Infectious pancreatic necrosis virus triggers antiviral immune response in rainbow trout red blood cells, despite not being infective

Background: Some fish viruses, such as piscine orthoreovirus and infectious salmon anemia virus, target red blood cells (RBCs), replicate inside them and induce an immune response. However, the roles of RBCs in the context of infectious pancreatic necrosis virus (IPNV) infection have not been studied yet. Methods: Ex vivo rainbow trout RBCs were obtained from peripheral blood, Ficoll purified and exposed to IPNV in order to analyze infectivity and immune response using RT-qPCR, immune fluorescence imaging, flow cytometry and western-blotting techniques. Results: IPNV could not infect RBCs; however, IPNV increased the expression of the INF1-related genes ifn-1, pkr and mx genes. Moreover, conditioned media from IPNV-exposed RBCs conferred protection against IPNV infection in CHSE-214 fish cell line. Conclusions: Despite not being infected, rainbow trout RBCs could respond to IPNV with increased expression of antiviral genes. Fish RBCs could be considered as mediators of the antiviral response and therefore targets of new strategies against fish viral infections. Further research is ongoing to completely understand the molecular mechanism that triggers this antiviral response in rainbow trout RBCs.


Introduction
Fish viral infections cause significant losses in aquaculture. Infectious pancreatic necrosis (IPN) is a highly contagious viral disease with a high impact on salmonid aquaculture industry. Infectious pancreatic necrosis virus (IPNV) is the causative agent of IPN and was the first fish virus isolated in cell culture 1 . IPNV outbreaks are usually related to high mortality rates in salmonid aquaculture, especially in young individuals 2,3 , highlighting the urgent necessity for the development of efficient strategies in vaccination. IPNV belongs to the Aquabirnavirus genus within the Birnaviridae family. Viruses of this family are nonenveloped particles with a double stranded RNA genome. This genome consists of two segments: the A segment contains the information to encode a protein that is post-translationally cleaved into VP2, VP3 and VP4 viral proteins; the B segment encodes the viral RNA polymerase VP1 4 . VP2 and VP3 are the major structural and immunogenic proteins, as they represent 64% of the total proteins of the virion 5 .
In contrast to mammals, fish, reptiles and avian red blood cells (RBCs) are nucleated. Typically, the role associated with RBCs has been the transport of O 2 to different tissues and gas exchange. However, a whole set of biological processes related to the immune response has been recently reported for nucleated RBCs from different species: recognition of pathogen associated molecular patterns 6,7 through expression of pattern recognition receptors, such as toll-like receptors (TLRs) 8 ; production of cytokine-like factors 7,9-11 ; phagocytosis 12 ; and formation of complement immune complexes 13 . Fish RBCs are known to be the target of some viruses, such as infectious salmon anemia virus (ISAV) 11 and piscine orthoreovirus (PRV) 14,15 . Furthermore, both viruses can induce immune responses in infected RBCs, characterized by the expression of genes related to IFN-1 (type I interferon) pathway. Besides, recently it has been shown that viral hemorrhagic septicemia virus (VHSV) halted replication in rainbow trout RBCs could induce cytokine production 16 .
In view of the aforementioned evidence, this study was aimed to evaluate the immune response of rainbow trout RBCs against IPNV, one of the most ubiquitous viral fish pathogens. To achieve this objective, we first analyzed the infectivity of IPNV in rainbow trout RBCs. Then, RBCs immune response was evaluated after ex vivo exposure to IPNV, by means of antiviral gene and protein expression analysis. Finally, we evaluated the ability of RBCs to confer protection against IPNV in CHSE-214 cells, which are susceptible to IPNV infection. To summarize, here we report the regulation of the immune response of rainbow trout RBCs by IPNV, a non-infective virus in this cell type. This immune response was characterized by the expression of genes related to the IFN-1 pathway, Mx production and induction of an antiviral state to IPNV in CHSE-214 cells.

Animals
Rainbow trout (Oncorhynchus mykiss) individuals of approximately 10 g were obtained from a commercial fish farm (PISZOLLA S.L., CIMBALLA FISH FARM, Zaragoza, Spain). Fish were maintained at the University Miguel Hernandez (UMH) facilities with a re-circulating dechlorinated-water system, at a stocking density of 1 fish/3L, at 14°C, and fed daily with a commercial diet (SKRETTING, Burgos, Spain). Fish were acclimatized to laboratory conditions over 2 weeks before experimentation. The number of fish used is indicated for each experiment/figure.

Amendments from Version 1
In this new version of the manuscript we have included the corrections of Dr. Rimstad and Dr. Bols along the manuscript, in order to improve manuscript understanding. The line of the manuscript has not changed, but the manuscript comprehension has been improved. We have corrected several mistyping and linguistic errors, as well as clarifying some points in the manuscript and eliminating some confusing citations. Therefore, some references have been eliminated and one reference previously unpublished is now updated. Besides, as Dr. Bols suggested, in Figure 3A, FFU/ml has been removed from the y axis and only mentioned in the legend of the figure.

REVISED
In time course experiments, the initial supernatant with IPNV was not removed. When each of the time points was reached, RBCs were washed with cell culture medium and CHSE-214 cells with PBS supplemented with calcium.

Viral titration assay
The virus titer in IPNV-exposed RBCs supernatants was quantified by TCID 50 and by RT-qPCR. Briefly, different dilutions of the supernatants (from 10 -1 to 10 -4 ) were added to CHSE-214 cell monolayers, and incubated at 14°C for 90 minutes. Then, the virus was removed and infected CHSE-214 cell monolayers covered with a solution of RPMI 2% FBS. Cell plates were incubated at 14°C for 7 days. For RT-qPCR titration, 30 µL of IPNV with known titer (10 9 TCID 50 /mL) and 30 µL of IPNV-exposed RBCs supernatants were used to extract RNA and synthetize cDNA, as explained hereafter. Ten-fold serial dilutions from 10 8 to 10 2 TCID 50 /mL were done to obtain IPNV cDNA and create a standard line. Gene expression by RT-qPCR cDNA was synthesized from RNA using M-MLV reverse transcriptase (Invitrogen, Thermo Fischer Scientific Inc.), as previously described 19 . Final concentration of cDNA was 6 ng/µL. RT-qPCR reactions were performed in a total volume of 20 µl using 12 ng of cDNA, 10 µl of TaqMan universal PCR master mix (Thermo Fischer Scientific), 900 nM final concentration of each primer (300 nM for IPNV segment A) and 300 nM of probe (150 nM for IPNV segment A). RT-qPCR was performed using the ABI PRISM 7300 System (Thermo Fischer Scientific). Cycling conditions were 50°C for 2 min and 95°C for 10 min, followed by 40 cycles at 95°C for 15 s and 60°C for 1 min.

RNA isolation and
Gene expression was analyzed by the 2 -ΔΔ Ct method 20 . The eukaryotic 18S rRNA gene (Cat#4310893E, Thermo Fischer Scientific) was used as endogenous control. Primers and probes are listed in Table 1.

Antibodies
Several antibodies were used to stain cells for cytokines and to measure polypeptides in RBCs extracts by western blotting. They are briefly described below and their Research Resource Identifiers (RRIDs) given. For intracellular staining, mouse polyclonal antibodies against rainbow trout IL1β (RRID: AB_2716269) 21,22 , IL8 (RRID: AB_2716272) 23 and TNF-α (RRID: AB_2716270) 24 were produced at the laboratory of Dr. Luis Mercado. Rabbit polyclonal antibody against rainbow trout Mx3 (RRID: ABA_2716267) 25,26 was produced at the laboratory of Dr. Amparo Estepa. Anti-IPNV-VP3 monoclonal antibody 2F12 (RRID: AB_2716296) was used for IPNV labelling 27 . Anti-rabbit IgG (H+L) CF™ 488 antibody produced in goat and anti-mouse IgG (H+L) CF™ 488 antibody produced in goat were used as secondary antibodies for proteins and anti-mouse IgG (H+L) CF™ 647 produced in goat to detect 2F12 antibody.

Western blot
Control and IPNV-exposed RBCs pellets (≈10 7 cells) were used for protein extraction. Cell pellets were washed 3 times with PBS and then resuspended in 30 µl of PBS with a cocktail of protease inhibitors (Sigma-Aldrich). Then, cells were frozen/ thawed 3 times and lysed using an eppendorf micropistile (Eppendorf, Hamburg, Germany). Samples were loaded in Tris-Glycine sodium dodecyl sulfate 12% polyacrylamide gels under reducing conditions. Electrophoresis was performed at 200 V for 60 min. For blotting, the proteins in the gel were transferred for 80 min at 100 V in transfer buffer (2.5 mM Tris, 9 mM glycine, 20% methanol) to nitrocellulose membranes (BioRad, Madrid, Spain). Then, membranes were blocked with 8% dry milk and 1% Tween-20 in PBS and incubated with eIF2α-P or α-Actin antibodies, at the recommended dilutions in PBS containing 0.5% dry milk and 0.5% Tween-20 at 4°C and overnight. Incubation with secondary antibody GAR-Po (Sigma-Aldrich) was done in 0.5% milk 0.5% Tween-20 in PBS for 45 min. Membranes were washed 3 times with PBS containing 1% dry milk 0.5% Tween-20 for 15 min after every antibody incubation. Finally, the membrane was washed 3 times with PBS before analysis of the peroxidase activity. Peroxidase activity was detected using ECL chemiluminescence reagents (Amersham Biosciences, Buckinghamshire, UK) and revealed by exposure to X-ray. Protein bands from western blotting were analysed by densitometry using the Scion Image 4.0.2 Software (RRID: SCR_008673) (www.scionorg.com).

Antiviral activity of conditioned medium
Conditioned medium (CM) was obtained from control and IPNV-exposed RBCs at MOI 0.5, during 3 days. The CMs were clarified at 1600 rpm for 5 min. IPNV titer in the supernatants of IPNV-exposed RBCs resulted in 10 TCID 50 /mL or less, therefore viral presence in the supernatants was obviated.
To test the antiviral activity of the CM, confluent CHSE-214 cells (7.8×10 4 cells/well), seeded in 96 well plates, were pre-treated with 100 µL of each supernatant at the indicated dilutions for 24 hours. After that, CHSE-214 cells were infected, as described previously, with IPNV at MOI 0.05, for 24 hours. Finally, intracellular staining of IPNV foci was carried out.
Intracellular staining of IPNV foci CHSE-214 cells were fixed with PFA diluted at 4% in PBS followed by a second fixation with cold methanol. Each fixation step lasted 15 minutes. Cells were washed with PBS after each fixation step. Blocking buffer containing 5% goat serum (Sigma-Aldrich) and 0.3% Triton X-100 (Sigma-Aldrich) was added to each well with the cells for 1 hour. Then, anti-VP3 2F12 antibody was diluted 1/500 in antibody dilution buffer (1% BSA (Sigma-Aldrich), 0.3% Triton X-100) and was incubated for 1 hour. FITC-labelled goat anti-rabbit was used as secondary antibody at 1/300 dilution. Cells were washed three times after each antibody incubation with PBS. IF images were taken INCell Analyzer 6000 imaging system. IN Cell Developer Toolbox 1.9.2 (RRID: SCR_015790; GE Healthcare, Little Chalfont, UK) was used to count number of IPNV foci (positive areas after image segmentation were selected when >21000 fluorescence units and >2500 µm 2 criteria was reached).
MTT assays Cell viability was tested using MTT (3-(4,5-dimethylthiazol-2-yl)-2,5-diphenyltetrazolium bromide) colorimetric assay 32 . Briefly, 25 µl of MTT at a final concentration of 1.9 mg/mL were added to previously treated CHSE-214 cells monolayers, seeded in 96 well plates. Cells were incubated for 3 hours at 21°C with the reagent. Then, the medium was removed from the wells. Formazan crystals were dissolved in 100 µl of 100% DMSO, incubated for 30 minutes. Absorbance was read at 570 nm in the EON TM microplate spectrophotometer (Biotek, Winooski, VT). Percentage of viable cells was calculated as follows: absorbance treated cells/absorbance non-treated cells) x100.

Ethics approval
Methodology was carried out in accordance with the Spanish Royal Decree RD 53/2013 and EU Directive 2010/63/EU for animals used in research experimentation. All experimental protocols involving animal handling were also reviewed and approved by the Animal Welfare Body and the Research Ethics Committee at the Miguel Hernandez University (approval number 2014.205.E.OEP; 2016.221.E.OEP) and performed by qualified research personnel.

IPNV did not infect rainbow trout RBCs
To evaluate the infectivity of IPNV in rainbow trout RBCs, RBCs were exposed to IPNV at MOI 0.5 and the viral RNA was evaluated by RT-qPCR in the cell pellet at different times post-exposure. IPNV infectivity was also evaluated in parallel in the CHSE-214 cell line, used as a positive control of infection. IPNV segment A (IPNV-A) RNA levels inside RBCs and CHSE-214 cell line were similar at 1 and 3 hours post-exposure (hpe) ( Figure 1A). After 6 hpe, IPNV-A RNA level was 3 logarithms lower in RBCs in comparison with CHSE-214 cells. On the other hand, the titer of IPNV in the supernatants from IPNV-exposed RBCs at a MOI of 0.5 and 5, was evaluated by TCID 50 , at 3 days post-exposure (dpe), and showed a recovered titer of 5 and 4 logarithms lower, respectively ( Figure 1B). Furthermore, the supernatants titrated by RT-qPCR, were below the lowest limit of detection 10 2 TCID 50 (Table 2). Moreover, FC analysis of control and IPNV-exposed RBCs for IPNV VP3 protein did not show significant differences ( Figure 1C and D). Therefore, IPNV did not infect rainbow trout RBCs.

IPNV exposure increased the expression of interferonrelated antiviral genes and proteins in rainbow trout RBCs
To determine if IPNV would induce an antiviral response in RBCs, RT-qPCR analysis of IFN-related antiviral genes was performed for IPNV-exposed RBCs. The results showed that mx1-3 and pkr genes were significantly expressed at 72 hpe. On the other hand, ifn1 gene presented a tendency to increase its expression after 6 hpe, having a peak at 24 hpe. Also, tlr3 gene expression tended to be upregulated at 24 hpe, whereas irf7 expression was upregulated at 72 hpe ( Figure 2A). Three and six dpe with IPNV, RBCs were stained intracellularly with an anti-Mx antibody and analyzed by FC and immunofluorescence imaging (IF). The results showed a significant increment in the expression of Mx protein at 6 dpe by both FC an IF ( Figure 2B and D). FC histograms showed, at 6 dpe, that RBCs depicted distinct peaks of Mx expression, showing that the expression of Mx in RBCs was heterogeneous in the total RBCs population ( Figure 2C).

Conditioned medium from IPNV-exposed RBCs protected CHSE-214 cells against IPNV infection
To analyze if IPNV-exposed RBCs could secrete factors that were capable to protect other fish cells against IPNV infection, conditioned medium (CM) from control and IPNV-exposed RBCs (with IPNV titer <10 TCID 50 /mL) were added to CHSE-214 cells prior to infection. Figure 3A shows a significant decrease in the number of IPNV infective focus forming units Recovered virus titer in supernatants from IPNV-exposed RBCs with an inoculum titer of 10 6 (MOI 0.5) and 10 7 (MOI 5) TCID 50 /mL obtained after 72 hpe (n = 5). Data is represented as mean±SD. Mann-Whitney test was performed among both conditions (*, p-value < 0.05). (C) MFI (mean fluorescence intensity) of viral protein VP3 in control and IPNV-exposed RBCs at MOI 0.5 and 3 dpe (n = 6) Mann-Whitney test was performed among both conditions. (D) Representative flow cytometry histograms of IPNV VP3 protein detection in control and IPNV-exposed RBCs at MOI 0.5 and 3 dpi.

Sample
Ct value ± SD   . Antiviral activity of the conditioned media from IPNV-exposed RBCs. (A) Viral titers (FFU/mL) in CHSE-214 cells infected with IPNV at MOI 0.05 previously non-treated (black) or treated with either supernatants from control RBCs (white) or IPNVexposed RBCs (grey), during 24 hours, at the indicated dilutions (n = 4, performing triplicates from each individual). Two-way ANOVA, with Sidak´s multiple comparison test, was performed among the different dilutions and conditions to test statistical differences. (B) Percentage of viable CHSE-214 cells pre-treated with conditioned medium from control and IPNV-exposed RBCs, during 24 hours, and relative to non-treated CHSE-214 cells. Percentage of viable cells was calculated as follows: absorbance treated cells/absorbance non-treated cells) x100. . Mann-Whitney test was performed among both conditions. (B) Phosphorylation of translation initiation factor eIF2α in IPNV-exposed RBCs. Representative western blot of eIF2α-P in control and IPNV-exposed RBCs from two individuals at MOI 0.5, 3 dpe. Densitometry ratios were done relativizing to α-actin. Mann-Whitney test was performed among both conditions.
(FFU/mL) when pre-treating with 1/5 diluted CM from IPNV-exposed RBCs. CHSE-214 cells viability, by means of an MTT colorimetric assay, was not affected by the exposure to CM ( Figure 3B).
IPNV exposure decreased the expression of cytokines in rainbow trout RBCs To evaluate whether ex vivo rainbow trout RBCs could produce cytokines in response to IPNV exposure, RBCs were exposed to IPNV and IL-1 β, IL-8 and TNF-α protein levels were evaluated by means of FC and IF in control and IPNVexposed cell cultures. The results showed a decrease in the protein expression of IL-1β, IL-8 and TNFα in IPNV-exposed RBCs ( Figure 4A).
IPNV exposure did not induce phosphorylation of the α-subunit of the eukaryotic translational initiation factor 2 (eIF2α) in rainbow trout RBCs The phosphorylation of the translation initiation factor eIF2α is a key mechanism of global inhibition of translational initiation 33 and it has been described to happen after IPNV infection in the permissive cell line CHSE-214 cells 34 . In this sense, since IPNV-exposed RBCs depicted a small downregulation of the evaluated cytokines protein levels, we further investigated whether IPNV exposure could reduce protein translation in RBCs by triggering the phosphorylation of eIF2α. However, the results revealed no changes in the phosphorylation of eIF2α ( Figure 4B). Related uncropped blots are included.

Discussion
Previously, we have demonstrated that rainbow trout RBCs can respond to VHSV, a ssRNA virus not targeting RBCs, halting its replication, downregulating type I interferonrelated genes, showing global protein downregulation in the cell and phosphorylation of the translation initiation factor eIF2α 16 .
It is known that IPNV primarily targets pancreatic and liver cells 35 . It has been also reported that IPNV was detectable in kidney hematopoietic tissue, corpuscles of Stannius, in Islets of Langherhans, in the lamina propria of the pyloric caeca, the gill arch connective tissue, the ventricle of the heart and dermis of the skin 35 . Our results showed that IPNV did not replicate in RBCs, although small amounts of IPNV were persistently found inside RBCs after 3 dpe (≈ 10 3 TCID 50 /mL). Similarly, IPNV has been shown to enter mammalian cells, without significant levels of replication, being this entry suggested to be receptor mediated 36 . From our results, the persistence of IPNV in RBCs after 72 hpe could point out the entry of the virus inside RBCs. However, we could not further verify the presence of the IPNV inside RBCs ( Figure 1).
Nevertheless, despite the lack of replication of IPNV in RBCs, IPNV could induce an antiviral gene expression mediated by the IFN pathway, as it has been observed in RBCs productive infections with ISAV 11 and PRV 14 . As shown by our results, ifn1 and IFN-1 related genes (irf7, pkr and mx) expression levels were increased time-dependently in response to IPNV-exposure. High inter-individual variability was observed, similarly to that found in the RBCs from salmons challenged with PRV 37 . In addition, although we could not verify the entry and uncoating of IPNV inside RBCs, we could observe an increment in the expression of the tlr3 gene in parallel to the expression of the other IFN-related genes in IPNV-exposed RBCs. This could indicate the activation of the TLR3/IFN pathway by the presence of intracellular viral dsRNA.
IFN-1 leads to the expression of interferon stimulated genes (ISGs) 38 . Among ISGs, the antiviral protein Mx has a well characterized antiviral role. Confirming those expectations, our results showed the significant upregulation of the Mx protein 6 dpe, after having a peak of its gene expression at 3 dpe.
Previously, a positive correlation between the expression of Mx protein and the inhibition of IPNV in CHSE-214 cells has been established 39 . Therefore, Mx protein production in IPNV-exposed RBCs could be involved in the low IPNV titers observed. The high basal levels of Mx protein detected inside RBCs ( Figure 2D), much elevated than those for CHSE-214 cells ( Figure S1), could be implicated in the early disappearance of IPNV inside RBCs. A similar hypothesis has been made in the abortive infection of VHSV in the RTS-11 cell line 40 and in rainbow trout RBCs 16 , where upregulation or high constitutive expression of mx genes was speculated to be related to the inhibition of the virus.
Moreover, our results showed that CM from RBCs exposed to IPNV could partially protect CHSE-214 cells from IPNV infection. Similar to other cell types, this antiviral activity has been also observed in CM of RTS11 and RTG-2 cells exposed to Poly (I:C) (polyinosinic:polycytidylic acid) and/or infected with chum salmon reovirus 41 . The fact that RBCs can secrete factors that confer protection against IPNV infection in other cell lines could indicate that RBCs, despite not being permissive to IPNV infection, may exhibit an antiviral response. Besides, we evaluated the production of cytokines in IPNVexposed RBCs. Previously, the expression of IL-1β in salmon gill and head kidney tissues 42 , IL-8 in rainbow trout head kidney tissue 43 and TNFα in zebrafish embryonic cells 44 have been implicated in the immune response against IPNV; therefore, we chose these cytokines to evaluate the immune response of rainbow trout RBCs to IPNV exposure. However, our results showed a reduction trend of these proteins in IPNV-exposed RBCs.
A shutdown in protein synthesis by phosphorylation of eIF2α has been reported in CHSE-214 cells infected with IPNV 34 . So far, in rainbow trout RBCs exposed to IPNV, although a trend to cytokine protein reduction was observed, no phosphorylation of eIF2α was detected and Mx protein expression was increased. IFN-1 has been reported to inhibit the production of IL-1β 45 , therefore, the cytokine reduction trend observed could have been a result of the related IFN-1 pathway upregulation. In contrast, in rainbow trout RBCs, VHSV rhabdovirus induced phosphorylation of eIF2α and a cell shut-off characterized by the downregulation of the proteome 16 .
Further studies are needed to completely understand the molecular mechanism through which IPNV triggers this immune response in rainbow trout RBCs. However, the lack of commercial antibodies against fish proteins involved in cell signaling networks limits the study of this area. The implication of RBCs during in vivo IPNV infection and the response against different strains of IPNV remains to be evaluated.
Finally, one of the potential applications of these results is that fish RBCs could be considered mediators of the antiviral response and therefore targets of novel DNA vaccines and of new strategies against fish viral infections.

Competing interests
No competing interests were disclosed.

Grant information
This work was supported by the European Research Council (ERC starting grant 2014 GA639249).
The funders had no role in study design, data collection and analysis, decision to publish, or preparation of the manuscript.

Acknowledgments
Special thanks to Remedios Torres and Efren Lucas for their technical assistance.

Figure S1. qPCR Amplification plot of mx1-3 gene expression in RBCs and CHSE-214 cells.
Click here to access the data. The authors have responded well to each individual point raised. We would recommend that fish populations used in experimental infections including immunological studies are routinely screened for pathogens before entering experimental infections, and not just assessed by measuring mortality. The lack of mortality may not warrant freedom of infection. The presence of low virulent variants of virus may be missed, but they could still have impact on parameters used for measuring immune responses.

Nombela I, Carrion
No competing interests were disclosed.

Competing Interests:
We have read this submission. We believe that we have an appropriate level of expertise to confirm that it is of an acceptable scientific standard. The manuscript by Maria de Mar Ortega-Villaizan Romo provides evidence that trout red blood cells can express antiviral mechanisms. The discovery is very intriguing, as red blood cells appear to have addition functions beyond carry gases, and this might open up new ways to enhance fish health. Therefore, I recommend indexing. As documented below, I do have many minor suggestions for improving the clarity and aesthetics of the manuscript.

Background
Remove "highly" I suggest "However, the roles of RBCs in birnavirus…." Methods: Remove "induced" Remove "induced" : Results Remove "-exposed RBCs" : I suggest recasting a few sentences as shown below. Conclusions "Despite not being infected, trout RBCs respond to IPNV with increased expression of antiviral genes." In the last sentence, I suggest "that triggers this antiviral response in trout RBCs"

INTRODUCTION
2 line remove "Among all of the viral diseases," 1 paragraph It sounds awkward and the point being made is made again in the 4 paragraph.

paragraph
Remove "This implies that nucleated RBCs can support viral replication." Over 40 years ago, enucleated cells were shown to support replication of some viruses . Therefore, having a nucleus in RBCs might have nothing to do with their capacity to support the replication of some viruses. For example, human RBCs might also have roles in viral infections. Therefore, the first 2 sentences of this paragraph could be eliminated and the next sentence, after removing "for example", would become the topic sentence.

paragraph
Remove the sentence that begins "Recently, it has been shown that viral hemorrhagic septicemia virus (VHSV) can induce …." I don't see how it fits the subject of the paragraph or point out in the sentence that RBCs as well as stromal cells were studied. 4 paragraph recast the last 2 sentences "However, we report that even though not infecting RBCs, IPNV induces in them the expression of antiviral genes in the INF-1 pathway." I am not sure that I would emphasize so strongly "protection against IPNV infection in CHSE-214 cells". This appears to be based on the 1/5 bar in Figure 3A. I would like to have seen if protection also arose from 1 to 3 supernatant dilution or 1 to 10 supernatant dilution. If either one or both of these dilutions did protect, it would strongly indicate that protection can be repeatability demonstrated.

Animals
Space between 1 and fish. Space after Spain)

RBC purification
Was no anticoagulant used in the collection of the blood? Please clarify. Would thrombocytes be present among the RBCs?
In the 3 last sentence I wonder if something should follow the 2%. RPMI with 2 % FBS? This also comes in the next section. Maybe define it here as RPMI with 2 % FBS (RPMI/FBS) and use in RPMI/FBS in subsequent section.

Viral infection assays
No need to repeat here as the genus species name are mentioned earlier.

Oncorhynchus mykiss
In the 2 last sentence, I believe that the RPMI was "removed" rather than "retired". Were RBC infected at 14 °C like CHSE-214 were?

IPNV exposure increased the expression of interferon-related antiviral immune genes and proteins in RBCs
The first sentence should be shortened as follows "to determine if IPNV would induce an antiviral response in RBCs." Are there antiviral genes that are not immune genes?
In other words, can the description be simplified to just 'antivrial genes"?

paragraph
The last two sentences could refer to the data of the ms for support.

paragraph
I wonder if the authors could briefly discuss the apparent constitutive expression of Mx in RBCs as seen in Panel D of Figure 2. Also could Mx be released from RBCs? In cattle Mx1 is found in exosomes (see Racicot K et al., 2012, Am J Reprod Immunol) and exosomes can be released by RBCs (see Danesh A et al., 2014, Blood).

Figure 1
In panel A write out hpe on the X axis. Could FFU be written out here in the legend so figure stands alone? The legend mentions doing a two-way ANOVA but no mention of a post test is given, although the 1/5 supernatant dilution is identified with an asterisk? Figure 4 Title legend has a spelling mistake and is a bit misleading because panel B is not covered by the current title. I suggest the following title. "Reduction in RBC cytokine levels by IPNV and a possible mechanism" The legend should give more information so the figure can better stand by itself. Write out MFI (Mean Fluorescence Intensity) and FC (Flow Cytometry). At two places in the legend, mention is made of the Mann-Whitney test but the outcome is not clearly stated.

If applicable, is the statistical analysis and its interpretation appropriate? Yes
Are all the source data underlying the results available to ensure full reproducibility? Yes

Are the conclusions drawn adequately supported by the results? Yes
No competing interests were disclosed. Author´s response: It has been corrected.

paragraph
Remove "This implies that nucleated RBCs can support viral replication." Over 40 years ago, enucleated cells were shown to support replication of some viruses . Therefore, having a nucleus in RBCs might have nothing to do with their capacity to support the replication of some viruses. For example, human RBCs might also have roles in viral infections. Therefore, the first 2 sentences of this paragraph could be eliminated and the next sentence, after removing "for example", would become the topic sentence. Author´s response: We have removed these sentences.

paragraph
Remove the sentence that begins "Recently, it has been shown that viral hemorrhagic septicemia virus (VHSV) can induce …." I don't see how it fits the subject of the paragraph or point out in the sentence that RBCs as well as stromal cells were studied. Author´s response: We have changed this sentence to: "Besides, recently it has been shown that viral hemorrhagic septicemia virus (VHSV) halted replication in rainbow trout RBCs could induce cytokine production" as another reference of immune response by RBCs against virus. 4 paragraph recast the last 2 sentences "However, we report that even though not infecting RBCs, IPNV induces in them the expression of antiviral genes in the INF-1 pathway." I am not sure that I would emphasize so strongly "protection against IPNV infection in CHSE-214 cells". This appears to be based on the 1/5 bar in Figure 3A. I would like to have seen if protection also arose from 1 to 3 supernatant dilution or 1 to 10 supernatant dilution. If either one or both of these dilutions did protect, it would strongly indicate that protection can be repeatability demonstrated. Author´s response: The protection conferred by the CM from IPNV-exposed RBCs diminished as the dilution increased from 1/5 to 1/25 and 1/125. However, we have not tried 1/3 or 1/10 dilutions. As Dr. Bols pointed out, it would have been interesting to check with lower dilution ranges. In relation to repeatability, this assay was done using the supernatants from 4 ex vivo infections, in triplicate each of them. Therefore, the repeatability could have been demonstrated by the assay itself?
We have rewritten the sentence to do not emphasize so strongly "protection" as follows: "the induction of an antiviral state to IPNV in CHSE-214 cells".

Animals
Space between 1 and fish. Space after Spain) Author´s response: It has been corrected.

RBC purification
Was no anticoagulant used in the collection of the blood? Please clarify. Would thrombocytes be present among the RBCs?
Author´s response: We did not use any anti-coagulant. The blood was immediately diluted in RPMI into the uterus and suggest involvement in basic cellular processes independent of its role during viral infections". Therefore, although it is an interesting topic of investigation we would rather not to mention the possible link between Mx and RBCs exosomes in this manuscript since it is unknown whether it could be related to Mx antiviral function.

Figure 1
In panel A write out hpe on the X axis. Author´s response: It has been written

Figure 3
Could FFU be written out here in the legend so figure stands alone? The legend mentions doing a two-way ANOVA but no mention of a post test is given, although the 1/5 supernatant dilution is identified with an asterisk? Author´s response: Yes, FFU/mL has been deleted from panel A, x axis, as it is indicated in the figure legend. On the other hand, Sidak´s multiple comparison test was performed and has been added to the legend. Figure 4 Title legend has a spelling mistake and is a bit misleading because panel B is not covered by the current title. I suggest the following title. "Reduction in RBC cytokine levels by IPNV and a possible mechanism" The legend should give more information so the figure can better stand by itself. Write out MFI (Mean Fluorescence Intensity) and FC (Flow Cytometry). At two places in the legend, mention is made of the Mann-Whitney test but the outcome is not clearly stated. Author´s response: Title of Figure 4 has been changed as follows: IPNV-exposure decreased cytokine levels in rainbow trout RBCs, and MFI and FC explained. In relation to the statistical analysis, Mann-Whitney test was performed for Fig4A and 4b, however, statistically significant differences were not observed. This is an interesting paper showing that purified rainbow trout RBC exposed to IPNV do not get infected This is an interesting paper showing that purified rainbow trout RBC exposed to IPNV do not get infected by the virus, but nevertheless raise an innate antiviral response. The latter is shown by the induction of IFN and a few IFN related genes and that conditioned media from IPNV exposed RBC inhibits IPNV infection of the susceptibe cell line CHSE-214. Many methods are used to approach the hypotheses and the results appear provide new information on the very intriguing role of fish red blood cells in interaction with viruses, in particular by showing that the cells can induce an antiviral immune response without being infected.
The main comments are related to how the authors have interpreted the results far beyond what they have shown, by drawing links to adaptive immune mechanisms and vaccination approaches, while they could have discussed antiviral protection mechanisms in further detail and with more scientific basis.
There are also some missing information regarding the experimental details, and far too extensive use of an unpublished paper in the discussion of the data.

Comments:
The title is generalizing beyond the scope of the paper, and should specify the virus and the host studied (IPNV and onchorhynchus mykiss). Abstract and Discussion: The statement that antiviral responses in RBCs could be targets for DNA vaccines is not linked to the data presented here. A vaccination effect requires antigen presentation and adaptive immune activation for long term protection, which is not touched upon at all in this paper. If this point should be made, the authors should in the discussion provide some explanation and scientific basis for how they think antiviral responses in RBCs to an uninfecting virus make them promising DNA vaccination targets, and how directing DNA vaccines to RBC should be performed. Introduction/Discussion: An unpublished paper: Nombela I, Puente-Marin S et al. is referred to in introduction and several times in the discussion. The content of this paper cannot be evaluated. The authors should primarily use published work in their discussion. Introduction. There are many language mistakes in the introduction (and elsewhere). Please have the manuscript corrected linguistically. Trout is not a specific annotation. Please specify that the work is performed for rainbow trout. Diseases are not agents. Diseases cause losses not loses. Viral titer not titters. genus not genera, it Aquabirnavirus Aquabirnaviridae is wrong suffix and it is not plural. etc. The role of IPNV -VP3 is not completely unknown, there are several papers describing its function. "The VP2 epitopes are the target for antibodies against different serotypes." Not but against distinguishing between. It should be noted that the serotyping system is based upon the use of rabbit antisera, it is not based on the immune response from fish. No observations of viral factories has been observed for ISAV. "The fact that these cells are nucleated implies that fish RBCs can respond to different stimulus Not really, having a nucleus doesn't really prove that the cell can and control cellular processes". react various stimuli, that have to shown. Materials and Methods: Number of fish used not given. Unclear if the fish used were tested for previous viral infections. 18S is given as the reference gene target in the text, but in the table EF1a primers are given. Results is not clear on what was used. Results: Unclear if the "n" in experiments is number of fish or parallel ex vivo infections. "IPNV segment A (IPNV-A) RNA levels inside RBCs and CHSE-214 cell line were similar at 1 and 3 hours post-exposure (hpe) ( Figure 1A). After 3 hpe, IPNV-A RNA level was 4 logarithms lower in RBCs in comparison with CHSE-214 cells." Do you mean from 6 hpe and not after 3 hpe? It is not 4 logs difference according to Fig 1 it is 3 log at 6 hpe and 5 log at 26 hpe. It is nevenr 4 logs? What is the detection limit for VP3 in flow cytometri? Discussion: 15. 1.

2.
Discussion: The small trend towards a difference in cytokine levels Fig 4A should not be discussed as "downregulation", and cannot be interpreted as a cell shut-off of the proteome by the infection. Most likely the slight trend towards down-regulation is just an effect of the relative upregulation of the IFN-I pathway (ISGs) Authors say in the discussion regarding the protective effect of conditioned media: "To elucidate which factors could be implicated in this protection, we evaluated the production " It is well known that IFN-I, and not to the same degree of cytokines in IPNV-exposed RBCs cytokines, confers a direct antiviral response on surrounding cells. The upregulation of IFN-I is already shown at the transcriptional level, and the upregulation of ISGs directly indicates IFN activity. The main role of cytokines is to attract and activate immune cells and mobilize an immune response in tissue in vivo. The authors should clarify this to show that they understand the difference between local antiviral protection and crosstalk with the immune system.

If applicable, is the statistical analysis and its interpretation appropriate? Not applicable
Are all the source data underlying the results available to ensure full reproducibility? Yes

Are the conclusions drawn adequately supported by the results? No
No competing interests were disclosed.

Competing Interests:
We have read this submission. We believe that we have an appropriate level of expertise to confirm that it is of an acceptable scientific standard, however we have significant reservations, as outlined above.

Maria del Mar Ortega-Villaizan Romo
Dear Dr. Espen Rimstad and Dr. Maria Dahle, We appreciate very much your revision and constructive comments on the manuscript. We have included your corrections in the new version of the manuscript hoping that now the manuscript will be suitable for publication. Please find below the response to your comments: 1. The title is generalizing beyond the scope of the paper, and should specify the virus and the