Single-cell profiling of African swine fever virus disease in the pig spleen reveals viral and host dynamics

Significance

African swine fever is arguably the most severe infectious disease threating the global pig industry. However, how African swine fever virus (ASFV) manipulates the infected cell population and controls host responses to drive high-efficient replication of the virus in vivo remains unclear. Our work suggested that the shift of ASFV infection from macrophages to an unusual subpopulation of immature monocytes and insufficient antiviral responses drive prolonged infection in vivo. Until now, the role of immature monocytes as an important target by ASFV has been overlooked. We identified molecular determinants of tropism and examined temporal dynamics in viral and host gene expression in these cells, which will help in rationale design of antivirals or vaccines and clarify ASFV pathogenesis.

Abstract

African swine fever, one of the major viral diseases of swine, poses an imminent threat to the global pig industry. The high-efficient replication of the causative agent African swine fever virus (ASFV) in various organs in pigs greatly contributes to the disease. However, how ASFV manipulates the cell population to drive high-efficient replication of the virus in vivo remains unclear. Here, we found that the spleen reveals the most severe pathological manifestation with the highest viral loads among various organs in pigs during ASFV infection. By using single-cell-RNA-sequencing technology and multiple methods, we determined that macrophages and monocytes are the major cell types infected by ASFV in the spleen, showing high viral-load heterogeneity. A rare subpopulation of immature monocytes represents the major population infected at late infection stage. ASFV causes massive death of macrophages, but shifts its infection into these monocytes which significantly arise after the infection. The apoptosis, interferon response, and antigen-presentation capacity are inhibited in these monocytes which benefits prolonged infection of ASFV in vivo. Until now, the role of immature monocytes as an important target by ASFV has been overlooked due to that they do not express classical monocyte marker CD14. The present study indicates that the shift of viral infection from macrophages to the immature monocytes is critical for maintaining prolonged ASFV infection in vivo. This study sheds light on ASFV tropism, replication, and infection dynamics, and elicited immune response, which may instruct future research on antiviral strategies.

African swine fever virus (ASFV) is the causative agent of African swine fever (ASF) which has caused substantial economic losses to the swine industry worldwide (1–3). ASFV infection causes acute fatal hemorrhagic disease in swine with high mortality rates up to 100% (4). Therefore, ASF is arguably the most severe emerging infectious disease threat for the global pig industry and listed as a “notifiable disease” by the World Organisation for Animal Health (5). No safe and effective vaccines or drugs are available for prevention or treatment of ASF until now. The mechanisms of viral entry, replication, assembly, and release as well as in manipulating host cell signaling and regulation of host cellular responses remain elusive that should be further investigated (6).

The clinical signs caused by ASFV infection in pigs include high fever and a variety of lesions in multiple lymphoid organs (4). However, the cellular and molecular understanding of how ASFV drives lymphoid organ pathology and function failure is limited. Although many cellular responses have been identified during ASFV infection, including inflammatory responses as well as virus-induced antagonistic responses against the host immune system (5, 7), most of these studies were performed ex vivo in macrophages (8). Multiple important questions related to the complexity of the in vivo ASFV infection remain to be answered. Especially, how ASFV manipulates the infected cell population and controls host responses to drive high-efficient replication of the virus and prolonged infection in vivo is still unclear.

High-throughput single-cell RNA-sequencing techniques (scRNA-seq) have been widely used to obtain comprehensive visualization of cellular responses and transcriptional profiles in various kinds of cells or animal tissues in recent years (9). In particular, the 10× Genomics scRNA-seq technology has further deepened our understanding of the cell compositions within tissues, single-cell transcriptional landscape, cell-to-cell communication, and cellular states (10–12). Meanwhile, by quantifying viral RNA within cells, scRNA-seq allows comparison of the gene expression pattern between infected and bystander cells in a diseased host, depicting a high-resolution transcriptomic landscape of the host during viral infection.

To offer a global and in-depth characterization of the pathogenic host responses throughout ASFV infection in vivo, we evaluated the virus dissemination by detailed necropsy studies of ASFV-infected pigs and performed a comprehensive scRNA-seq investigation of the lymphoid organ spleens from ASFV-infected and uninfected pigs in the present study. We characterized the transcriptomes for 49,924 single cells from 12 spleens in ASFV-infected or mock-infected pigs, and explored the changes in cell-type abundance throughout ASFV infection. By using scRNA-seq technology with multiple methods, we determined that monocytes and macrophages comprise the main infected cell population in the spleen. ASFV causes massive death of macrophages but shifts its prolonged infection into a rare subpopulation of immature monocytes during the infection. The immature monocytes with reduced interferon (IFN) signaling and antigen-presentation capacity which are more sensitive to ASFV infection than conventional monocyte subsets significantly arise after the infection. The emerged immature monocytes are critical for maintaining the prolonged infection of ASFV in vivo. Moreover, we detected viral RNA within individual cells, allowing us to investigate ASFV tropism in vivo and identify ASFV-associated transcriptional changes in putative pro-viral or antiviral genes. These data demonstrate that the overall transcriptional induction to ASFV is aberrant. Our study depicts a high-resolution transcriptomic landscape of spleen immune cells during disease progression of ASF and highlights the power of simultaneous single-cell measurements of both hosts and viral transcriptomes in delineating a comprehensive map of in vivo infection, which will facilitate a better understanding of the protective and pathogenic immune responses of ASF.

Results

Rapid Virus Dissemination to Multiple Organs in Pigs After ASFV Infection.

A preliminary titration study was performed to establish a proper clinical course but not overwhelming challenge dose, and the dose of 10 HAD₅₀ that led to 100% mortality in 7 to 10 d was selected in the present study (SI Appendix, Fig. S1A). A series of necropsy studies were performed in 18 pigs on day 0, 1, 3, 5, 7, or 9 after ASFV infection (4, 13) (SI Appendix, Fig. S1B). The viral DNA could be detected in the serum and rectal swab samples from one of the challenged pigs but not in its nose and mouth samples at 3 dpc (SI Appendix, Fig. S1C). The viremia occurred in the majority of pigs (7/9) on day 5 following ASFV infection, meanwhile, the virus could be detected from the mouth (7/9), nose (4/9), and rectal swab (6/9) at the same time (SI Appendix, Fig. S1C). The temperature began to rise at 3 dpc and reached the highest temperature at 7 dpc (SI Appendix, Fig. S1D). The onset of clinical signs occurred at 3 dpc including pyrexia and obtundation. The overall highest mean clinical score was observed at 9 dpc (14) (SI Appendix, Fig. S1E). No significant increase of p30 antibody was observed as ASFV infection progressed (SI Appendix, Fig. S1F).

Viral DNA levels in tissues were quantitated using a sensitive quantitative PCR (qPCR) assay. The highest levels of viral DNA were detected in the spleen (SI Appendix, Fig. S1G), and it also revealed the most severe pathological manifestation (SI Appendix, Fig. S1H). Splenic red pulp was enlarged and hyperemic, with lymphocyte infiltration (SI Appendix, Fig. S1I). In pigs necropsied on days 0, 1, and 3, no detectable levels of viral DNA in the collected tissues were observed. However, high levels of viral DNA were observed in all of the collected tissues of the animals necropsied on days 5, 7, and 9 (SI Appendix, Fig. S2A). A progressively severe lesions occurred in animals, which were characterized by extensive hemorrhaging and multiple immune organs’ enlargement (SI Appendix, Fig. S2B). Pathological changes occurred in almost all major organs of the challenged pigs, including extensive necrosis, degeneration, and inflammatory cell infiltration (SI Appendix, Fig. S3).

Dissecting In Vivo ASFV Infection Using Single-Cell Mapping of Host Transcriptome in the Spleen.

Given that the principal gross lesions were observed in the spleens that contained the highest levels of viral loads, we used the scRNA-seq to comprehensively characterize the cellular dynamics in the spleens of the challenged animals as ASFV infection progressed (Fig. 1A). We sorted live spleen (red pulp) cells from pigs on 0, 3, 5 and 7 d after infection with ASFV and performed single-cell gene-expression profiling. A total of 49,924 cells from the spleens of ASFV-infected or mock-infected pigs euthanized according to the approved procedure by the Gansu Animal Experiments Inspectorate and the Gansu Ethical Review Committee (SYXK [GAN] 2014-003) were analyzed, ~107.2 million unique transcripts were obtained from all of the isolated cells. Among these cells, 12,064 cells (24.2%) were from the pigs euthanized at 3 dpc, 9,131 cells (18.3%) were from the pigs euthanized at 5 dpc, 9,264 cells (18.5%) were from the pigs euthanized at 7 dpc, and 19,465 cells (39.0%) were from the mock-infected pigs (at 0 dpc). All high-quality cells were incorporated into an unbatched and comparable dataset and followed by principal component analysis (Fig. 1B) after correction for read depth and mitochondrial read counts (Dataset S1). The samples were well distributed across cell-type clusters (Fig. 1C) but separated by DPC (Fig. 1D), suggesting dynamic cell states throughout ASFV infection.

Fig. 1.

Study design and differences in cell composition in the spleen throughout ASFV infection. (A) Schematic illustration of the experimental workflow. The scRNA-seq and bulk RNA-seq was applied to obtain comprehensive visualization of transcriptional profiles in the spleen throughout ASFV infection and the output data were used for integrative analyses and verified by multiple experiments. (B) Overview of the cell clusters in the integrated single-cell transcriptomes. Clusters were labeled by FindClusters with resolution 0.8. Each dot represents a single cell. (C) Cellular populations identified. Each dot corresponds to a single cell, colored according to cell type. The Right panel of violin plots represents the expression distribution of selected canonical cell markers in the seven clusters. The columns represent selected marker genes and the rows represent clusters with the same color as in the Left panel. (D) UMAP projection of the samples from day 0, 3, 5, or 7. Each dot represents a single cell, colored according to cell type. (E) Average proportion of each cell type derived from the samples collected on days 0, 3, 5, and 7, respectively. (F) The cell compositions at a single sample level. (G) Condition preference of each cell type. y axis, average percentage of samples across four conditions (days 0, 3, 5, and 7). The time points are shown in different colors. Each bar plot represents one cell type. Error bars represent ±SD for the samples collected at the indicated time point. (H and I) The top 20 viral genes in infected cells, ranked by the expressed frequency on day 5 (H) and day 7 (I).

Using graph-based clustering of uniform manifold approximation and projection (UMAP), the transcriptomes of seven major cell types were captured in accordance with the expression of canonical gene markers (Fig. 1C and SI Appendix, Fig. S4A). These cells included: B cells (14,183 cells), T cells (21,819 cells), NK cells (1,598 cells), macrophages (5,095 cells), monocytes (3,407 cells), DC (1,529 cells), and neutrophils (1,953 cells). Based on this analysis, we comprehensibly defined the composition of cell populations in the spleen throughout ASFV infection.

Dynamics of Cell Type Compositions in the Spleen as ASFV Infection Progresses.

Changes in cell-type abundance in the spleen could reflect cell proliferation, death, and movement into and out of the spleen. To investigate the differences in cell composition at different time points as ASFV infection progresses, we evaluated the relative proportion of the seven major cell types in the spleen of each animal according to the scRNA-seq data (Fig. 1 E–G). The relative proportion of T cell cluster initially increased on 3 dpc before declining gradually between 3 and 7 dpc. The relative abundance of B cells and DCs stayed relatively constant during the viral infection. Of note, the relative percentage of NK cells and macrophages precipitously decreased >1.5-fold, and monocytes increased >fivefold between 3 and 7 dpc with disease severity, suggesting that the contributors to cellular variability were the progression of infection. The data have also been presented as cell numbers, and the tendency for the changes of cell numbers was similar with the changes of proportions as well (SI Appendix, Fig. S4B). This confirmed the dynamics of cell types compositions in the spleen during ASFV infection. The dynamics of monocytes/macrophages, CD21⁺ B cells, and T cells were verified by fluorescence-activated cell sorting (FACS) method, the changes in abundance of monocytes/macrophages, B cells and T cells were identical with the results of scRNA-seq data (SI Appendix, Fig. S4C).

The temporal expression patterns of 159 previously annotated ASFV genes were analyzed as well. The top-ranked viral genes based on expressed frequency in cells with viral fragments were explored. The CP204L, MGF_100-1L, MGF_110-5L-6L, I73R, and DP238L genes were the top-ranked viral genes at 5 and 7 dpc during infection (Fig. 1 H and I). The high expression of MGF_100-1L, MGF_110-5L-6L, MGF_110-3L, MGF_110-4L, and I73R genes was in accordance with a recent study demonstrating that these viral genes are the top-ranked viral genes in ASFV-infected PAMs (8). These genes may contribute to the pathogenicity of ASFV and serve as important diagnostic markers in vivo.

Monocytes and Macrophages Are the Major Cell Types that Infected by ASFV In Vivo.

The scRNA-seq analysis indicated that all the cells harbored no detectable viral loads on 0 and 3 dpc, which is consistent with our previous observation that no viral DNA was detected in the spleen tissues from pigs necropsied on days 0 and 3 by qPCR (SI Appendix, Fig. S1G and Dataset S1). The viral genes were detectable at 5 and 7 dpc, and the relative proportion of ASFV-positive cells and viral loads increased from 5 to 7 dpc confirming that the viral loads increased as infection progressed (SI Appendix, Fig. S4 D–F). We then divided the group of cells with viral mRNA into three categories of viral-load states based on the unique molecular identifiers (UMIs): low (<0.5%), medium (between 0.5% and 5%), and high (>5%) (15). A clear distinction was found in monocytes and macrophages compared to the rest of the cell types. We found that monocytes manifest the widest range of viral load states, with 32.5%, 3.6%, and 63.9% of viral fragment containing monocytes characterized respectively by low, medium, and high viral-load states (Fig. 2A and SI Appendix, Fig. S4G). Macrophages ranked second, with 72%, 12%, and 16% of viral fragment containing macrophages characterized respectively by low, medium, and high viral-load states. The rest of the cell types mainly displayed low heterogeneity of viral-load states, suggesting that viral-load heterogeneity is higher in monocytes and macrophages.

Fig. 2.

Characterization of the generic host response to ASFV infection. (A) Single-cell heterogeneity of intracellular viral loads within the spleens of ASFV-infected pigs. Shown are the percentages of low (light blue), medium (steel blue), and high (dark blue) viral-load states (y axis) within the population of cells containing viral fragments, as identified for each of the seven cell types (x axis). (B) Monocytes and macrophages comprise the main infected cell population in the spleen. (C) The density plot of viral UMIs in the infected macrophages and monocytes on day 5 and day 7. (D) Comparison of log₂fold-change (log₂FC) of 10,071 intersected genes between scRNA-seq and bulk RNA-seq for 5 dpc vs. 0 dpc. (E) Comparison of log₂FC of 10 selected genes determined by scRNA-seq and qPCR analysis in monocytes/macrophages. (F and G) Expression of IFNs, IFN receptors and ISGs, and infection status in macrophages and monocytes during ASFV infection respectively. (H) Boxplot of ISG score in macrophages/monocytes at baseline, uninfected bystanders or infected cells during the infection (5 and 7 dpc). (I) Association between ISG score and the number of viral UMI. (J) Heatmap of IFNs, IFN receptors, and ISGs between 0 and 5 dpc in bulk RNA-seq data. (K) Viral DNA, MX2, or MX1 expression in uninfected bystanders or infected monocytes/macrophages of spleen from ASFV-infected pigs at 5 or 7 dpc determined by qPCR.

Not all cell types support ASFV entry and replication. The prevalence of infected cells was investigated on 5 dpc and 7 dpc, respectively. We used scRNA-seq to identify infected cells in vivo based on the presence of ASFV poly-adenylated mRNA. According to a statistical model considering ambient RNA contamination (16), the viral UMIs was used to infer infected and bystander cells. The infected cells and bystander cells in each sample were defined by the Otsu’s thresholding after logarithmic transformation (SI Appendix, Fig. S5A and Dataset S2) (16, 17). Based on the cutoff, we calculated for each cell type the percentage of infected cells among the cells derived from the spleens in the ASFV-infected pigs. We found that relatively large percentages of infected cells were observed in monocytes (ranging between 41.0 and 57.4%) and macrophages (ranging between 11.4 and 30.0%), with the highest percentage of virus-infected cells (57.4%) observed in monocytes (Fig. 2B and Dataset S3). The density plot of viral UMIs in each infected cell also showed that the viral loads increased from 5 to 7 dpc in macrophages and monocytes (Fig. 2C). B cells, T cells, neutrophils, and NK cells were not identified as infected more often than expected by chance (Fig. 2B). None or only a small proportion of these cells were detected to be infected ranging from 0 to 1.8%. We did not observe any infected DCs in the spleen in vivo, though infected DCs have been observed in culture and in lymph nodes (18, 19).

The single-cell suspensions from the spleens were prepared, and the collected cells were incubated with ASFV. The infected cells at different time points were sorted by the FACS method. We found that the monocytes/macrophages were highly susceptible and permissive to ASFV infection, and B-cell and T-cell subsets were nonpermissive to the infection (SI Appendix, Fig. S5B). Besides, the single-cell suspensions from the spleens were prepared, and the lymphocyte (B and T cells) populations were sorted by the FACS method and followed by ASFV infection. The permissiveness of B cells and T cells to ASFV infection was assessed, which also demonstrated that B cells and T cells were nonpermissive to ASFV infection (SI Appendix, Fig. S5C).

Lower Levels of IFN Expression and IFN Response Were Observed in ASFV-Infected Monocytes and Macrophages Than that in Bystander Cells.

To further confirm the scRNA-seq data, the expression consistency of the intersected genes between scRNA-seq and bulk RNA-seq for 5 dpc vs. 0 dpc were analyzed, which showed that both estimates were in general agreement (Fig. 2D). The monocytes/macrophages population in the spleens throughout ASFV infection was sorted by FACS, and the mRNA transcription of several upregulated genes in monocytes/macrophages was detected by qPCR, which was also in agreement with the expression pattern identified by scRNA-seq (SI Appendix, Fig. S6 A and B). Meanwhile, the qPCR analysis verified the expression of five negatively regulated genes (SI Appendix, Fig. S6C), suggesting that the scRNA-seq data provided reliable expression patterns of various host genes (Fig. 2E).

We next explored the antiviral responses in monocytes and macrophages. A recent study indicates that porcine IFNs are effective to suppress ASFV replication (20). The expression of IFN receptors, IFNs, and IFN signature genes (OAS2, MX1, and MX2) in monocytes and macrophages from the spleens of mock-infected or ASFV-infected pigs were evaluated. ASFV CP204L and EP402R were used as indicators of ASFV infection (Fig. 2F). Both monocytes and macrophages show no significant IFN receptors, IFN-β or IFN-γ expression but instead display moderate levels of IFN-stimulated genes (ISGs) after the infection, and the expression of ISGs decreased on 7 dpc compared to that on 5 dpc (Fig. 2F). More importantly, the expression of these ISGs were significantly lower in infected cells than that in uninfected (bystander) cells (Fig. 2G), and ISG scores were lower in infected cells than bystander cells (Fig. 2H). When we examined ISG scores along with viral loads, a significant negative correlation was detected (Fig. 2I), suggesting that ASFV antagonizes IFN-mediated antiviral response to benefit viral replication. The expression of IFNs and ISGs in the tissue of the spleen during ASFV infection was determined by bulk RNAseq (Fig. 2J) and qPCR (SI Appendix, Fig. S6D) as well, which indicated that ASFV infection induces limited IFNs expression and moderate IFN response in the spleen. Meanwhile, the ASFV-infected and bystander monocytes/macrophages in the spleens on 5 and 7 dpc were sorted by FACS, and the mRNA transcription of MX2 and MX1 in monocytes/macrophages were detected by qPCR, which was consistent with the expression pattern identified by scRNA-seq (Fig. 2K). These results confirmed that ASFV infection induces lower levels of IFNs expression and IFN response in ASFV-infected monocytes and macrophages than that in bystander cells.

Characterization of Macrophage Responses during ASFV Infection In Vivo.

Macrophages are of great interest as they are well known to be central to ASFV infection and pathogenesis (7, 8, 14, 21, 22). Consistent with two recent reports (8, 23), a large quantity of host genes were differentially regulated in macrophages after the infection. Importantly, we extend this evaluation to include the case of in vivo infection (Fig. 3A and Dataset S4). We observed that the genes associated with monocyte-to-macrophage differentiation [lysozyme (LYZ) was upregulated, and complement C1q chain (C1Q)A, C1QB as well as C1QC were downregulated], translation, antigen processing and presentation, inflammatory response, and immune response were significantly regulated by infection (Fig. 3 A and B). We characterized genes that were differentially expressed between infected and bystander macrophages, as these could delineate host entry factors, restriction factors, or genes that are modulated by infection within a cell. Surprisingly, there were only 21 differential expressed genes (DEGs) and all these DEGs were downregulated in the infected cells, of which Zinc finger protein 36 (ZFP36), MAF BZIP transcription factor B (MAFB), and dual specificity phosphatase 1 (DUSP1) were the top three DEGs with far more fold-change than others (Fig. 3C and Dataset S5). ZFP36 regulates turnover of MHC class II and other immune-related mRNAs (24). MAFB plays pivotal roles in promoting macrophage M2 polarization (25). DUSP1 has emerged as a critical regulator of innate immune response and inflammation (26, 27), and is involved in regulation of adaptive immune response during Vaccinia virus infection (27). We found that DUSP1 mRNAs increased considerably by day 3 after infection, but dramatically decreased by days 5 and 7 (Fig. 3D), and the expression of DUSP1 was significantly lower in infected cells than in bystander cells (Fig. 3E). Reduced DUSP1 expression on 5 and 7 dpc might be a direct consequence of ASFV infection of macrophages.

Fig. 3.

Characterization of macrophage responses in the spleens of ASFV-infected pigs. (A) Volcano plots of DEGs of 5 dpc vs. 0 dpc, or 7 dpc vs. 0 dpc for macrophages. (B) Dot map of the KEGG pathway and GO terms for DEGs between 5 dpc or 7 dpc and 0 dpc in macrophages. (C) Volcano plot of DEGs between infected and bystander macrophages during the infection. (D) DUSP1 expression in macrophages across ASFV infection in vivo. (E) DUSP1 expression in uninfected bystanders or infected macrophages during the infection. (F) qPCR analysis of ASFV B646L expression in DMSO- or BCI-treated PAMs after ASFV infection ex vivo. (G) Evaluation of ASFV replication in DMSO- or BCI-treated PAMs after ASFV infection ex vivo by the FACS method. The Right panel represents the percentage of infected PAMs treated by DMSO or BCI. (H) Evaluation of ASFV replication in NC siRNA or DUSP1 siRNA-transfected PAMs after ASFV infection ex vivo by the qPCR method. Upper: DUSP1 mRNA expression level; Lower: ASFV B646L mRNA expression level. (I) Western blotting analysis of ASFV B646L expression levels in NC siRNA or DUSP1 siRNA-transfected PAMs (Left), and DMSO- or BCI-treated PAMs (Right) after ASFV infection ex vivo, respectively. (J) The replication of ASFV-GFP in NC siRNA or DUSP1 siRNA-transfected PAMs. (K) Association between DUPS1 expression and the number of viral UMIs according to scRNA-seq data analysis.

To investigate the role of DUSP1 in ASFV infection, an allosteric DUSP inhibitor BCI was used to block DUSP1 activity, and its effect on cell viability was evaluated by Cell Counting kit-8 (SI Appendix, Fig. S7A). PAMs infected with ASFV in the presence of DMSO or BCI were used for analysis of viral loads. We found that the viral gene B646L mRNA expression levels (used as a viral replication indicator) were increased in BCI-treated cells compared to the DMSO-treated cells, indicating that DUSP1 plays an antiviral effect against ASFV infection (Fig. 3F). We confirmed this pattern by the FACS method (Fig. 3G). Moreover, we determined that downregulation of DUSP1 expression in PAMs considerably enhanced ASFV replication at the viral mRNA level (Fig. 3H), suggesting that DUSP1 is a restriction factor for replication of ASFV in macrophages. This antiviral effect was also verified at the protein level by western blotting analysis (Fig. 3I). The densitometric analysis on the western blotting was performed and the statistics have been analyzed to demonstrate consistency of results over independent replicates, which confirmed that DUSP1 is a restriction factor for replication of ASFV in macrophages (SI Appendix, Fig. S7B). The expression of endogenous DUSP1 was actually decreased in PAMs as ASFV infection progressed (SI Appendix, Fig. S7C). Besides, a GFP-tagged ASFV (ASFV-GFP) was used to evaluate the antiviral role of DUSP1, which showed that the replication of ASFV was enhanced by BCI treatment or DUSP1 siRNA transfection (Fig. 3J and SI Appendix, Fig. S7D). The infectious virus titrations were further determined, which also revealed an antiviral role of DUSP1 against ASFV infection (SI Appendix, Fig. S7E). When we examined DUSP1 abundance along with viral loads according to the scRNA-seq data, a significant negative correlation was also detected (Fig. 3K).

Characterization of Monocyte Responses during ASFV Infection In Vivo.

Highest percentage of virus-infected cells were observed in monocytes on 5 and 7 dpc (Fig. 2B and Dataset S3). The host genes differentially regulated in monocytes after ASFV infection were evaluated (Fig. 4A and Dataset S4). We observed that the genes associated with translation, antigen processing and presentation, oxidative phosphorylation, and apoptosis were significantly regulated by the infection (Fig. 4B). The DEGs between infected and bystander monocytes were subsequently investigated. 956 and 1,346 genes were differentially expressed between infected and bystander monocytes at days 5 and 7 after ASFV infection (|log₂FC| > 0.25) of which 218 and 238 changed by >twofold, respectively (Fig. 4C and Dataset S5). A large amount of DEGs were involved in immune response, NOD-like receptor signaling pathway, apoptosis, necroptosis, and response to virus in ASFV-infected pigs (Fig. 4D). The expression of monocytes markers CD14, CD16, and LYZ were lower in infected monocytes than that in bystander monocytes (Fig. 4C). Immune-related genes, including BTG anti-proliferation factor (BTG)1, BTG2, NFκB inhibitor Zeta (NFKBIZ), and C-X-C motif chemokine ligand 2 (CXCL2), were down-regulated in infected cells (Fig. 4C).

Fig. 4.

Characterization of monocyte responses in the spleens of ASFV-infected pigs. (A) Volcano plots of DEGs of 5 dpc vs. 0 dpc, or 7 dpc vs. 0 dpc for monocytes. (B) Dot map of KEGG pathway and GO terms for DEGs between 5 dpc or 7 dpc and 0 dpc in monocytes. (C) Volcano map highlighting the most differentially expressed genes between infected and bystander monocytes. (D) Dot map of KEGG pathways and GO terms for DEGs between infected and bystander monocytes. (E) The top-ranked upregulated host genes in the infected monocytes. (F) The mRNA expression levels of the VAPB gene in uninfected bystanders and infected monocytes. (G) Association between VAPB expression and the number of viral UMIs. (H) Heatmap of GSVA t values for hallmark pathways between each infection stage and baseline (Day 0) in monocytes and macrophages.

The most significantly upregulated gene in infected cells was vesicle-associated membrane protein-associated protein B/C (VAPB) (Fig. 4E). VAPB is a multi-functional protein involved in lipid transport and homeostasis, and membrane trafficking (28). Intriguingly, VAPB was previously shown to positively regulate the replication of rhinoviruses, norovirus, and tombusvirus (29–31). Our data show that VAPB mRNA was upregulated in infected cells compared to the bystander monocytes on both 5 and 7 dpc (Fig. 4F), implying a significant positive correlation with viral loads in vivo (Fig. 4G).

ASFV Causes Massive Death of Macrophages but Shifts Its Infection into a Rare Subpopulation of Immature Monocytes Resulting from Emergency Myelopoiesis.

Macrophages precipitously declined at days 5 and 7 after infection, while monocytes considerably increased and maintained prolonged ASFV infection (Fig. 1G), suggesting that ASFV infection causes the death of macrophages and then shifts its infection into monocytes. Gene Set Variation Analysis (GSVA) results showed that apoptosis, TNFA signaling pathway via NFκB and p53 pathway were activated in the macrophage after infection but were inhibited in monocyte after infection (Fig. 4H). Thus, we speculated that the regulation of these apoptosis-related pathways probably triggered the depletion of macrophages in the spleen following ASFV infection.

The spleen is identified as a critical reservoir of monocytes (32). Human monocytes have been subdivided into three major populations based on relative surface expression of CD14 and CD16 (33, 34), which are the dominant blood monocytes in healthy individuals (34). Monocytes were of particular interest in ASFV-infected animals as the highest percentages of infected cells were observed in monocytes (Fig. 2B and Dataset S3). We focused on investigation of the characteristics of different monocyte subsets during ASFV infection and found that there was substantial heterogeneity within the infected monocytes, with 48.7%, 43.0%, 6.4%, and 1.9% of infected monocytes characterized respectively by expression levels of CD14 and CD16 (Fig. 5A), suggesting that different monocyte subsets show different susceptibility to ASFV infection. This phenomenon could be observed at both day 5 and day 7 after ASFV infection (SI Appendix, Fig. S8A).

Fig. 5.

Immature monocytes are more permissive to ASFV infection than conventional monocytes. (A) Percentage of different subsets of monocytes in all infected monocytes during the infection (5 and 7 dpc). (B) Smoothed expression of CD14 and CD16 for monocytes during ASFV infection. Boxes: CD14⁺CD16⁻, CD14⁻CD16⁺, DN, and DP subsets described in the text; numbers: percentage of cells in each subset at different time points after infection. (C) The infection rates of each subset of monocytes at 5 or 7 dpc determined by the scRNA-seq data. (D) Monocytes marker genes CD14, CD16 and LYZ expression in uninfected bystanders or infected monocytes during the infection. (E) The infection rates of each subset of monocytes at 5 or 7 dpc determined by FACS analysis.

In the spleens from uninfected pigs, these subpopulations were mainly detectable as three clusters, of which >80% of cells manifest high expression of CD16 (Fig. 5B). However, monocyte subpopulations changed dramatically during ASFV infection. CD14⁺ CD16⁺ monocytes (double positives [DPs]) declined, while CD14^– CD16^– cells (double negatives [DNs]) increased (Fig. 5B). While 34.4% of cells fell into DPs bin at baseline (day 0), this dropped to 18.7% and 17.6% on days 5 and 7 after ASFV infection respectively. In contrast, DN monocytes increased from 13.1% at baseline to 29.5% and 32.1% at 5 and 7 dpc, respectively.

During the infection, we found that the DN subpopulation harbored a disproportionately high percentage of ASFV-infected cells, with ~79.6% of DNs being infected on 7 dpc. CD14⁻ CD16⁺ ranked second, with ~57.7% of cells being infected at day 7 (Fig. 5C). This indicated that CD14⁻ monocytes are more susceptible to ASFV than CD14⁺ monocytes, consistent with the fact that CD14 was lower in ASFV-infected monocytes than in bystanders (Fig. 5D). We confirmed this phenomenon by the FACS method (Fig. 5E). DNs were more sensitive to ASFV infection, with 35.97%, and 42.37% of infected cells by day 5 and day 7, respectively. CD14⁻ CD16⁺ ranked second, with 18.97%, and 22.87% of infected cells by day 5 and day 7, respectively. Our FACS data verified that CD14⁻ monocytes are more sensitive to ASFV infection than CD14⁺ monocytes.

After ASFV infection, conventional CD14⁺ monocyte subsets considerably decreased and were replaced by CD14⁻ monocytes. The most frequent subpopulations were DN cells and CD14^– CD16⁺ cells, which rose to make up 76%, and 76.7% of monocytes on days 5 and 7, respectively (Fig. 5B). As most of the infected monocytes did not express CD14 (91.5% on 5 dpc, and 91.9% on 7 dpc) (Fig. 5C), we confirmed that their gene-expression profiles were most correlated with conventional monocytes and not other cell types (SI Appendix, Fig. S8B). These data implied that monocyte subtypes played distinct roles in ASFV infection. Besides, the fixed spleen tissues were stained with anti-P72 (a late ASFV capsid protein) and anti-CD14 antibodies, which also showed that CD14⁺ cellular subsets considerably decreased and were replaced by CD14⁻ subsets after ASFV infection (SI Appendix, Fig. S9A), and virus factory formation could be clearly observed at days 5 and 7 (SI Appendix, Fig. S9B). The ISG scores in CD14⁻ and CD14⁺ monocytes were evaluated and compared to investigate the antiviral immune responses in the two subpopulations. The ISG scores increased moderately for both CD14⁻ and CD14⁺ monocytes on 5 and 7 dpc. Interestingly, the ISG scores were lower in CD14⁻ monocytes than CD14⁺ monocytes after ASFV infection (SI Appendix, Fig. S10A). Besides, GSVA pathway analyses showed that apoptosis-related pathways were less activated in CD14⁻ monocytes than CD14⁺ monocytes (SI Appendix, Fig. S10B). This might explain why CD14⁻ monocytes are more sensitive to ASFV infection than CD14⁺ monocytes.

Immature myeloid cells can be released from the bone marrow under stress conditions, a process known as emergency myelopoiesis (35–37). Bone marrow monocyte precursors (BM-MPs) have lower expression of CD14 and CD16 and higher expression of S100A8 and S100A9 compared to circulating monocytes (38). S100A8 and S100A9 are constitutively expressed by myeloid cells, including myeloid precursors, polynuclear neutrophils (PNN), monocytes, and early-differentiation cells of the myeloid lineage, but not by lymphocytes (39–43). If DNs result from emergency myelopoiesis, their gene expression may be more similar to BM-MPs than circulating monocytes. In our data, the expression of CD14, CD16, S100A9, and S100A8 in DNs during ASFV infection relative to baseline monocytes mirrored BM-MPs, suggesting their similarity (SI Appendix, Fig. S10C). We therefore hypothesized that the CD14 negative populations may derive from emergency myelopoiesis.

MHC Class II Genes Are Downregulated Throughout ASFV Infection Independent of Infection Status.

After elucidating the antiviral immune responses in monocytes and macrophages, we next investigated and compared the changes in gene-expression profiles of different cell types throughout ASFV infection. We first compared transcriptomes from the infection stage to mock-infection stage for each cell type and found that monocytes had significant gene-expression changes among the seven cell types, which ranked second in all ASFV infection stage versus mock-infection stage (6.8%, 9.4% and 10.7% of genes tested) (SI Appendix, Fig. S11A). DCs ranked first on 3 and 5 dpc (7.1% and 10.6% of genes tested, respectively), and B cells ranked first on 7 dpc (11.1% of genes tested). The relative proportion of monocytes was extremely higher than DCs on days 5 and 7 (SI Appendix, Fig. S11B). We therefore focused our attention on characterizing monocytes.

One prominent feature of the monocyte differential expression profile was the decline in several major histocompatibility complex (MHC) class II gene expression levels by day 5 and day 7 after ASFV infection, including swine leukocyte antigen (SLA)-DQA1, SLA-DQB1, SLA-DRA, SLA-DRB1, SLA-DMA, SLA-DMB, and SLA-DOA (SI Appendix, Fig. S11C and Dataset S4). Previous study showed that genotype I ASFV does not modulate MHC class II levels in monocytes in vitro (7). However, we observed widespread changes in MHC class II but not MHC class I gene expression throughout genotype II ASFV infection in vivo (SI Appendix, Fig. S11C).

Widespread changes in MHC class II expression were observed throughout ASFV infection (SI Appendix, Fig. S11C). The most significant decreases occurred in MHC class II genes of monocytes, with slight changes in many MHC class I genes (Dataset S4). B cells displayed modest reductions in MHC class II genes. DCs showed modest decreases in MHC class II genes as well. Macrophages manifest smaller changes in MHC class II genes. The decreased expression of MHC class II genes was observed in both permissive and nonpermissive cells throughout the infection (SI Appendix, Fig. S11C). Meanwhile, no significant change in MHC class II gene expression was observed between the infected cells and bystander monocytes (SI Appendix, Fig. S11D).

The genes correlated with MHC class II in monocytes were identified, as they may be part of a co-regulated expression program. Many of the most correlated genes are functionally involved in antigen presentation, including CD74, PLVAP, C1QB, C1QA, C1QC, CD83, and DNAJB1 (SI Appendix, Fig. S11C). One of the most associated genes was CD74, which is identified as a transcription regulator (44). Thus, MHC class II and other genes involved in antigen presentation may be part of a single transcriptional module, co-regulated by CD74 and/or other genes.

To reveal the coregulated expression relationship between host and virus in monocytes in vivo, the interplay between viral genes and the host genes in monocytes during ASFV infection was investigated. First, the differentially expressed genes (DEGs) between cells at 5 dpc or 7 dpc and at 0 dpc were identified, and the union set of the DEGs on day 5 and day 7 was obtained. Then the 159 viral genes and the above union set of DEGs with average expression values were clustered using R package TCseq. Six time-dependent expression patterns were identified and their biological significance was determined (SI Appendix, Fig. S11E and F, and Dataset S4). The antigen processing and presentation, and Toll-like receptor signaling pathways were impaired, and the oxidative phosphorylation and mitochondrial respiratory associated genes associated with viral infection were upregulated during viral infection (SI Appendix, Fig. S11E and F). We therefore hypothesized that ASFV infection downregulates many host antiviral genes and upregulates putative pro-viral genes in monocytes.

Discussion

The outbreak of ASF can cause drastic economic losses to the animal industry and leads to a disruption in the livestock supply chain and meat consumption structure (45). The multi-organ dysfunction or failure in pigs with ASF is greatly correlated with the high-efficient replication of ASFV in cells composing these organs. However, how ASFV manipulates the infected cell population and controls host transcriptomes to drive high-efficient replication of the virus and prolonged infection in vivo remains unclear. In this study, we used scRNA-seq approach to comprehensively survey the molecular correlates of disease progression and viral replication in the immune cells of spleens from ASFV-infected pigs. The study determined the changes in cell-type and -state abundance throughout ASFV infection, characterized the ASFV-infected cell populations in the spleen, and identified genes regulated by the cytokine milieu or by direct ASFV infection (SI Appendix, Fig. S11, SI Appendix, Fig. S12), which draw the factors that drive splenic lesions by ASFV infection in vivo.

Hemorrhagic lesions in ASF are accompanied by impaired hemostasis (4). These pathological changes have been well described which suggest that high levels of virus in blood are associated with the development of hemorrhagic disease (46–49). By day 3 post-infection, no virus transcripts were detected in the spleen although clinical signs and decreased numbers of spleen cell populations were observed. It appears that initial virus replication is higher in blood or other tissues. The viral DNA could be detected in the blood of one infected pig at day 3. During the period between the third day and fifth day, the virus might replicate quickly and reach an extremely high level and then spread to various tissues. As the largest lymphatic organ and a reservoir for blood in the pig, spleen functions in the removal of aging erythrocytes, recycling iron, eliciting immunity, and production of antibodies (50–52). However, a massive destruction of the spleen was observed after ASFV infection with severe hemorrhagic lesions and diffuse enlargement, and the spleen harbored the highest viral loads among various tissues from infected pigs. We investigated transcriptional-level changes in the immune cells in spleens of pigs during ASFV infection, some of which reflect their dynamics and suppression of their physiological antiviral function. We observed that the relative abundance of monocytes and macrophages dramatically changed during the viral infection in vivo. Monocytes had over fivefold increase throughout the infection, while macrophages reflect more than threefold decrease in abundance. Macrophages are at the front line of defense against viruses that play important roles in the clearance of viruses (53). The reduction of macrophages might block viral clearance in vivo and result in severe disease. Monocytes increase in the spleen by days 5 and 7 after the infection, suggesting that ASFV infection causes the death of macrophages and then shifts its infection into monocytes to affect cellular function and cause splenic lesions. Lymphopenia was described earlier in the case of genotype II ASFV infection in pigs (54). Our study also proved a tendency toward lymphopenia. The role of apoptosis in depletion of T and B lymphocytes in peripheral blood, lymph nodes, and spleen has been reported in previous studies (54–59). ASFV also induces apoptosis in the infected cells including macrophages and monocytes (60–62). In the present study, GSVA results showed that apoptosis, TNFA signaling pathway via NFκB and p53 pathway were activated in the macrophage after infection. This explains why massive death of macrophages was observed. In contrast, monocytes considerably increased and maintained prolonged ASFV infection in the spleen. An increase in monocytes from blood of ASFV-infected domestic pigs at late infection stage had been observed previously (58, 63). The apoptosis-related pathways were less activated in the increased monocyte subset (CD14⁻ monocytes) in the spleen. Besides, these immature monocytes may derive from emergency myelopoiesis. This may be the reason for observation of increased monocyte counts in the spleen after the infection. The characteristics of the monocytes from blood of ASFV-infected pigs at late infection stage should be further investigated.

Monocytes are a type of immune cell that originates in bone marrow. Although initially perceived as a homogeneous population (64), heterogeneity of the monocyte population has long been recognized and, in part, is a result of the specialization of tissue monocytes in particular microenvironments. Different monocyte subsets seem to show developmental stages with distinct physiological roles, such as recruitment to inflammatory lesions or entry to normal tissues (64, 65). As the main cells targeted by ASFV, monocytes had higher viral-load heterogeneity than other cell types.

ASFV infection-induced IFN activates monocytes, which could further upregulate MHC class II expression (66–68). Notably, the main MHC class II expression was dramatically decreased in both infected and bystander monocytes, indicating that the reduction was induced by the cytokine milieu but not direct infection of the cells by ASFV. DCs displayed modest reductions in MHC class II as well. However, the relative proportion of monocytes was extremely higher than DCs after ASFV infection. Therefore, decreased monocytes antigen processing and presentation might make it easy to understand why a failed or delayed adaptive immune response is a characteristic of fatal ASF in pigs.

Among the cells targeted by ASFV, we observed that monocytes sustain the main viral infection during the infection at acute infection stage in the spleen. More than half of monocytes (up to 57.5% of monocytes) were infected by day 7 after the infection, compared to ~30% for macrophages, which ranked second. The relative abundance of monocytes dramatically rose from day 5 onward that corresponded with the increase of ASFV-positive cells. As the infection progressed, conventional CD14⁺ monocyte subsets, which play vital roles in other viral infections (69–71), considerably decreased and were replaced by CD14⁻ monocytes. Intriguingly, we observed that more than 90% of the infected monocytes belong to the CD14⁻ subsets, an unexpected tropism. The CD14⁻ CD16⁻ monocytes were immature monocytes, which are more homogeneous to bone marrow monocyte precursors but not circulating monocytes. This indicates that the increased abundance of CD14⁻ CD16⁻ monocytes is due to the emergency myelopoiesis (36). This points at the advantage of scRNA-seq: regardless of limited detection of CD14 or CD16, other specific RNA markers can be used to define these immature cells as monocyte precursors. PAMs are more susceptible to ASFV infection in vitro in comparison with cultured monocytes (72), while our data show that monocytes maintain the prolonged infection in the spleen in vivo. This might be due to that the infectivity strongly correlates with the physiological variability of monocyte differentiation state. CD14⁺ monocytes comprise the main cultured monocytes in vitro, but CD14⁻ monocytes are more permissive for ASFV infection compared to CD14⁺ monocytes, and CD14⁻ monocytes comprise the main infected cell population in vivo.

Until now, roles of immature CD14⁻ monocyte in ASFV infection have been overlooked due to that they do not express classical monocyte marker CD14. Previous publications demonstrated that susceptibility to in vitro infection with ASFV of monocytes/macrophages was restricted to intermediate or later stages of differentiation (73). In our study, the proportion of monocytes increased from 3% at baseline to 12.1% at 5 dpc, and 16.2% at 7 dpc. The monocytes comprised the main infected cell population at 5 and 7 dpc with high levels of viral loads, and >90% infected monocytes belonged to CD14⁻ monocyte-lineage, suggesting that the immature CD14⁻ monocytes are critical for prolonged infection of ASFV in vivo. The suppression of apoptosis, IFN response, as well as antigen-presentation capacity by ASFV in CD14⁻ monocytes has contributed to the efficient and prolonged infection of ASFV. The immature population of monocytes represents the major population infected in the spleen at days 5 and 7 post-infection of pigs. High levels of ASFV P72 protein could be observed in the CD14⁻ cells (SI Appendix, Fig. S9). This indicated that the infection progressed to a late stage in CD14⁻ cells. However, this may not lead to an amplification of infectious virus. High expression levels of ASFV proteins in these CD14⁻ cells can affect cellular function and have an impact on bystander cells’ function, resulting in splenic lesions. Thus, a strategy targeting CD14⁻ monocytes could be a future target for antiviral drugs against ASFV infection. Besides, this will provide insights for establishment of sustainable ASFV-sensitive cell lines. Our data indicate that the primary infected population in the spleen is a CD14⁻ monocyte subset. Myeloid cells have been recognized as the main targets of ASFV (7), including DCs ex vivo and in lymph nodes in vivo (18, 19). However, in the spleen in vivo the viral RNA were only detected in monocytes and macrophage among immune cells. The cell-to-cell contact or specific cytokines treatment might be required for DC infection.

DUSP1 shows an antiviral function against ASFV. DUSP1 switches off mitogen-activated protein kinase (MAPK) to maintain IRF1 expression and leads to increased expression of IRF1-dependent genes in pulmonary A549 cells (74). IRF1 contributes to innate immunity against many pathogens (75). We hypothesize that DUSP1 restricts ASFV replication through the DUSP1-MAPK-IRF1 axis in porcine immune cells. The agonists targeting this pathway might have antiviral effects against ASFV. Besides, our scRNA-seq and bulk RNA-seq data determined several molecular commonalities between ASF and Ebola virus disease, which also reveals significant decrease of MHC class II expression in monocytes and emergency myelopoiesis (38).

In summary, this study broadens our understanding of ASF and provides a far more nuanced view of host and viral gene expression. We found that ASFV causes massive death of macrophages but shifts its infection into CD14⁻ immature monocytes that maintain the prolonged infection of ASFV in vivo. Future studies need to be focused on the important roles of CD14⁻ monocytes or key molecules in ASFV infection and pathogenicity, which may guide research on antiviral strategies.

Materials and Methods

Infections.

All experiments with live ASFV in the present study were performed within the biosafety level 3 (BSL-3) facilities. The animals were inoculated intramuscularly with the indicated amounts of ASFV CN/GS/2018 or equal amounts of solvent control. The animals were slaughtered and dissected at indicated time. All the challenged animals were monitored daily for clinical signs. The spleens collected at 0, 3, 5, and 7 d after infection were used for preparing the single-cell suspensions.

Single-Cell Preparation and scRNA-seq.

The spleens were finely minced with scissors and digested in RPMI medium containing 3 mg/mL Collagenase D (Roche) for enzymatic dissociation. Cell suspensions were subjected to red blood cell lysis before enrichment. For each sample, cell viability exceeded 90%. The scRNA-seq libraries were generated using Chromium Single Cell 3′ Library & Gel Bead Kit v.2 (10× Genomics) according to the manufacturer’s protocol. Libraries were sequenced on Illumina HiSeq X Ten.

scRNA-seq Data Processing.

A hybrid genome of Sus scrofa (genome assembly Sscrofa11.1) and ASFV was constructed. The raw sequencing data were processed by the Cell Ranger pipeline (v6.1.2). The Seurat objects were merged and used for downstream analysis. The integrated matrix was scaled, and the top 30 dimensions resulted from the principal component analysis (PCA) were used for the uniform manifold approximation and projection (UMAP). The shared nearest neighbor graph-based clustering was performed on the PCA-reduced data to identify cell clusters. The cell identities were determined based on the expression of canonical marker genes.

GO Term and Pathway Enrichment Analysis of DEGs.

GO term and pathway enrichment analysis of DEGs was performed using DAVID (version 2021; https://david.ncifcrf.gov/). The results were visualized using the ggplot2 R package (v3.3.5).

Gene Set Score Analysis.

The “AddModuleScore” function from Seurat (v4.1.0) was used to calculate the gene set score for each cell. We then applied the two-tailed Wilcoxon rank-sum test between the scores of different groups. R package GSVA (v1.44.3) were performed to assess the relative pathway activities in the macrophages and monocytes. The hallmark pathways for GSVA were obtained from the MSigDB database.

Western Blots.

Cellular lysates were detected using anti-p72 or anti-DUSP1 antibodies. The appropriate HRP-conjugated secondary antibodies were used to detect the antigen–antibody complexes, which was then visualized using enhanced chemiluminescence detection reagents (Thermo Fisher Scientific).

Hematoxylin-Eosin Staining.

The tissue samples were collected and fixed with 4% paraformaldehyde for at least 24 h, dehydrated, and paraffin embedded. The treated tissues were sectioned at 4 µm and stained with hematoxylin and eosin (H&E). The score takes into account each of the five rating areas (inflammation, necrosis, congestion, fibrosis, and other histological changes) and is out of 20 total points.

Flow Cytometry (FACS).

For analysis of the composition of ASFV-infected cells, pig spleen cells were stained with antibodies of the indicated proteins in PBS containing 2% (wt/vol) fetal bovine serum (FBS). The staining was performed following the manufacturer’s instruction (BD Biosciences). At least 10,000 to 20,000 cells were collected by BD LSR Fortessa (BD Biosciences) and analyzed using FlowJo software (BD Biosciences).

Detailed Description of the Materials and Methods.

Detailed procedures of the aim and study design, animals, cells and virus, infections, blocking ELISA, single-cell preparation and scRNA-seq, scRNA-seq data processing, differential expressed genes (DEGs) analysis, GO term and pathway enrichment analysis of DEGs, comparison with bulk RNA-seq dataset, categorization of infected versus bystander cells, Gene set Score analysis, identifying time-dependent transcriptional program in ASFV-infected monocytes, construction of the ASFV-GFP using the homologous recombination method, measuring viral DNA in various samples, qPCR, DUSP1 inhibitor assay, siRNA knockdown assay, western blots, hematoxylin–eosin staining, flow cytometry (FACS), tissue immunofluorescence, HAD₅₀ test assay, and statistical analysis are described in SI Appendix, Materials and Methods.

Supplementary Material

Appendix 01 (PDF)

Dataset S01 (XLS)

Dataset S02 (XLS)

Dataset S03 (XLS)

Dataset S04 (XLS)

Dataset S05 (XLSX)

Data, Materials, and Software Availability

The bulk RNA-seq data have been deposited on the Figshare website (76), and raw 10× Genomics scRNA-seq data have been deposited at the NCBI SRA database with accession number of PRJNA879060 (77). All original codes have been deposited at Github (78). All other study data are included in the manuscript and/or supplementary data.