 h at 34 °C, the supernatant in the six-well plate was collected for viral hemagglutination titer detection. PBS (25 μL) was added to the wells of columns 2 to 12 of the U-shaped 96-well plate, followed by 50 μL of cell supernatant in the first column. Next, 25 μL of supernatant from column 1 was taken into the corresponding wells of column 2, thus, diluted two-fold. All wells were introduced with 25 μL PBS containing 1% chicken RBCs and observed in 30 min. The erythrocyte agglutination titer was determined by the highest dilution showing complete agglutination, and the reciprocal of this dilution was the erythrocyte agglutination titer of the virus.

2.8. Plaque assay

H1N1 influenza virus was infected into MDCK-traf3^−/− and wild-type MDCK cells at MOI 0.01. The culture supernatants were harvested at 24 h, 48 h, 72 h and 96 h after infection and used for plaque assay to determine the virus titers. The supernatant was diluted 10 times gradually before being added to a 6-well culture plate containing MDCK cells. One hour later, the supernatant was discarded and then the solution containing 1 μg/mL TPCK-trypsin (Sigma-Aldrich, St. Louis, MO, USA) and 1% penicillin/streptomycin liquid (Gibco, Grand Island, NY, USA) was prepared by mixing 2 × DMEM medium into an equivalent microcrystalline cellulose solution. After culturing for an additional 96 h, the live cells were stained with 1 mL neutral red staining solution (Beyotime, China) to observe plaque formation.

2.9. RNA sequencing of the transcriptional profile of MDCK cells

Harvested cells were wild-type MDCK, H1N1 influenza virus-infected wild-type MDCK (48 h), MDCK-TRAF3^−/−, H1N1 influenza virus-infected MDCK-TRAF3^−/− (48 h). Each group's samples were named WT-uninfected, WT-infected, TRAF3-ko-uninfected, and TRAF3-ko-infected, respectively. The total cellular RNA was sent to Wuhan Biobank (China) for RNA sequencing. The main process was as follows: cDNA synthesis, library preparation, amplification and quality control, sequencing, and data analysis. Subsequently, clustering, Gene Ontology (GO) functional annotation, Kyoto Encyclopedia of Genes and Genomes (KEGG) pathway enrichment was performed (Supplementary file 4).

2.10. Statistical analysis

Results were presented as mean ± SEM and analyzed using GraphPad Prism 9.0. The unpaired t-test was utilized to compare the two groups. Statistical significances were marked as: *p < 0.05, **p < 0.01 and ***p < 0.001, respectively.

3. Results

3.1. MDCK-TRAF3^−/− cells were identified by PCR, DNA sequencing, qRT-PCR, and western blot

PCR identification primers of MDCK-TRAF3^−/− were marked red in Fig. 1A. Full-length PCR (the primers were F1 and R1) results showed only one band at the knock-out cells which was about 1200 bp longer than that at the wild type cells (Fig. 1B), suggesting that the puromycin resistance gene sequence was inserted between the 5′ and 3′ homologous arm and indicating the cells were a biallelic knockout. However, this could be caused by the residue of the pY75 plasmid in the cells. Therefore, another set of identification PCR was included, with one primer designed on the puro sequence and the other on the lateral side of the homologous arms (primer F2 and R2, or primer F3 and R3). The results showed that only the MDCK-TRAF3^−/− cells showed bands (Fig. 1C). These results above indicated that homozygous cells with TRAF3 gene knockout were obtained.

Fig. 1.

PCR, qRT-PCR, and western bolt to verify MDCK-TRAF3^−/− monoclonal cells. (A) Part of the TRAF3 exon1 sequence (158 bp) in MDCK-TRAF3^−/− was replaced with the puromycin resistance gene sequence (Pruo). PCR identification primers of MDCK-TRAF3^−/− were marked red. Primer F1 and R1 were targeted the 5′ and 3′ homologous arms, respectively. Primer F2 and R2 were targeted upstream of the 5 ′homologous arm sequence and puromycin sequence, respectively. Primer F3 and R3 were targeted downstream of 3′ homologous arm sequence and the puromycin sequence, respectively. (B) Gel electrophoresis was conducted to separate the full-length PCR products with primers F1 and R1. Lanes 1 and 2 indicate wild-type and MDCK-TRAF3^−/− full-length PCR products. (C) Gel electrophoresis was conducted to separate the targeting site PCR products. Lane 1 indicates the DNA ladder，Lane 2 and 3 indicate located PCR products using F2 and R2 as primers. Lanes 4 and 5 indicate located PCR products using F3 and R3 as primers. (D) qRT-PCR analysis on TRAF3 mRNA expression in wild-type MDCK and MDCK-TRAF3^−/−. (E) Western blot of TRAF3 protein in wild-type MDCK and MDCK-TRAF3^−/−.

Subsequently, DNA sequencing of the full-length PCR products showed that the TRAF3 gene had a 158 bp knockout in exon 1, and 1383 bp were inserted (including PGK promoter, Puro ORF, and SV40 poly A).

Further, the mRNA expression of TRAF3 was detected by qRT-PCR, and the expression level of TRAF3 protein was detected by Western blot. The qRT-PCR results (Fig. 1D) showed that the TRAF3 mRNA was markedly decreased (F = 0.0828, p < 0.0001). Similarly, Western blot results (Fig. 1E) showed that TRAF3 protein was significantly reduced in MDCK-TRAF3^−/− than wild type MDCK with Calnexin as a reference protein. Protein expression fell by 97.1%. Therefore, MDCK-TRAF3^−/− cells are suitable for subsequent functional studies of TRAF3.

3.2. TRAF3 knockout can upregulate NF-κB pathway in MDCK cells

NF-kB proteins can regulate the expression of hundreds of genes, which regulate important physiological processes such as inflammation, immunity, proliferation, and cell death [29]. Earlier studies showed that TRAF3 negatively regulated the NF-κB pathway in immune cells [19], which was verified in MDCK cells in the present study. To check the changes of NF-κB pathway in MDCK cells after TRAF3 knockout, the RNA transcriptomics of MDCK-TRAF3^−/− and WT MDCK cells were sequenced, respectively. A total of 4276 genes were found to be significantly different between the two groups, including 2176 downregulated and 2700 upregulated genes (Fig. 2A). According to the transcriptomic data (Fig. 2B), after TRAF3 knockout, multiple gene expression in the NF-κB signaling pathway was upregulated, including a series of critical genes such as nuclear factor kappa B subunit 2 (NFKB2) and RELB. Meanwhile, the cell senescence signaling pathway was found to be significantly inhibited, and the expression of several differential genes was downregulated in MDCK-TRAF3^−/− cells (Fig. 2C). These results suggest that MDCK-TRAF3^−/− cells may be altered in terms of proliferation and cell death, which may affect the production of influenza viruses in MDCK and need to be further investigated.

Fig. 2.

Differentially expressed genes in MDCK-TRAF3^−/− cells compared with WT counterparts. (A) MDCK-TRAF3^−/− vs. WT MDCK volcanic plot, with |log2FoldChange(FC)| > 1 and p < 0.05 being the criteria. Red, blue, and black dots stand for genes with upregulation, downregulation, and non-differentiation, respectively. (B) Heatmap showing DEGs related to NF-κB pathway. (C) Heatmap showing DEGs related to cell senescence signaling pathway.

3.3. The production of influenza A virus was increased in MDCK-TRAF3^−/− cells

H1N1 influenza virus with different MOI was infected in MDCK-TRAF3^−/− and WT counterparts to test whether MDCK cells would be more sensitive to influenza virus after TRAF3 knockout. The cell supernatant was collected at different intervals for hemagglutination assay and plaque assay. The results showed that the hemagglutination titer of influenza in MDCK-TRAF3^−/− cells increased compared with WT MDCK cells at different MOI and times (Fig. 3A). The highest hemagglutination titer was eight times that of the control group after infection for 96 h at 0.001 MOI. Meanwhile, the results of plaque assay showed that the number of plaque formation in the MDCK-TRAF3^−/− cells group were significantly higher than those in the control group after infection for 48 h, 72 h and 96 h (Fig. 3B). The virus titer of the control group peaked after 48 h of infection, while that of the MDCK-TRAF3^−/− group peaked after 72 h. These results suggested that MDCK-TRAF3^−/− cells were more sensitive to the influenza virus and could produce higher titer of the virus.

Fig. 3.

The results of hemagglutination test and plaque assay. (A) The hemagglutination titer of influenza in MDCK-TRAF3^−/− and WT counterparts. MDCK-TRAF3^−/− cells and WT counterparts were infected with H1N1 with MOI = 0.1, 0.01, and 0.001, respectively. Then influenza virus erythrocyte agglutination test was performed at 48 h, 72 h, and 96 h after infection. (B) The results of plaque assay. The MDCK-TRAF3^−/− and wild-type MDCK cells were infected with H1N1 influenza virus at MOI 0.01. The cultural supernatants were harvested at 24 h, 48 h, 72 h and 96 h postinfection and then applied for plaque formation assay.

3.4. Interferon-related signaling pathways were inhibited in MDCK-TRAF3^−/− cells

Transcriptome sequencing of MDCK-TRAF3^−/− cells and WT counterparts 48 h after H1N1 infection was performed to understand changes in the interferon-related signaling pathways. After software analysis, screening was performed with the criteria of p-value <0.05 and |log2(FoldChange)| > 1, and a total of 8208 differentially expressed genes with significant relationships were obtained, including 3356 downregulated and 4582 upregulated genes (Fig. 4A). In KEGG enrichment analysis, the most differentially expressed genes (DEGs) were in the anti-viral-related signaling pathway (Fig. 4B). Changes in NOD-like receptor (Fig. 4C), Toll-like receptor (Fig. 4D), RIG-I-like receptor (Fig. 4E) and JAK-STAT signal pathways (Fig. 4F), which were associated with type I interferon production and function was observed to understand the effects of TRAF3 knockout on interferon-related pathways. Transcriptomic data showed that many DEGs (such as DDX58, TLR3, IFIH1, ATG5 and TBK1) related to the above pathways were downregulated, indicating that pathways related to type I interferon production and function were inhibited. The expressions of TKR3, IFNB1, JAK1, JAK2 and various anti-viral related genes such as ISG15, EIF2AK2, OAS1, OAS2, MX2 and RNASEL decreased significantly. The immunity of MDCK-TRAF3^−/− cells to the influenza virus decreased significantly. This was echoed by the downregulation of most differential genes in the signaling pathways associated with influenza A virus infection (Fig. 4G). Thus, influenza viruses proliferated better in MDCK-TRAF3^−/− cells.

Fig. 4.

Differentially expressed genes between MDCK-TRAF3^−/− and WT counterparts following infection with influenza virus 48 h. (A) MDCK-TRAF3^−/− vs. Wild-type MDCK volcanic plot, with | log2FC| > 1 and p < 0.05 being the criteria. Red, blue and black dots stand for genes with upregulation, downregulation, and non-differentiation, respectively. (B) MDCK-TRAF3^−/− vs. Wild-type MDCK KEGG enrichment dot plot. (C) Heatmap showing DEGs related to NOD-like receptor pathway. (D) Heatmap of DEGs related to Toll-like receptor pathway. (E) Heatmap showing DEGs related to RIG-I-like receptor pathway. (F) Heatmap showing DEGs related to JAK-STAT pathway. (G) Heatmap showing DEGs related to pathways associated with influenza A virus infection.

3.5. The differentially expressed genes were further validated by qRT-PCR

To confirm the accuracy of the transcriptomic results, qRT-PCR was performed on type I interferon-related genes. According to the production and action process of interferon, we selected the following genes: genes involved in the production of interferon, such as DDX58, TLR3, IFIH1, TBK1, ATG5 and IFNB1; genes involved in the activation of ISGs (interferon-stimulated genes) by interferon, such as JAK1, JAK2 and STAT2; and genes involved in viral resistance, such as ISG15, EIF2AK2, OAS1, OAS2, MX2 and RNASEL. The results revealed that the expression of these genes in MDCK-TRAF3^−/− were reduced to different degrees compared to the control (Fig. 5), which is consistent with the results of transcriptome. Therefore, the Interferon-related signaling pathway may be inhibited, and the anti-viral ability of MDCK cells may be reduced following TRAF3 gene knockout.

Fig. 5.

Transcription level of some differentially expressed genes verified by qRT-PCR. (A)Transcription level of ISG15. (B) Transcription level of EIF2AK2. (C) Transcription level of OAS1. (D) Transcription level of OAS2. (E) Transcription level of MX2. (F) Transcription level of RNASEL. (G) Transcription level of DDX58. (H) Transcription level of TLR3. (I) Transcription level of IFIH1. (J) Transcription level of IFNB1. (K) Transcription level of ATG5. (L) Transcription level of TBK1. (M) Transcription level of JAK1. (N) Transcription level of JAK2. (O) Transcription level of STAT2. (P) Transcription level of IFNGR2. The MDCK- TRAF3^−/− cells and WT counterparts were infected with H1N1 at MOI = 0.01. Total RNA was extracted at 48 h post-infection, and then cDNA was synthesized by reverse transcription, followed by qRT-PCR.

4. Discussion

In the present study, TRAF3 knockout in MDCK cells improved their sensitivity to the influenza virus. Previous studies have shown that TRAF3 promoted type I interferon production through RLRs and TLRs [30,31]. By knocking out the TRAF3 gene in both dendritic cells (DCs) and macrophages, type I interferon production was significantly reduced [22,32]. A similar phenomenon was observed in MDCK cells in the present study. Transcriptomic data from MDCK with TRAF3 gene knockout revealed that type I interferon, and multiple interferon-related signaling pathways were significantly inhibited compared to WT counterparts after being infected with influenza virus H1N1. Several vital genes in interferon-related signaling pathways were downregulated, such as TLR3 and TBK1. As a receptor mainly in the endosomes, TLR3 can be activated by recognizing a variety of dsRNA molecules [[33], [34], [35]]. Subsequently, TLR3 can recruit TIR domain-containing adapter-inducing interferon-β (TRIF) to trigger the downstream signal transduction cascade and eventually promote interferon production [36]. Decreased TLR3 expression in MDCK-TRAF3^−/− cells indicated that TLRs-mediated type I interferon production might have been inhibited. The RLR family mainly includes melanoma differentiation-associated gene 5 (MDA5) and RIG-1 that can identify short and long double-stranded RNA, respectively [[37], [38], [39]]. Activation of RLRs can activate the mitochondrial antiviral signaling pathway (MAVS), which causes the phosphorylation of TBK1 and triggers downstream anti-viral innate immune signaling pathways, including type I interferon production [39,40]. The decreased TBK1 expression in MDCK-TRAF3^−/− cells declared that RLRs-mediated type I interferon production was also inhibited. The above results indicated that TRAF3 knockout in MDCK cells inhibited both TLRs and RLRS-mediated type I interferon production. Interestingly, we also found decreased expression of JAK1, JAK2, STAT2, and STAT6 genes, suggesting that the JAK-STAT signaling pathway, in which type Ι interferon induces upregulation of various anti-viral proteins, may also be inhibited by TRAF3 knockout in MDCK cells. This is not entirely unexpected, as several studies have demonstrated that TRAF3 has different functions in various cells in relation to the JAK/STAT pathway [41]. In anaplastic large cell lymphoma (ALCL) cells, TRAF3 is a positive regulator of the JAK/STAT pathways through upregulation of phosphatase and tensin homolog protein [42]. In contrast, in regulatory T (Treg) cells, TRAF3 dampened interleukin-2 (IL-2) signaling by facilitating recruitment of T cell protein tyrosine phosphatase (TCPTP) to the IL-2 receptor complex, resulting in dephosphorylation of the signaling molecules JAK1 and JAK3 and negative regulation of JAK-STAT5 signaling [43].

Interferons are known to activate ISGs and produce various anti-viral proteins that inhibit multiple steps of viral replication [44]. After infection with H1N1, transcriptome results also suggested that the expression of various anti-viral genes (such as ISG15, EIF2AK2, OAS1 and OAS2) decreased in MDCK-TRAF3^−/− cells, which reduced the resistance of MDCK cells to influenza virus. This phenomenon was further confirmed by increased hemagglutination titers in MDCK-TRAF3^−/− cells.

Simultaneously, gene expression in the non-classical NF-κB pathway was increased in MDCK-TRAF3^−/− cells. NF-κB, as a crucial nuclear transcription factor, was expressed in almost all types of cytoplasm. Its functions are extensive and closely related to cell survival, mutation, and proliferation [45]. Several disorders, such as malignant tumors, and inflammatory and autoimmune diseases, have been suggested to be associated with an imbalance in the regulation of NF-κB [46,47]. The upregulation of NF-κB after TRAF3 knockdown suggested that MDCK-TRAF3^−/− cells may have alterations in proliferation, differentiation, metabolism, and tumorigenicity changes, the details of which need to be further investigated.

The study contributed to our understanding of the role of TRAF3 in MDCK cells. The results showed that TRAF3 knockout could simultaneously inhibit several interferon I-related pathways and downregulate the expression of various anti-viral related genes. Thus, MDCK-TRAF3^−/− cells can be used to obtain a higher yield of influenza virus to produce influenza vaccines and the isolation of strains.

Author contribution statement

Yang Le: Performed the experiments; Wrote the paper.

Jiayou Zhang: Conceived and designed the experiments; Analyzed and interpreted the data; Contributed reagents, materials, analysis tools or data.

Zheng Gong; Xuanxuan Nian; Xuedan Li; Daiguan Yu; Ning Ma; Rong Zhou; Guomei Zhang; Bo Liu; Lu Yang: Analyzed and interpreted the data.

Zhegang Zhang: Conceived and designed the experiments; Analyzed and interpreted the data.

Baiqi Fu; Xiuqin Xu: Contributed reagents, materials, analysis tools or data.

Xiaoming Yang: Conceived and designed the experiments; Contributed reagents, materials, analysis tools or data.

Data availability statement

Data will be made available on request.

Declaration of competing interest

The authors declare that they have no known competing financial interests or personal relationships that could have appeared to influence the work reported in this paper.

Appendix A. Supplementary data

The following are the Supplementary data to this article: