Host Gene Expression of Macrophages in Response to Feline Coronavirus Infection

Feline coronavirus is a highly contagious virus potentially resulting in feline infectious peritonitis (FIP), while the pathogenesis of FIP remains not well understood, particularly in the events leading to the disease. A predominant theory is that the pathogenic FIPV arises from a mutation, so that it could replicate not only in enterocytes of the intestines but also in monocytes, subsequently systemically transporting the virus. The immune status and genetics of affected cats certainly play an important role in the pathogenesis. Considering the importance of genetics and host immune responses in viral infections, the goal of this study was to elucidate host gene expression in macrophages using RNA sequencing. Macrophages from healthy male cats infected with FIPV 79-1146 ex vivo displayed a differential host gene expression. Despite the virus uptake, aligned viral reads did not increase from 2 to 17 h. The overlap of host gene expression among macrophages from different cats was limited, even though viral transcripts were detected in the cells. Interestingly, some of the downregulated genes in all macrophages were involved in immune signaling, while some upregulated genes common for all cats were found to be inhibiting immune activation. Our results highlight individual host responses playing an important role, consistent with the fact that few cats develop feline infectious peritonitis despite a common presence of enteric FCoV.


Introduction
Feline coronavirus (FCoV) is a highly contagious virus that is distributed worldwide and is ubiquitous in virtually all cat populations, particularly in multi-cat environments, such as shelters and rescues [1][2][3][4]. FCoV exists as two pathotypes, feline enteric coronavirus (FECV) and feline infectious peritonitis virus (FIPV). The enteric virus, FECV, commonly causes an asymptomatic infection presenting with mild gastrointestinal signs, and can persist in certain individuals [5]. FECV is readily transmitted via the fecal-oral route; therefore, the prevalence of FECV infection is generally associated with the number and density of cats housed together [1,4,6].
Sporadically and unpredictably, the infection can turn pathogenic, in which case the FIPV infection is associated with the highly fatal, systemic immune-mediated disease, feline infectious peritonitis (FIP) [5]. In this form, the virus develops a 100% fatal syndrome with two possible presentations: A pyogranulomatous disease of the visceral serosa and omentum, with consequent cavitary effusions, termed the "wet" or "effusive" form; and the dry form with granulomatous

Animal Procedures
All animal procedures were conducted and approved under the guidelines of the IACUC of Western University of Health Sciences, protocol approval number R10/IACUC/017. Peripheral blood for transcriptome studies was taken from six healthy male, specific pathogen-free (SPF) cats residing in an existing colony at the University of California, Davis. 30-40 mL of blood, equivalent to 1% of body weight or less was collected in heparinized tubes. The ages of cats at the time of blood draw were five months up to two years (five cats), and four years (one cat).

Monocyte Isolation
Monocytes from peripheral blood were isolated as previously described for canine monocytes [28] with some modifications. Briefly, the gradient centrifugation steps occurred at 450× g without break at deceleration and the subsequent washes to remove platelets were performed at 200× g for a total of three washes. PBMCs were counted, resuspended in RPMI 1640, containing 10% of FBS, 1% of penicillin and streptomycin, and 1× of non-essential amino acids, and plated at 5 × 10 6 in 6-well plates. Non-adherent cells were removed after 24 h by vigorously washing with a culture medium and cells were infected the following day.

Viral Infection for Transcriptome
For host transcriptome studies, macrophages were infected with FIPV 79-1146 (ATCC VR2128). The viruses were incubated at a multiplicity of infection (MOI) of 2 in a serum-free OptiMEM (Gibco, Thermo Fisher Scientific, Waltham, MA, USA) for 1 h for virus attachment, washed with OptiMEM, and incubated with fresh supplemented RPMI1640 for an additional 2 or 17 h. Technical replicates for the control, 2 and 17 h for macrophages from each cat were plated and incubated with PBS or the virus, respectively. CRFK cells (including technical replicates) were also infected as the control at an MOI of 1 in OptiMEM, followed by incubation in a supplemented DMEM. Uninfected controls underwent the same process with PBS without the virus. After incubation, the cell culture medium was completely removed and 600 µL of TRIzol (Invitrogen, Thermo Fisher Scientific, Waltham, MA, USA) was added to each well, followed by RNA extraction with the ZymoResearch RNA kit (ZymoResearch, Irvine, CA) according to the manufacturer's instructions. The RNA quality was evaluated via the Bioanalyzer (Agilent, Santa Clara, CA) and sent (1 µg per sample) for mRNA sequencing to Novogene, Inc. (Sacramento, CA, USA).

Quality Control of RNA Sequence Data
RNA paired-end sequencing quality control was assessed through FastQC (www.bioinformatics. babraham.ac.uk/projects/fastqc). An average of 35 million paired reads were sequenced per sample. Both adapters and low-quality bases (QV < 20) were trimmed from the reads' extremities with Trimmomatic [29].

Alignment Against Reference Transcriptomes
Kallisto [30] was the algorithm of choice for performing the alignment of all paired reads against the whole Felis catus reference transcriptome (F. catus NCBI-RefSeq-9.0). An average of 87.5% of the total reads from each sample was mapped onto the cat's annotated transcriptome. Alternatively, we also attempted to retrieve viral reads for both macrophages and CRFK from the sequenced libraries using Kallisto to align reads against the 11 protein-coding genes from the feline coronavirus (FCoV). An average of 0.03% of the total reads per sample was aligned against the FCoV annotated transcripts. Viral read counts were normalized as fragments per kilobase per million (FPKM): [read_counts / (gene_length_in_kb × total_reads_in_sample)] × 1,000,000.

Differential Expression
Tables generated by Kallisto were used as input for differential expression (DE) analyses. Due to the unique host responses in the macrophage dataset, the NOISeq version 2.14.1 [31] (Ctrl, 2, and 17 h) was employed to assess differentially expressed genes (DEGs) from each cat from which the ex vivo infected macrophages were derived. NOISeq output tables contained DEGs for each comparison (2 h versus control and 17 h vs. control) per cat. vennCounts and vennDiagram functions from the limma R package [32] were used for combining DEGs from each cat. Datasets were submitted to a multidimensional scaling (MDS) analysis, with the plotMDS function from EdgeR, to identify distinct samples clustered in a two dimensions-reduction landscape prior to the start of DE analyses. All tools described above for differential expression were run within the R environment version 3.5.2.

Gene Enrichment Analyses Using Both GO Terms and KEGG Pathways
After generating a list of differentially expressed genes (DEGs), we used ClueGO [33] under the Cytoscape version 3.7.1 [34,35] for a gene enrichment analysis relying on the Felis catus annotation from both gene ontology (http://geneontology.org/) and KEGG pathways (https://www.genome.jp/ kegg/pathway.html) consortia. Both enrichment analyses adopted the Hypergeometric test along with the Benjamini and Hochberg p-value adjustment method. A 0.05 threshold was set for the latter.

Host Responses Are Unique to Individual Cats with Several Clusters Present
To analyze host responses to FIPV, mRNA from feline macrophages (n = 6), infected ex vivo, was isolated and processed for next-generation sequencing. A unique bioinformatic analysis of differentially expressed genes (NOISeq) was necessary to compare gene expression. Individual cats displayed very different patterns of gene activation, with macrophages from some cats showing a clear separation of uninfected control versus infected samples (#1 and 2), while macrophages from other cats (#5 and 6) did cluster closely together independent of infection status or infection time ( Figure 1). Cells 2020, 9, x FOR PEER REVIEW 4 of 20 limma R package [32] were used for combining DEGs from each cat. Datasets were submitted to a multidimensional scaling (MDS) analysis, with the plotMDS function from EdgeR, to identify distinct samples clustered in a two dimensions-reduction landscape prior to the start of DE analyses. All tools described above for differential expression were run within the R environment version 3.5.2.

Gene Enrichment Analyses Using Both GO Terms and KEGG Pathways
After generating a list of differentially expressed genes (DEGs), we used ClueGO [33] under the Cytoscape version 3.7.1 [34,35] for a gene enrichment analysis relying on the Felis catus annotation from both gene ontology (http://geneontology.org/) and KEGG pathways (https://www.genome.jp/kegg/pathway.html) consortia. Both enrichment analyses adopted the Hypergeometric test along with the Benjamini and Hochberg p-value adjustment method. A 0.05 threshold was set for the latter.

Host Responses Are Unique to Individual Cats with Several Clusters Present
To analyze host responses to FIPV, mRNA from feline macrophages (n = 6), infected ex vivo, was isolated and processed for next-generation sequencing. A unique bioinformatic analysis of differentially expressed genes (NOISeq) was necessary to compare gene expression. Individual cats displayed very different patterns of gene activation, with macrophages from some cats showing a clear separation of uninfected control versus infected samples (#1 and 2), while macrophages from other cats (#5 and 6) did cluster closely together independent of infection status or infection time ( Figure 1). Macrophages were derived from six different cats that are represented by distinct symbol shapes on the chart: Cat #1 (circle), cat #2 (square), cat #3 (asterisk), cat #4 (triangle), cat #5 (diamond), and cat #6 (cross symbol).

Analysis of the Presence of Viral RNA in Infected Macrophages and CRFK
Cultured macrophages were infected with the FIPV 79-1146 for 2 and 17 h before RNA was collected for the expression analysis of host genes and viral reads. The viral presence was very low in the samples as indicated by the aligned reads for most macrophage samples at 2 h. However, the presence in most samples did indicate that viral particles were taken up by the cells (Table 1). No viral reads were detected in macrophages from cat #3.

Analysis of the Presence of Viral RNA in Infected Macrophages and CRFK
Cultured macrophages were infected with the FIPV 79-1146 for 2 and 17 h before RNA was collected for the expression analysis of host genes and viral reads. The viral presence was very low in the samples as indicated by the aligned reads for most macrophage samples at 2 h. However, the presence in most samples did indicate that viral particles were taken up by the cells (Table 1). No viral reads were detected in macrophages from cat #3. Table 1. Viral RNA load per sample in both raw read count summation and normalized fragments per kilobase million (FPKM) average from all virus' genes combined, obtained in cellular extracts of either macrophages or CRFK cells infected with FIPV, at 2 and 17 h post-infection. Cats #1-6 were healthy cats from the UC Davis cat colony, ages five months (5M) to four years (4Y). In macrophages, there was a limited viral presence/uptake of FIPV at 2 h, or not detected (nd), and no evidence of significant replication at 17 h (cat #6 n/a (no sample available)). In CRFK cells, viral RNA increased several log-fold from 2 to 17 h. Comparing 2 to 17 h samples, there was no indication of viral amplification despite viral uptake, as no increase of viral reads was observed in macrophages at the later time point. The viral reads at 17 h were either similar or lower than the number of reads at 2 h. Macrophages from one cat (cat #3) did not show any viral presence at 2 h. In contrast to the macrophages exposed to the virus, RNA sequencing showed several log-fold increases of viral isolates in the CRFK cells used as replication controls from 2 to 17 h (Table 1). This indicated that very few viral particles entering CRFK cells result in a significant amplification of the virus. This is also associated with pronounced cytopathic effects, which was observed in CRFK, but not macrophages.

Host Transcriptome Analysis Shows Cat Specific Responses to FIPV
Macrophages from each of the individual cats did differentially express several hundred genes at 2 and 17 h, compared to uninfected macrophages, highlighting the unique host responses of the cells from individual cats. Cat #3 was excluded from the bioinformatic analysis due to the lack of viral RNA in the macrophages. Since no control sample was available for cat #4, no NOISeq individual analysis of 2 or 17 h samples of cat #4 was possible in comparison to its own uninfected control. At 2 h after infection, macrophages from cats #1, 2, 5, and 6 expressed 1787, 499, 455, and 608 differentially downregulated genes, while 1620, 673, 416, and 639 genes were differentially upregulated ( Figure 2).
It is of interest that, while the macrophages clearly show a differential regulation of genes, there are relatively few common DEGs expressed in all of the macrophages from the four cats that were compared. Only 10 genes were significantly upregulated in macrophages from all cats, while one gene was downregulated ( Figure 2). The gene significantly downregulated in all macrophages was ATPase phospholipid transporting 8B4 (ATP8B4), which is involved in the cation transport and biosynthesis of ATP. Among the 10 upregulated genes, several are involved in immune signaling of the viral infection, such as SMAD4, ATP binding cassette, HAUS8, FCH domain, ubiquitin peptidase 20, and heteronuclear Ribonuclear Protein C (hnRNP C). The remaining genes in this group are involved in biosynthesis and other non-immune cellular functions (Table 2). It is of interest that, while the macrophages clearly show a differential regulation of genes, there are relatively few common DEGs expressed in all of the macrophages from the four cats that were compared. Only 10 genes were significantly upregulated in macrophages from all cats, while one gene was downregulated ( Figure 2). The gene significantly downregulated in all macrophages was ATPase phospholipid transporting 8B4 (ATP8B4), which is involved in the cation transport and biosynthesis of ATP. Among the 10 upregulated genes, several are involved in immune signaling of the viral infection, such as SMAD4, ATP binding cassette, HAUS8, FCH domain, ubiquitin peptidase 20, and heteronuclear Ribonuclear Protein C (hnRNP C). The remaining genes in this group are involved in biosynthesis and other non-immune cellular functions (Table 2).   Monocyte isolations yielded different amounts of cells for each cat resulting in sufficient macrophages from only four cats to conduct the 17 h infection time point, with cat #3 not showing viral reads and thus not included in the analysis. Therefore, macrophage gene expression from cats #1, 2, and 5 was compared. At 17 h, macrophages from cats #1, 2, and 5 differentially downregulated 573, 1203, and 1452 genes, respectively, with 120 of the same genes downregulated in macrophages from all three cats ( Figure 2). The number for upregulation of genes was similar, with 582, 1407, and 1611 genes for macrophages from cats #1, 2, and 5, respectively. One hundred and thirty-two genes were differentially upregulated in macrophages from all three cats ( Figure 2). Downregulated immune genes at 17 h tumor necrosis factor (TNF) receptor superfamily 4, interleukin 10, transforming growth factor (TGF)-beta, signal transducer and transcription activator 3 (STAT3), and transcription factors interferon regulatory factor (IRF) 4, and SMAD family member 7 (Table A1). One hundred and thirty-four upregulated genes common for all cats included several other proteins involved in cell cycle and metabolism. Among these were several in the centromere or centrosomal category, nuclear body protein SP140, or 7-dehydrocholesterol reductase. There were only two classical cytokine genes among the upregulated, namely leukemia inhibitory factor (LIF), from the interleukin 6 cytokine family, and the transcription factor IRF 3 (Table A2).

Gene Enrichment for Pathway Analysis and Ontology
As indicated in Figure 2, at 2 h after infection, there was no significant overlap of DEGs common to all infected macrophages in order to allow gene enrichment analyses (both GO terms and KEGG pathways). There was more commonality of DEGs expressed in response to FIPV at 17 h when macrophage responses of individual cats were compared, therefore gene enrichment was performed. Overall, 120 genes were differentially downregulated in all cats in response to FIPV with one KEGG pathway enriched, "Arginine and proline metabolism" (Table 3).
Gene ontology for enriched genes downregulated at 17 h showed the terms "extracellular matrix binding" and "collagen binding" for the molecular function, and "regulation of endothelial cell proliferation", "regulation of epithelial cell proliferation", "positive regulation of vasculature development", "positive regulation of angiogenesis", and "regulation of peptidase activity" for biological processes (Table 4).
The Cytoscape network analysis (by clueGO) shows how several of these pathways interact with each other, yielding two networks ( Figure 3).

Discussion
While isolates of both FCoV serotypes I and II have been shown to be pathogenic, only serotype II has been shown to efficiently replicate in the cell culture, using CRFK or felis catus whole fetus (FCWF) cell lines. Therefore, serotype II strains have been the subject of expanded in vitro and ex vivo investigations, including those focusing on viral entry and replication in monocytes/macrophages. It remains to be satisfactorily explained how both of these serotypes can arise from FECV in individual cats, with a yet unidentified mutation potentially resulting in a differential replication in monocytes and subsequently, leading to de novo pathogenesis in the affected cats. It is, however, likely that host immune responses play a significant, but not sufficiently

Discussion
While isolates of both FCoV serotypes I and II have been shown to be pathogenic, only serotype II has been shown to efficiently replicate in the cell culture, using CRFK or felis catus whole fetus (FCWF) cell lines. Therefore, serotype II strains have been the subject of expanded in vitro and ex vivo investigations, including those focusing on viral entry and replication in monocytes/macrophages. It remains to be satisfactorily explained how both of these serotypes can arise from FECV in individual cats, with a yet unidentified mutation potentially resulting in a differential replication in monocytes and subsequently, leading to de novo pathogenesis in the affected cats. It is, however, likely that host immune responses play a significant, but not sufficiently elucidated, role in the pathogenesis, and the question remains: What are the responses by the macrophages when infected with FCoV?
The presence of the virus has been evaluated directly in studies in macrophages [8,36,37] using PCR or immunofluorescence. Most studies showing an increased replication of FIPV in vitro, particularly in comparison to FECV, use indirect methods such as TCID50 in CRFK cells incubated with virus-infected macrophage extracts or supernatants [8,37]. However, to our knowledge, no studies so far have verified replication in macrophages by more state-of-the-art sensitive techniques, such as RNA sequencing.
In this study, RNA sequencing and analysis of viral reads indicated that at 2 h, the virus is taken up by the macrophages, as evidenced by the presence of viral RNA. However, there was no significant increase in viral RNA at 17 h, a time close to peak replication of the virus, which is usually at 24 h. Instead of a significant increase in viral reads, which would be expected with viral amplification, a similar amount, or even decrease of the virus was observed in the samples. In contrast to this, just a limited uptake of the virus into CRFK cells at 2 h led to several log-fold of replication, indicated by several thousands to hundreds of thousands of viral transcripts at 17 h. While the samples were not enriched for the pathogen, but rather the host RNA, there is a striking difference between replication of the virus in the macrophages compared to CRFK cells. Since subtle differences of replication in macrophages might not be recognized and could be below the threshold of detection, further investigations are warranted.
Macrophages of one cat #3 did not show the presence of viral RNA at all. This might be a technical issue, such as a higher MOI needed for successful infection. It is also possible that this particular host was more resistant to infection, which has been shown before both in vivo and in vitro [38,39]. Macrophages are competent immune cells and it has been shown that in vitro or ex vivo infection of macrophages with other viruses is challenging even if viremia exists in vivo. For example, there is evidence that Marek's disease virus, a herpesvirus, is phagocytized by macrophages in vivo and then disseminated to infect T and B lymphocytes. In contrast, in vitro/ex vivo infection of macrophages has been shown to be difficult [40]. The investigators were successful when virally infected target cells were incubated with the macrophages, which then phagocytized the infected cells and therefore, took up the virus much more effectively. Similarly, it might be worth exploring the incubation of feline macrophages with infected epithelial target cells for studies requiring a higher ratio of macrophages positive for the viral antigen. In addition, methods such as RNA sequencing could be explored to better quantify differences between FIPV and FECV uptake, viral presence, and amplification in macrophages.
Gene expression of monocytes/macrophages following exposure to the feline coronavirus so far has not been investigated with next-generation sequencing. The PCR analysis of mRNA expression of individual cytokines and other immune-related proteins has been done in vitro and in vivo [19][20][21]41], but next-generation sequencing is poised to give a more complete picture of gene expression after infection. In particular, differences in responses to FIPV and FECV can be better defined.
Transcriptome analyses of FIPV-infected feline macrophages were done in this study at 2 h (early phase of infection) and 17 h (closer to the known peak of amplification of FCoV), and samples were positive for a limited presence of viral transcripts. However, no enrichment of pathways or gene ontology common for all cats was possible in the early phase after infection. The only gene downregulated in all analyzed samples at 2 h was ATPase phospholipid transporting 8B4 (ATP8B4). This protein is involved in phospholipid transport, but there are no studies to its involvement in viral infection and replication. From the 10 upregulated genes, several are involved in immune-related host responses (Table A1). SMAD4 is a transcription factor involved in immune signaling, particularly in TGF-beta signaling. TGF-beta is an anti-inflammatory or regulatory cytokine that has been demonstrated to act as a proviral factor in epithelial cells during an influenza infection [42]. Haus8 is involved in maintaining cellular spindle integrity, but appears to have a role in the RIG-1 like antiviral signaling pathway in a Sendai virus infection [43]. ATP binding cassette subfamily A member 1, alternatively named RNase L inhibitor, blocks ribonuclease L. The interferon-regulated 2-5A/RNase L pathway plays a major role in the antiviral innate immune response. Consequently, several viruses are known to inhibit this pathway, including HIV which also induces the RNase L inhibitor [44]. The FCH domain is a phosphoprotein, associated with a viral infection. It is linked to endocytosis in an avian influenza virus replication [45]. Heterogeneous nuclear ribonucleoprotein C associated with pre-mRNA processing, RNA metabolism, and transport is a host factor important in the replication of positive-strand RNA viruses [46,47]. Finally, ubiquitin-specific peptidase 20 is an enzyme that has been shown to negatively regulate NFkB in an HTLV infection [48]. Other genes in this group are involved in cellular processes not linked to viral replication. Taken together, we identified several genes that are downregulated in all macrophages infected with FIPV and are linked to antiviral signaling. However, further studies are needed to elucidate the role of these genes in viral host interactions.
Interestingly, several downregulated genes that did overlap among cats were involved in immune signaling, which might confirm that the immune responses of macrophages are negatively altered by FIPV. While a few of these genes are easily identifiable for involvement in antiviral or inflammatory responses, such as interferon regulatory factor 4 (IRF 4) or TGF beta, a further investigation shows that several of the transcripts do, in fact, play a role in responses to infection. For example, heterogeneous nuclear ribonucleoprotein A2/B1 (hNRP A2/B1), which is involved in RNA binding and trafficking, has previously been shown to bind to the NP of Avian Influenza [46]. Other proteins downregulated are purinergic receptors (P2Y), that are involved in antiviral responses, affecting cytokine responses and T-cell activation [49]. These receptors also contribute to the direct elimination of the virus by inhibiting their intracellular replication [50]. Intersectin 2 plays important role in the regulation of the adaptive immune response in viral infection [51]. Ubiquitin-like proteins have been shown to modify proteins, thus conferring functions related to programmed cell death, autophagy and regulation of the immune system [52], and IRF 4, which is critical to T-cell effector function [53].
In regards to the viral infection of macrophages, it would be expected to see a strong inflammatory response of the cell in response to the virus, such as upregulation of Toll-like receptor pathways, interferon type I signaling, etc. In our study, the absence of typical inflammatory signals is notable. While immune molecules IRF3 and LIF were upregulated, no classical virally activated signaling is discernable or enriched in KEGG pathways. LIF has not been well investigated in regards to viral infections, but it has been implicated in suppressing the replication of HIV [54]. IRF3 is necessary for IFN beta induction and SARS-CoV has been shown to block a step between the nuclear transport of IRF3 and its phosphorylation, which is necessary for an IFN induction, but not with the induction of IRF3 itself [55]. Our data is consistent with FCoV employing a similar strategy, however, further studies are needed. Some differentially upregulated genes common for all cats, after a macrophage infection, were found to be within unexpected cellular functions, such as centromere proteins and TPX2 that are involved in specific phases of the cell cycle. Recent studies have indicated that centrosomes, as well as spindle organization, are actually involved in responses to a viral infection [56][57][58]. On the other hand, some upregulated transcripts included the protein tyrosine phosphatase, which is involved in attenuating T-cell activation [59] or nuclear body protein SP140. The latter has been shown to act as important repressor of genes involved in the regulation of cytokine production, inflammatory response, and cell-cell adhesion [60].
The picture emerging from these differentially expressed genes is that genes involved in antiviral responses and immune activation are depressed by FIPV uptake, while other genes involved in cell cycle and proteins repressing immune responses are upregulated. Even without an active viral replication, these changes most likely influence the pathogenesis and might explain how the monocytes carrying the virus potentially affect other cells, specifically lymphocytes.

Conclusions
The FIPV exposure and uptake leads to a limited differential gene expression in feline macrophages that might affect cell function. It will be of importance to further investigate cellular responses to FCoV isolates, in order to understand virus interactions with the host macrophages. A comparison between infection with FIPV and FECV strains focusing on host responses might yield insights into the pathogenesis of the virus. On the other hand, individual cat responses may be found to be of significant relevance to the pathogenesis and the number of samples required for analyses might be very high. The problem remains that very few cells take up the virus, and thus it will be difficult to analyze the transcriptome of a low number of infected cells. However, with high sample numbers, it could still be possible to identify differential gene expression that can provide valuable insight into the pathogenesis. An alternative is to use bone marrow-derived macrophages for viral infection since these macrophages have been shown to support replication at a higher rate, especially at a high MOI. An attempt to increase the number of virus-positive cells by incubating them with infected target cells may also identify more infected macrophages. Another possible source of infected macrophages is to isolate and analyze cells directly from viremic cats for transcriptome studies, however, immunologically those cells are representative of an ongoing disease process, and gene expression probably would be very different than early in the pathogenesis. In any case, further analyzing the host responses with next-generation molecular techniques will increase our understanding of FIPV pathogenesis.

Conflicts of Interest:
The authors declare no conflict of interest.