Abstract

   African swine fever virus (ASFV), causing an OIE-notifiable viral
   disease of swine, is spreading over the Eurasian continent and
   threatening the global pig industry. Here, we conducted the first
   proteome analysis of ASFV-infected primary porcine monocyte-derived
   macrophages (moMΦ). In parallel to moMΦ isolated from different pigs,
   the stable porcine cell line WSL-R was infected with a recombinant of
   ASFV genotype IX strain “Kenya1033”. The outcome of the infections was
   compared via quantitative mass spectrometry (MS)-based proteome
   analysis. Major differences with respect to the expression of viral
   proteins or the host cell response were not observed. However,
   cell-specific expression of some individual viral proteins did occur.
   The observed modulations of the host proteome were mainly related to
   cell characteristics and function. Overall, we conclude that both
   infection models are suitable for use in the study of ASFV infection in
   vitro.

   Keywords: African swine fever virus, proteome, primary cell culture,
   mass spectrometry

1. Introduction

   African swine fever virus (ASFV, family Asfarviridae) is a highly
   virulent virus that infects swine (family Suidae) for which, currently,
   neither a vaccine nor a treatment is available. Its double-stranded
   DNA-genome contains more than 150 open reading frames (ORFs). However,
   for many of them, information regarding functions and interactions is
   still limited. For roughly one third of the ASFV ORFeome, no
   information about the expression of a corresponding protein in
   mammalian cells is available [[36]1]. This knowledge gap, which is
   partially due to the lack of appropriate immunologic reagents, has been
   partially filled by the application of MS-based proteomics using
   infected stable cell cultures [[37]2,[38]3] but expression data are
   still missing for ASFV-infected macrophages, the natural target cell of
   ASFV. The genome of ASFV isolates is quite variable. While homologs of
   a core set of conserved viral genes encoding for viral structural
   proteins or proteins involved in virion morphogenesis can be found in
   all 24 genotypes currently defined, the composition of a genome can
   vary with respect to the presence or absence of other genes [[39]4];
   among them are sets of paralogous genes known as multi gene families
   (MGFs). MGF proteins have been described as virulence factors, host
   range determinants or modulators of host response to infection
   [[40]5,[41]6,[42]7,[43]8]. Like other viruses, ASFV is known to evade
   the antiviral response by modulating immune-related pathways, such as
   NFκB-mediated signaling or the interferon response. The viral proteins
   directly involved in immune evasion are only known for some of these
   pathways (reviewed in [[44]9]).

   Macrophages, the primary target cells of ASFV [[45]10], are phagocytic
   cells belonging to the myeloid lineage that differentiate from
   monocytes in response to growth factors such as granulocyte-monocyte
   colony-stimulating factor (GM-CSF) [[46]11,[47]12]. As professional
   antigen-presenting cells, macrophages express swine leukocyte antigens
   (SLAs) I and II, homologues of human major histocompatibility complexes
   I and II, on their surface. This complex is formed by SLA-DR and SLA-DQ
   proteins, both consisting of alpha and beta chains encoded by SLA-DRA
   and SLA-DRB or by SLA-DQ1 and SLA-DQB genes, respectively [[48]13].
   Apart from SLA complexes, swine macrophages are characterized by the
   expression of specific clusters of differentiation (CD) proteins such
   as CD14, CD172a (SWC3), CD68 and the scavenger receptor CD163
   [[49]11,[50]14].

   Many studies about ASFV replication and host interactions have been
   performed using established cell culture models, such as Vero cells
   infected with cell culture adapted viruses such as BA71V [[51]15]. In
   this way, some pitfalls that come with the use of primary cell cultures
   in infection experiments (animal health issues, higher variability) can
   be avoided, but presumably at the cost of a more artificial infection
   model.

   While the transcriptomes of ASFV-infected cell lines [[52]16] or
   primary macrophages have been studied in vitro [[53]16,[54]17] and in
   vivo [[55]18], proteome studies of ASFV-infected cells [[56]2] or ASFV
   particles [[57]3,[58]19] using shotgun MS platforms are rare [[59]20].
   A recent proteome study with ASFV-infected porcine alveolar macrophages
   (PAMs) noted a differential regulation of host proteins involved in
   immune processes and defense responses [[60]21]. Older studies of the
   ASFV proteome using 2D electrophoresis, optionally combined with
   MALDI-TOF MS, identified a limited number of proteins that were
   influenced by the infection, among them heat shock proteins,
   redox-related proteins, and nucleoside diphosphate kinases
   [[61]22,[62]23,[63]24,[64]25,[65]26,[66]27,[67]28].

   Here, we conducted the first MS-based quantitative proteome analysis of
   in vitro ASFV-infected primary porcine monocyte-derived macrophages.
   For comparison, the stable wild-boar-derived cell line WSL-R [[68]29],
   which is commonly used for in vitro infection experiments with ASFV,
   was analyzed in parallel to monitor any differences in the expression
   patterns of ASFV proteins in primary cells and a stable cell line. The
   modulation of the host proteome after infection was also monitored via
   MS analysis and compared between the two cell types using
   biostatistical means. ASFV-Kenya1033 was isolated from domestic pigs in
   western Kenya as a virulent genotype IX strain (L. Steinaa & R. Bishop,
   personal communication). For this study, the recombinant
   ASFV-Kenya1033-ΔCD2v-dsRed was chosen for its proven ability to
   replicate well in both WSL-R cells [[69]30] and macrophages [[70]31],
   which is prerequisite for a proteome analysis focusing on the
   cell-specific expression of viral and host genes after infection with
   ASFV.

2. Materials and Methods

2.1. Safety Statement

   All ASFV-infection experiments were carried out in a biocontainment
   facility fulfilling the safety requirements for ASF laboratories and
   animal facilities (Commission Decision 2003/422/EC, Chapter VIII).

2.2. Cells and Viruses

   The wild boar lung-derived cell line WSL-R (hereafter ‘WSL’ cells)
   [[71]29], obtained from the Biobank of the FLI (CCLV-RIE #0379), was
   cultured in Iscove′s modified Dulbecco’s medium mixed with Ham’s F-12
   nutrient mix (1:1; v/v) supplemented with 10% fetal bovine serum (FBS)
   (during cultivation) or 5% FBS (after inoculation).

   Virus stock of ASFV-Kenya1033-ΔCD2v-dsRed [[72]30] was grown on WSL
   cells. Titers were determined on WSL cells and peripheral blood
   monocytic cells (PBMCs) as TCID[50]/mL [[73]32].
   ASFV-Kenya1033-ΔCD2v-dsRed, derived from highly pathogenic genotype IX
   isolate ASFV-Kenya1033 (isolated February 2013), expressed dsRed as a
   fluorescent reporter under the late p72 promotor in the CD2v locus. The
   generation and partial characterization of the recombinant has been
   described elsewhere [[74]30]. Both viruses replicate well in WSL cells,
   but titers of the recombinant are slightly reduced in comparison to the
   parental virus.

2.3. Isolation and Cultivation of Porcine Macrophages

   Blood was drawn from 6- to 12-month-old female domestic pigs kept at
   the FLI animal facility (permission: LALLF-Nr. 7221.3-2-041/17). PBMCs
   were obtained using Pancoll animal density gradient medium (PanBiotech,
   Aidenbach, Germany) with a density of 1.077 g/mL, in accordance with
   the manufacturer’s recommendations. For the magnetic cell separation of
   CD172a^+ PBMCs, anti-mouse IgG1 magnetic particles (BD Biosciences,
   Franklin Lakes, NJ, USA) in combination with α-SWC3 hybridome
   supernatant (clone 74-22-15) were used according to established
   procedures. CD172a^+ PBMCs were cultured in Primaria^TM cell culture
   dishes (Corning, New York, NY, USA) with PC-1^TM Medium (Lonza, Basel,
   Switzerland) containing 1% penicillin/streptomycin solution (v/v,
   ThermoFisher Scientific, Waltham, MN, USA). Cells were incubated at 37
   °C and 2.5% CO[2].

   Differentiation into mature moMΦ was induced using a modified protocol
   based on Carlson et al., 2016 [[75]33] and Carlson et al., 2020
   [[76]31]. Briefly, one day after isolation, cells were washed three
   times with PBS before treatment with 5 ng/mL GM-CSF (KingFisher
   Biotech, Saint Paul, USA, #RP0940S) applied in fresh culture medium.

2.4. Phenotypic Characterization of Monocyte Derived Macrophages

   For fluorescence analysis, moMΦ were grown on glass cover slides and
   infected with ASFV-Kenya1033-ΔCD2v-dsRed 24 h post differentiation
   using a MOI of 2. At 6 or 18 hpi, moMΦ were fixed in 3.75% (v/v)
   formaldehyde in PBS, permeabilized with 0.5% Triton X-100 in PBS and
   stained for actin using an Alexa568-conjugated phalloidin derivate
   (0.33 µM, Invitrogen). Viral proteins p30 and p72 were detected using
   monospecific rabbit antisera [[77]30,[78]34] and AlexaFluor 647
   conjugated secondary antibodies. Nuclei were stained using Hoechst33342
   (Invitrogen, Waltham, MA, USA). Samples were analyzed using a Leica
   DMi8 microscope. Pictures were processed using ImageJ (v1.52 h, U. S.
   National Institutes of Health, Bethesda, MD, USA) software.

   To assess the expression of CD14, CD172a, CD163, SLA II and CD68 in
   moMΦ via flow cytometry, the cells were harvested 24 h after
   differentiation, washed in FACS buffer (0.1% sodium azide, 0.1% FCS, 1
   mM EDTA in PBS), sedimented (350× g, 4 °C, 5 min) and stained for
   surface epitopes CD14, CD172a, CD163 and SLAII (10 min, 4 °C, dark)
   using appropriate antibodies; for details, see [79]Table S1. Cell
   viability was checked with Zombie Aqua staining reagent (Biolegend, San
   Diego, CA, USA) in accordance with the manufacturer’s directions. After
   fixation and permeabilization with True Nuclear Transcription Factor
   Buffer Set (Biolegend), intracellular CD68 was stained for 15 min at
   room temperature (RT) in the dark. Samples were measured with a BD
   FACSCanto™ II (BD Biosciences, BD FACSDiva™ Software v9.0.1, BD
   Biosciences, Franklin Lakes, NJ, USA) flow cytometer and the results
   processed via FlowJo™ v10.7 (BD Biosciences, Franklin Lakes, NJ, USA)
   software. See [80]Figure S1 for details of the gating strategy.

2.5. Infection with ASFV-Kenya1033-ΔCD2v-dsRed

   To enhance infection rates, cells were centrifuged (60 min, 37 °C, 600×
   g) during inoculation with ASFV-Kenya1033-ΔCD2v-dsRed with MOI2
   ([[81]30], JH Forth, L Käbisch, R Portugal, S Blome, GM Keil,
   manuscript in preparation).

   For proteome analysis of moMΦ, approximately 5 × 10^6 cells were
   detached from the cell culture plates 24 hpi via incubation with 4 °C
   cold PBS supplemented with 50 mM EDTA, pelleted (250× g, 5 min, 4 °C)
   and harvested in 250 µL of lysis buffer (2% SDS in 100 mM Tris-HCl, pH
   8.0). WSL cells were washed with PBS three times and harvested 48 hpi
   in 500 µL of lysis buffer. The longer incubation time for WSL cells was
   chosen to ensure that all cells were infected and had progressed to the
   late stage of infection [[82]2,[83]30]. For lysis and the inactivation
   of ASFV, the mixtures were incubated at 95 °C for 10 min, cooled to RT
   and clarified by centrifugation (5 min, 10,000× g, 20 °C). This
   supernatant is referred to as ‘lysate’.

2.6. Generation and Analysis of MS-Samples

   For shot-gun proteomic analysis, lysates were supplemented with 0.5%
   dithiothreitol for reduction and digested by Filter Aided Sample
   Preparation (FASP) [[84]35] using Trypsin (Promega #V5111, Madison, WI,
   USA) at an enzyme to substrate ratio of 1:50. Desalted peptides were
   separated using nano reversed phase liquid chromatography. Per sample,
   1360 fractions were spotted to a MALDI target plate and analyzed via
   MALDI-TOF/TOF MS with an Ultraflextreme instrument (Bruker, Bremen,
   Germany). For details see [85]Supplementary Methods.

   Protein abundances in mol percent (%mol) were calculated on the basis
   of the exponentially modified protein abundance index (emPAI) [[86]36].
   MS data are available at JPOST under the identifier JPST001339
   ([87]https://repository.jpostdb.org/entry/JPST001339).

2.7. Statistics and Software

   All calculations were performed using the statistical language R
   [[88]37]; graphs were generated with the R-package ggplot2 [[89]38] and
   with BioRender.com (accessed on 30 September 2021).

   The following study design was applied: moMΦ were isolated from three
   pigs. From one animal, mock-infected and ASFV-infected cells were
   analyzed in triplicate together with mock and infected triplicates of
   WSL cells. Infected moMΦ from the other two animals were also analyzed
   in triplicate to assess the variability of ASFV protein expression in
   cells from different animals. For the biostatistical analysis of host
   protein expression levels, porcine protein identifiers were referenced
   to the corresponding genes using the R-package gprofiler2 (version
   0.2.1) [[90]39]. Gene products were only considered for statistical
   analysis if they were expressed in at least one of the three replicates
   from two of the three pigs or in two of the three WSL cell samples.
   Genes were only considered as differentially expressed (differentially
   expressed genes, DEG) under different conditions if expression levels
   changed more than 2-fold and p-values of a two-sided t-test were <0.05.

   Host gene lists for enrichment analysis of Gene Ontology (GO) terms
   [[91]40] or Kyoto Encyclopedia of Genes and Genomes (KEGG) pathways
   [[92]41] were compiled from the output lists of ProteinScape software
   using R scripts. These lists contained DEGs and also the genes that
   were identified only under one of the compared conditions (infected
   versus naïve or moMΦ versus WSL). GO term and KEGG pathway enrichment
   analysis was performed using CytoScape (version 3.8.2) [[93]42]
   together with the package ClueGO (version v2.2.5.7) [[94]43].

3. Results

   The goal of this study was to compare the expression of ASFV genes and
   host response in primary porcine moMΦ and WSL cells. Briefly, moMΦ were
   differentiated from PBMC-derived, magnetically sorted CD172a^+
   monocytes, characterized and infected with ASFV. Proteomes of
   ASFV-infected moMΦ and WSL cells were compared via HPLC-MS analysis
   following the workflow outlined in [95]Figure 1.

Figure 1.

   [96]Figure 1
   [97]Open in a new tab

   Workflow for generation, processing and analysis of samples for
   proteome analysis. (A) Isolation, selection and differentiation one day
   after isolation and quality control of primary moMΦ, (B) ASFV
   infection, sample generation and processing for MS analysis, (C)
   analysis of peptides by HPLC-MALDI-TOF/TOF MS workflow and data
   processing.

   In order to achieve the high infection rates required for the unbiased
   comparison of both cell types, moMΦ and WSL cells were harvested at 24
   hpi and 48 hpi [[98]2], respectively.

3.1. Characterization of Primary Macrophages

   To avoid the accumulation of BSA from the medium by the moMΦ, BSA- and
   serum-free culture conditions had to be used. Under these conditions,
   the prepared cells expressed a panel of moMΦ-specific CD markers which
   were assessed via flow cytometry ([99]Figure 2A), maintained their
   characteristic morphology, and remained susceptible to ASFV infection,
   as confirmed using immunofluorescence microscopy ([100]Figure 2B).
   Additionally, macrophage characteristic CD markers were detected by MS
   in moMΦ but not in WSL cells ([101]Figure 3A).

Figure 2.

   [102]Figure 2
   [103]Open in a new tab

   Characterization of moMΦ. (A) Phenotypic characterization of moMΦ by
   flow cytometry. Surface markers SLA class II, CD14, CD68, CD163 and
   CD172a are expressed on >95% of all moMΦ. (B) Morphologic
   characterization by fluorescence microscopy. Representative data from
   different preparations are shown. moMΦ were infected with ASFV at MOI2,
   fixed at 6 or 18 hpi, and stained for actin with phalloidin (green) to
   visualize the cell cortex. Early (p30, 6 hpi) or late (p72, 18 hpi)
   ASFV proteins were immunostained (violet) with appropriate antibodies.
   Bar indicates 25 µm.

Figure 3.

   [104]Figure 3
   [105]Open in a new tab

   Comparison of host protein expression between WSL and moMΦ and in
   response to ASFV infection. Expression of CD markers and SLA II in
   naïve WSL and moMΦ (A) and infected and mock-infected moMΦ (B). Volcano
   plots representing host gene expression levels in response to infection
   in moMΦ (C). The horizontal line corresponds to p = 0.05 (two-sided
   t-test), vertical lines to fold-changes of 2 and 0.5, respectively, red
   dots indicate host genes highlighted in previous ASFV-related
   publications.

   To confirm the moMΦ phenotype using protein expression profiling, we
   performed GO term enrichment and GO term clustering analysis of genes
   that were differentially expressed in naïve moMΦ and WSL cells using
   CytoScape and ClueGO software ([106]Supplementary file
   ‘S5_Gene_lists.xlsx’ sheet ‘Gene lists for EA’). Typical biological
   pathways related to phagocytic antigen-presenting cell functions such
   as endocytosis (GO:0006897; KEGG:04144), phagocytosis (GO:0006909;
   KEGG:04145), signal transduction (GO:0007165; GO:0007186) and
   immune-related pathways (KEGG:04612) were over-represented
   (Bonferroni-corrected p-value < 0.05) in moMΦ ([107]Supplementary Table
   S5 ‘S5_Gene_lists.xlsx’, sheet ‘Results enrichment analysis’). DEGs
   overexpressed in moMΦ included the macrophage mannose receptor (MRC1),
   the superoxide dismutase (SOD2), biliverdin reductase (BLVRB) and 11
   genes related to the lysosome (KEGG:04142), such as Cathepsin A (CTSA)
   and ß-hexosaminidase subunits α and β (HEXA, HEXB). Vice versa, some
   proteins linked to the fatty acid metabolism (FASN, FABP3) or the TNF
   receptor-associated protein 1 (TRAP1) were overexpressed in WSL
   ([108]Supplementary File ‘S4_MS-statistics.xlsx’, sheet ‘Q_Cell type’,
   Figure S4a).

   These results indicated that the typical characteristics of protein
   expression in moMΦ had been maintained during cultivation and that the
   proteome analysis reflected cell-type-specific characteristics.

3.2. Infection of moMΦ by ASFV

   As animal-to-animal variation is a major general caveat of experiments
   using primary cells, the reproducibility of host and viral protein
   expression after the infection of moMΦ was of prime importance. To test
   this, moMΦ from three different animals were cultured and infected, and
   the proteomes were analyzed 24 hpi. Prior to MS analysis, ASFV
   infection was confirmed via the expression of the dsRed reporter and
   immunoblot analysis. Additionally, sample homogeneity was verified via
   SDS PAGE of the proteins followed by Coomassie staining ([109]Figure
   S2).

   Protein expression levels were calculated for all three experiments and
   are compared in the scatterplots shown in [110]Figure 4. Although some
   degree of variation was observed, the mean expression levels of ASFV
   and host proteins correlated well between different animals with
   correlation coefficients ranging from 0.89 to 0.93 for viral proteins
   and 0.85 to 0.86 for host proteins. Together with the observed mean
   relative standard deviations of protein expression levels (viral
   proteins: 0.36 to 0.45; host proteins: 0.32 to 0.36) ([111]Table S2),
   the reproducibility of viral expression levels ([112]Figure S3) of
   biological replicates from different pigs, the good accordance of the
   proportion of viral proteins in relation to the whole cell proteome
   ([113]Figure 5A), and the high infection rates ([114]Figure S2a)
   indicated that moMΦ could be experimentally infected in a reproducible
   way.

Figure 4.

   [115]Figure 4
   [116]Open in a new tab

   Comparison of viral and host gene expression levels. Scatter plots of
   viral (upper row) and host (lower row) protein levels representing moMΦ
   preparations from different pigs and WSL cells (far right panels). Mean
   expression levels (%mol concentrations) of at least 3 biological
   replicates are shown after log10 transformation (data taken from
   [117]Supplementary Table ‘S7_MS-quantification.xlsx’). N = 3 for the
   comparison of individual pigs. In the right panels, the means of 9 moMΦ
   and 3 WSL samples are shown. The dissecting lines are shown as
   references in gray.