Influenza infection rewires energy metabolism and induces browning features in adipose cells and tissues.

Like all obligate intracellular pathogens, influenza A virus (IAV) reprograms host cell’s glucose and lipid metabolism to promote its own replication. However, the impact of influenza infection on white adipose tissue (WAT), a key tissue in the control of systemic energy homeostasis, has not been yet characterized. Here, we show that influenza infection induces alterations in whole-body glucose metabolism that persist long after the virus has been cleared. We report depot-specific changes in the WAT of IAV-infected mice, notably characterized by the appearance of thermogenic brown-like adipocytes within the subcutaneous fat depot. Importantly, viral RNA- and viral antigen-harboring cells are detected in the WAT of infected mice. Using in vitro approaches, we find that IAV infection enhances the expression of brown-adipogenesis-related genes in preadipocytes. Overall, our findings shed light on the role that the white adipose tissue, which lies at the crossroads of nutrition, metabolism and immunity, may play in influenza infection.

Fig 5b-Reviewers 1 & 2 previously pointed out my major concern the authors are not measuring IAV in adipocytes.
The actual measurement appears to be of viral RNA and antigens being trafficked into the adipose tissue by infiltrating immune cells. It seems an unlikely coincidence that the kinetics match perfectly. The authors are harvesting fat tissue from the intraperitoneal area where these immune infiltrates are present and increase on the same time and amplitude as presented in Fig 5b. No attempt has been made or indicated to remove infiltrating immune cells from the tissue. OMICS of mixed cell populations should be referred to as such and not presented as if pure SCAT or EWAT was studied. Further, IAV has been shown to induce hemorrhages and hematomas in adipose tissue (DOI: 10.1128/JVI.01235-07) which as reviewer 1 pointed out only high path IAV would only transverse via blood in viremia but that doesn't exclude immune cells arriving via this route (which is supported by some of these data). Either way, these data indicate immune cells are present in these tissues and must be acknowledge in the text and taken into account in the interpretation of these data. IT IS NOT APPROPREATE to refer to these cells as "cells in a Depot" or in any other ambiguous terms when the authors have defined them. Again, reviewers 1 (Q3) & 2 (Q1) are correct, these are extremely low levels of virus, especially for lungs during this time, ignoring the peak is delayed. In their rebuttal the authors argue "the kinetic (days 0, 2, 4, 7, 14 and 28 postinfection) of viral RNA detection in the fat depots that has been done in the revised version of the manuscript clearly show a time-dependency of viral RNA accumulation in fat tissues ( Figure 5b). Importantly, the profiles in SCAT and EWAT were similar to that observed in the lungs, with a sharp peak at day 7 post-infection." This is troubling. The method of quantification the authors are using reports their values as RNA copies per cell with a limit of detection at 10^3 copies and assumes a cell contains 10 pg of RNA (PMID: 21185869). Fig 5b peak lung is approximately 10^7 copies per ug RNA. Applying Kawakami et al assumptions that roughly translate into 100 copies of M gene RNA per cell in the lung at peak. That is unbelievably low considering typical IAV lung titer is peaking at about 7 log10 TCID50 per ml lung homogenate. Further the authors point to the time delay as being specific to IAV accumulation in adipose, yet they appear to show a time delay in the lungs that is contrary to IAV kinetics in the lung. This, and the HA expression, supports the view that the authors are measuring viral antigen or RNA from immune cells. The authors present STAT1 and IFN antiviral defense as evidence for direct viral infection, which is untrue given the presence of virus harboring activated immune cells. Fig 5c attempts to address the first issue by presorting the tissue and the results are clear, the adipocytes have no appreciable virus (i.e. AF are below the LOD, AF is not defined I assume adipose). That is probably why we see no effort to determine the significance. However, it is disconcerting that the reported adipose value is below 50 M1 copies/ugRNA even though the authors clearly state this value is far below the standard curve they generated to extrapolate their data See 715-716: "draw a standard curve (Ct value/viral copy number) ranging from 3.8 x109 (Ct ~5) to 380 copies (Ct ~25); allowing virus quantification in samples by interpolation of the Ct values." Thus, this figure meaningless. One assumes this extrapolation is well below the limit of detection elsewhere (PMID: 21185869) where quantification via this method uses a standard curve but he atypical unit copies per ugRNA makes it hard to compare to other methods papers that use copies per cell. Perhaps the authors used an unorthodox normalization that diminished their results (?). The PCR reaction conditions should be reported with cycles as this impacts results and threshold cycle (Ct) values of samples considered positive. Overall, the Authors' responses gloss over and appears to ignore the elephant in the room that both reviewers are pointing out. It is somewhat alarming that in response to reviewer 2 Q1 that authors repeat the above arguments, then state "iv) despite the low RNA levels that were detected in SCAT and EWAT, transcriptome analysis showed significant induction of STAT1-dependenttype I IFN-mediated antiviral immune defense in both depots, and iv) HAexpressing immune cells and HA-expressing adipose cells were detected in the adipose tissues of infected animals." The later argument explains why cytokines are elevated here (claimed in the paper to be adipose secreted without support), how antigen got into adipose cells the authors can't infected, and why the adipose has induction of antiviral immune factors. These factors are well documented to alter metabolism in all cells, including adipose that could easily explain the "browning" effect. TCID50 would be helpful, but Supplementary Figure 6 is a-d, there is no "e" so I can't comment on resolution of the issues related to adipose infectivity in terms of kinetics and MOI. However, the copy levels appear within the LOD. Figure 5c clearly demonstrates the immune cell containing SVF fraction is where the viral RNA are. Thus, Lines like 246 must be corrected as they are misleading. "Viral RNA was detected in the SCAT and EWAT (but not in 247 pancreas)" or 266 "Since virus-bearing adipose cells were detected in the adipose tissues of IAV-infected mice (see Fig. 5c and 5d". "IAC cannot replicate in adipose (in vitro)" implies it is infecting the adipose, and as Referee 1 points out, the cells are not being acid washed after in vitro infection so Fig 6a is  . Statements like this are not substantiated without proper controls: "On the basis of this dataset, we conclude that influenza virus enters, replicates its genome, and expresses its viral proteins, in preadipocytes and adipocytes." Line 336 Claim 2-IAV may promote white adipose tissue browning by inducing preadipocytes to commit to brown-like/beige adipocytes. ii.IAV induces thermogenic brown-like/beige adipocytes in SCAT To claim this expression of UPC1 should be quantified. This is shown in Fig 2b-c. However, the error bars are expressed in SEM and do not reflect the variability of the representative blot. Change to Std Dev, note exact significance test performed, and change to scatter plot to represent the number of blots and mice used in graphical representation. Further, claims of "increased thermogenic profile" are not corroborated by the unknown statistical test and are indicated as the "most notable" evidence lines 166-170.
ii. " Figure 7b shows that infection led to elevated expression of browning related genes in preadipocytes butstrikingly -not in adipocytes." How do the authors reconcile the kinetics of IAV and preadipocytes turning brown to mediate the effects seen in the in vivo figures given this result? The claim that "We next used tissue explant cultures to investigate the impact of influenza infection on SCAT and EWAT's secretory functions" is misleading. These cytokines are systemically increased with IAV infection and the authors are not measuring secretion SCAT or EWAT. The authors also claim "leptin….was elevated in SCAT and low in EWAT" then "Importantly, IAV infection led to depot-specific differences in the adipose tissue's lipid metabolism, and leptin secretion.." IF the authors are trying to use Fig1d to bolster their argument that SCAT and EWAT are different then they should graph them together and use the appropriate statistical method to test this. at the heart of the claim that IAV promotes subcutaneous adipose tissue browning and is specific to SCAT not EWAT. Yet the histology slides were not stained for IAV and there is no indication as to blinding of the authors during data collection. This paper would benefit immeasurably from sending these slides out for independent analysis by a blinded histopathologist. If Supplemental 2a-b is a summary from a blinded expert, then this addresses some essential claims of the paper. These data should be presented in a meaningful way and the appropriate statistical analysis should be used to determine if IAV induces differences in browning of the SCAT and/or EWAT. The difference between IAV induced browning of SCAT and EWAT should be presented, I see no IAV induced difference between adipocyte area (i.e. In 2a the difference between SCAT and EWAT is constant for all 5 size groups irrespective of IAV). This should be moved to main figures and supported by raw numbers and representative histology.
All lines that state things like: "From the core set of the 148 genes that were commonly downregulated upon infection in SCAT and EWAT,. This is one example of many, all should be changed to accurately reflect the data presented in the paper, namely IAV infection causes these local and systemic changes as opposed to IAV infection of SCAT or EWAT. Further, as the previous reviewers point out "OMICS" of these tissues include activated immune cells.
Claim 4-IAV infection rewires energy metabolism (transcriptomics): As outlined above the transcriptomics data is on a mixed cell population and infiltrating immune cells secrete cytokines that can alter adipose metabolism (PMID: 17118983; PMID: 30249998). This claim is valid as it refers to IAV infection in general and does not attempt to claim IAV infection of adipose cells alter their metabolism. Correct all text accordingly, except in the discussion where qualified reasonable claims may be made.
i. IAV rewires mitochondrial bioenergetics (cell intrinsic)-Cell intrinsic changes in bioenergetics between SCAT and EWAT were not measured or supported by these data and can't be claimed. The authors could easily perform bioenergetics on in vitro infections of pure cultures.
ii. IAV rewires mitochondrial bioenergetics (cell extrinsic at SCAT & EWAT tissue level)-Bioenergetics of adipose ex vivo is simple and would substantiate these claims (i.e. mice infected and tissue removed (days PI equal to transcriptomics) and subjected to mitochondrial stress tests). Alternatively, a relative comparison between IAV SCAT and EWAT transcriptomes could answer this question normalized by housekeeping gene. As it stands the authors are presenting only downregulated SCAT transcriptomic data (Fig 3d) without showing EWAT. Additionally, these "more through" analysis are presented incoherently. The authors present only one table allowing readers to compare SCAT & EWAT Table S1 and it is of transcripts upregulated with IAV yet the majority of the text refers to a figure without an EWAT comparison and infers extensively about TCA cycle and OXPHOS without measuring it.

*no explanation of DEG in text
Claim 5-"novel aspects of the metabolic control of influenza virus-host interactions at the molecular, cellular and tissue levels."-Claims of "influenza-associated modifications in SCAT and EWAT resulted in changes in the host's metabolism." are overreaching. The infection is systemic and has well documented systemic changes in glucose/insulin etc, that are referred to here as "metabolic." The persistence long after infection is indicative of these systemic changes. The authors could specifically infect adipose in vivo and repeat or soften claims and make distinctions between their metabolic claims in terms of host-tissue-SCAT/EWAT-cellular.
6. Minor claims i. "influenza infection has a long-lasting impact on whole-body glucose metabolism, and that the impact is independent of the host's dietary status. ii. IAV changes host's glucose metabolism that persist long after infection has been cleared and IAV alters wholebody glucose metabolism long after resolution (>2wk). This has previously been reported, here is one example: PMID: 30203067, please refer to human studies too.
Are the claims novel? There are previous publications so lines 103-104 are not accurate, as was previously pointed out by referee 1, and should be changed to accurately reflect previous publications and claims of novelty including adipocytes as "previously unrecognized targets for influenza" should be stricken. Further, WAT browning has been seen with other viral infections so lines 477-8 need changed.

General Considerations
The paper be of interest to others in the field and could be marginally influential.
Are the claims convincing? No. As they stand many of the claims lack evidence. As detailed above proper controls are missing, statistical analysis is either lacking or undefined, necessary comparisons to support claims are missing, and data undermining claims are present.
Are there other experiments that would strengthen the paper further? Ex vivo and in vitro bioenergetics would improve metabolic claims for extrinsic metabolic changes in tissue and SCAT vs EWAT respectively. They are easy and would dramatically support unsubstantiated claims from transcriptomics. Previous reviewers requested RNA-Seq & Ex Vivo tissue measure that were not performed. Quantitation of existing data could also help. These would be necessary to support the current claims of the paper. However, I recommend adjusting claims to support the current data.
Are the claims appropriately discussed in the context of previous literature? No, there is a substantial amount of literature in this area that is not cited enough.
If the manuscript is unacceptable in its present form, does the study seem sufficiently promising that the authors should be encouraged to consider a resubmission in the future? Yes.
Is the manuscript clearly written? No, the manuscript has several figures that provide little if any evidence and the presentation of the transcriptomics doesn't aid the reader in assessing the claims. It should be shortened and focused on claims based on these data and only relevant figures used.
Have the authors done themselves justice without overselling their claims? No, the authors have actually undermined their findings by overreaching on nearly every major claim and putting discussion worthy adjectives/descriptive details in results. These data are compelling even if adipocytes are not directly infected.
Statistical analysis-Unknown how sound, the statistical section of the methods makes it hard to determine what tests were applied to these data and legends lack test information.

Responses to Reviewer's comments and suggestions
First, we would like to thank the Reviewer for the constructive criticisms, comments and suggestions. After completion of the suggested edits, the revised manuscript has significantly gained in precision and clarity.
All changes that have been made to fulfill Reviewer's requests are highlighted in the revised version of the manuscript.
In its revised form, the manuscript contains 8 Figures While the Reviewer acknowledged the presence of viral-antigen-positive preadipocytes in the fat tissues of influenza-infected mice (yet at lower magnitude than viral-antigen-positive immune cells), clarification for the detection of viral RNA in adipocytes was requested.

As suggested, the limit of detection of our qPCR assay has been added to all Figures that illustrate viral RNA quantification in tissues (SCAT, EWAT, lungs, pancreas), fat tissue-cell-fractions (stromal vascular cells vs. adipocytes), or cells (preadipocytes vs. adipocytes).
We agree with the fact that we could not write that adipocytes present in the fat tissues of influenza-infected mice were bearing viral RNA. This was inappropriate and, in fact, untrue. We thank the Reviewer for this insightful remark.

Viral RNA-and viral antigen-harboring cells are detected in the adipose tissue of influenza-infected mice.
Next, we sought to quantify viral RNA in the adipose tissues from IAV-infected mice (M1 negative strand specific qPCR assay). Albeit at lower levels than in the lungs, viral RNA was detected in SCAT and EWAT (but not in pancreas) at 7 dpi (Fig.  5a). To get some insight into the dynamics of viral RNA accumulation, viral RNA levels were quantified in the lungs, SCAT and EWAT at 2, 4, 7, 14 and 28 dpi (Fig. 5b). In the lungs, the viral load rapidly increased from 2 dpi, peaked at 7 dpi, and rapidly declined thereafter. In SCAT and EWAT, we also observed a time-dependency of viral RNA accumulation, which started from 2 dpi, culminated at 7 dpi and decreased afterwards, as noticed in the lungs. WAT's cellularity is highly heterogenous [18][19][20]22 . Mature, lipid-filled adipocytes make up approximately 30% of adipose tissue cells. The remaining stromal vascular fraction (SVF) cells include e.g. vascular cells, fibroblasts, innate and adaptive immune cells, and preadipocytes. We therefore separated mature adipocytes from SVF cells through enzymatic fractionation of the SCAT and EWAT from mock-treated and IAV-infected mice. As shown in Supplementary Fig. 5a, viral RNA was detected in SVF cells isolated from SCAT and not in SVF cells isolated from EWAT. No viral RNA was detected in adipocytes isolated from either fat depot (not shown). Thus, immune cells (identified as CD45 + cells) and preadipocytes (identified as CD45 -CD31 -CD29 + CD34 + Sca1 + cells) from the SVF of the SCAT and EWAT from mock-treated and IAV-infected mice were sorted by fluorescent-activated cell sorting (FACS) (see gating strategy Supplementary Fig. 5b) and stained for the viral protein hemagglutinin (HA) (Fig. 5d) (Fig. 6).
As suggested, illustration of the detection of the viral proteins NP, NS1 and HA viral in in vitro infected preadipocytes and adipocytes (immunofluorescent staining, and confocal microscopy) has been removed in the revised version of the manuscript.
In fact, the description of in vitro infection of preadipocytes and adipocytes has been completely reorganized, and shortened, as follows:

In vitro influenza virus infection is abortive in preadipocytes and triggers a brownlike/beige adipogenic program. Preadipocytes share numerous phenotypic and functional
properties with macrophages 37 . IAV replication in macrophages is most often abortive (i.e. no release of new infectious virions), yet certain virus strains can replicate productively in these cells 38,39 . Thus, the detection of HA-positive preadipocytes in the WAT from infected mice (see Fig. 5d) led us to question the IAV-infection ability of murine preadipocytes in vitro. We used mouse preadipocyte 3T3-L1 cells, which can differentiate into adipocytes upon induction with a standard hormone cocktail 40 . Hence, undifferentiated 3T3-L1 cells (from now on referred to as preadipocytes) were exposed to graded doses of IAV (at MOI of 0.01, 0.25, 0.5, 1 or 2) and viral RNA levels were evaluated at 6, 24 and 48 hpi. As shown in Fig. 6a, Fig. 6a). Unlike most RNA viruses, IAVs transcribe and replicate their genome inside the nucleus; this induces major nuclear and nucleolar ultrastructural changes 41 . As shown in Fig. 6b, highresolution transmission electron microscopy revealed the marked accumulation of viral M1associated rod-like tubular structures in the nucleus, and the complete segregation of nucleolar components in infected preadipocytes (MOI of 5, at 24 hpi), relative to non-infected cells. In addition, the particles budding from the preadipocytes' plasma membranes appeared to be empty (Fig. 6b, right panel) Fig. 6b). We next sought to determine preadipocytes' response to infection (MOI of 1, at 24 hpi) by quantifying the expression of the antiviral innate-immune-related genes Tlr3, RigI, Mda5, Mx1 and Vip (Fig. 6c) Fig. 6c). Importantly, infection enhanced the expression of genes involved in the browning process, such as Ucp1, Pgc1a (coding for peroxisome proliferator-activated receptor gamma coactivator 1 alpha), Fgf21 (coding for fibroblast growth factor 21), Apln (coding for apelin), Metrnl (coding for meteorin-like protein) and Tmem26 (coding for transmembrane protein 26) [42][43][44][45] in preadipocytes (MOI of 1, at 24, 48 and 72 hpi) (Figure 6d). Accordingly, adipocytes that had differentiated in vitro from IAV-infected preadipocytes had higher Ucp1 and Pgc1a transcript levels than adipocytes that had differentiated from mock-treated preadipocytes (Supplementary Fig. 6d).

. Relative to mock-treated cells, the expression of these genes was enhanced in infected preadipocytes. It is worth mentioning that in vitro infection of preadipocytes at much lower MOI (0.01) also led to increased Tlr3, RigI, Mda5, Mx1 and Vip transcript levels (Supplementary
Corresponding to the following revised Figure 6 and Supplementary Figure 6:

In addition, the transition for justifying in vitro infection of adipocytes (in the Results section), has been changed, as follows:
Although no viral RNA-harboring adipocytes were detected in either fat depots from IAVinfected mice, previous literature reported that adipocytes might be permissive to IAV infection in vitro 46,47 Supplementary Fig. 7 Fig. 7a), nuclear and nucleolar changes featuring efficient viral genome replication, and budding of empty particles ( Supplementary  Fig. 7b), as well as increased expression of the antiviral innate-immune-related genes Tlr3, RigI, Mda5, Mx1 and Vip (Supplementary Fig. 7c). As opposed to preadipocytes (see Fig. 6d), IAV infection of adipocytes did not increase the expression of Ucp1, Pgc1a, Fgf21, Apln, Metrnl and Tmem26 (Supplementary Fig. 7d). Fig. 7e).

Corresponding to the following revised Supplementary Figure 7:
Claim 2-i. IAV may promote white adipose tissue browning by inducing preadipocytes to commit to brown-like/beige adipocytes.ii. IAV induces thermogenic brown-like/beige adipocytes in SCAT * The first Reviewer's concerns regarding claim 2 is: i. IAV may promote white adipose tissue browning by inducing preadipocytes to commit to brown-like/beige adipocytes (Discussion).

We took into account the different remarks made by the Reviewer (such as infiltration of immune cells egressing from the infected lungs) to rewrite the Discussion on how influenza infection may lead to SCAT browning.
The text (Discussion) has been changed as follows:

Strikingly, we found that immune cells harboring viral proteins (HA) were present in the WAT (mostly SCAT) of influenza-infected mice. Immune cell egress from the lungs has been reported in the context of influenza infection 74,75 . Thus, immune cells (either infected or harboring remnant viral antigens and RNA) may have entered the bloodstream through the capillaries of damaged alveolar wall of the infected lungs and reached fat depots, participating to the observed changes in WAT's secretory function and to the induction of antiviral-innate-immunity-related pathways. The phenotypic characterization of immune cells that are present in the fat depots of IAV-infected mice is currently under investigation. Alongside immune cells, viral antigen-positive preadipocytes were also found, although less frequently, in the WAT from infected mice. How preadipocytes acquired viral antigens from immune cells that have migrated from infected lungs to WAT, remains to be established.
* The second Reviewer's concerns regarding claim 2 is: ii. IAV induces thermogenic brown-like/beige adipocytes in SCAT (Fig. 2a, 2b (histology and  morphometry), and Fig. 2c, 2d (UCP1)).
- Fig. 2a, 2b. (histology & histomorphometry of SCAT from mock-treated mice vs. SCAT from IAV-infected mice). Histology of the fat tissues has been made elsewhere than in our laboratory (see Acknowledgement section). For determination of adipocyte-size in the fat depots, this was done using an automatic program that measure the adipocyte size of ~ 1000 adipocytes per slide. Crude data of the morphometry analysis have been provided as Supplementary data (4 Excel files): (#8 to 14), and means), and (4) Crude data morphometry analysis EWAT from IAV-infected mice (values for each individual mouse (#1 to 7), and means).

(1) Crude data morphometry analysis SCAT from mock-treated mice (values for each individual mouse (#8 to 14), and means), (2) Crude data morphometry analysis SCAT from IAV-infected mice (values for each individual mouse (#1 to 7), and means), (3) Crude data morphometry analysis EWAT from mock-treated mice (values for each individual mouse
Besides, as stated in the Introduction, the differences between SCAT and EWAT have been repeatedly described (Reference 22), these fat depots notably differed regarding the adipocyte size: the adipocytes within SCAT are smaller than adipocytes within EWAT (Reference 28). Thus, the comparison of adipocyte sizes between SCAT vs. EWAT from mocktreated and IAV-infected animals confirmed what is known from literature, that is the reason why these results are presented as Supplementary Fig. 2c. The appearance of brown-like/beige adipocytes in fat tissues was only observed in SCAT and not in EWAT. This was confirmed by the increased number of small adipocytes in the SCAT of IAV-infected mice vs. the SCAT of mock-treated mice (Fig. 2b), the increased expression of the thermogenic, brown-specific, protein UCP1 in the SCAT of IAV-infected mice vs. mocktreated mice (Fig. 2c, 2d), and the increase local surface temperature (Fig. 2e).

All these observations were absent when comparing the EWAT from IAVinfected mice vs. the EWAT from mock-treated mice.
We apologize if this was unclear. We therefore reformulate the text (Results), as follows:

Influenza infection is associated with subcutaneous adipose tissue browning.
To determine whether SCAT and EWAT undergo qualitative and/or quantitative changes following influenza infection, we performed histologic and histomorphometric analyses at 7 dpi. The EWAT samples from mock-treated and IAV-infected mice had a similar histological appearance i.e. white adipocytes containing large unilocular lipid droplets (a typical WAT histology) (Supplementary Fig. 2a). In contrast, the SCAT samples from mock-treated and IAV-infected mice showed histological differences: numerous pockets of dense, small and multilocular adipocytes -a feature of brown-like adipocytes, also known as beige or brite adipocytes 27 -were found interspersed within the SCAT from infected animals (Fig. 2a). As reported in literature 28 , histomorphometric analyses confirmed that adipocytes are smaller in SCAT than EWAT (Supplementary Fig. 2c). In accordance with the appearance of small brown-like/beige adipocytes, the SCAT from IAV-infected mice showed higher numbers of small adipocytes and reduced numbers of large adipocytes than the SCAT from mock-treated mice (Fig. 2b). The differences in adipocyte size between mock-treated and IAV-infected mice were less pronounced for the EWAT (Supplementary Fig. 2b). Moreover, levels of the mitochondrial uncoupling protein 1 (UCP1, a marker of the brown adipose tissue 29 ) were higher in SCAT protein lysates from infected mice than in those from mock-treated controls. In contrast, no UCP1 was detected in the lungs or EWAT lysates from IAV-infected mice (Fig. 2c,d). In line with increased levels of thermogenic UCP1, the surface temperature generated from the SCAT was higher in IAV-infected mice than in mock-treated mice (Fig. 2e).
These results emphasized the depot-specific nature of the effects of influenza infection on fat, with the appearance of thermogenic brown-like/beige adipocytes in the subcutaneous depot being the most notable.
Corresponding to the following revised Figure

In general, all bar graphs have either been converted to box-and-whisker plots, dot plots (with mean +/-SD (instead of SEM)), or dot-plots overlaid on the bars (+/-SD (instead of SEM)).
This change also concerned the data illustrating the increased expression of the thermogenic protein UCP1 in the SCAT upon influenza infection (combining the data from 4 Western-blots covering a total of protein extracts from 5 SCAT from mock-treated mice, 35 SCAT from IAV-infected mice, 5 EWAT from IAV-infected mice, and 5 Lungs from IAV-infected mice) showed on Fig. 2c. (representative blot), 2d. (quantification of all blots). In addition, as requested from the Editorial Board, Western-blot 's crude data (gels and quantification) have been provided as Supplementary data.
Claim 3-IAV infection (of the host) leads to profound changes in the major WAT depots (i.e. the inguinal SCAT and the visceral (epididymal) adipose tissue (EWAT)) ( Fig. 1) In general, the statistics used have been detailed in the revised version of the manuscript (Materials & Methods section).
- Fig. 1b. (fat masses). As requested, SCAT and EWAT masses from mock-treated vs. IAV-infected mice, formerly presented as bars, have been converted to box-and-whisker plots.
- As stated in the Introduction, the differences between SCAT and EWAT have been repeatedly described, these fat depots notably differed regarding their secretory functions (e.g. Reference 22). Here, we intended not to compare SCAT and EWAT in general but to compare the SCAT's response to IAV infection vs. the EWAT's response to IAV infection. Compared to mock-treated controls, infection was associated to increased secretion of leptin by the SCAT, and to decreased secretion of leptin by the EWAT (while SCAT and EWAT both enhanced their production of IL-1b, IL-6 and IL-10 upon infection). We apologize for this misunderstanding, and we reformulated this part of the Results section as follows: Adipose tissue is composed of several cell types that include mature adipocytes, preadipocytes, and a range of innate and adaptive immune cells [17][18][19][20] . As an endocrine organ, WAT secretes numerous cytokines, metabolites and hormones that originate from all these cell types 18 . By preserving the physiological in vivo cross-talk between cells, explant cultures of WAT allow analyzing the secretory function of the whole tissue 25 . We therefore used tissue explant cultures to investigate whether influenza infection was associated with changes in SCAT and EWAT's secretory function (Fig. 1d). When compared with controls, the adipose tissues collected from infected mice produced higher levels of the proinflammatory cytokines IL-1b and IL-6, and of the anti-inflammatory cytokine IL-10. Following IAV infection, the release of leptin -a WATderived hormone that regulates feeding behavior, body weight 26 , and innate and adaptive immune responses 21 -was enhanced for SCAT and decreased for EWAT.

Claim 4-IAV infection rewires energy metabolism (transcriptomics):
As stated by the Reviewer, transcriptomic analyses on whole fat tissues correspond to a mixture of numerous, and different cell-types that composed the tissue. As suggested, the revised version of the manuscript now emphasized more the heterogeneity of the adipose tissue, regarding its cellular composition. The text has been changed according to this essential comment, as follows: -Results- Fig. 1.
Adipose tissue is composed of several cell types that include mature adipocytes, preadipocytes, and a range of innate and adaptive immune cells [17][18][19][20] . As an endocrine organ, WAT secretes numerous cytokines, metabolites and hormones that originate from all these cell types 18 . By preserving the physiological in vivo cross-talk between cells, explant cultures of WAT allow analyzing the secretory function of the whole tissue 25 . We therefore used tissue explant cultures to investigate whether influenza infection was associated with changes in SCAT and EWAT's secretory function (Fig. 1d).
WAT's cellularity is highly heterogenous [18][19][20]22  Besides, 1. We agree that our work did not address the effect of influenza infection on fat tissue, in a cell-intrinsic way. To study cell intrinsic changes in adipose tissue, one should isolate adipocytes and preadipocytes (but also the different innate and adaptive immune cells) from the fat tissue of IAV-infected mice and do bioenergetics on these cells. As stated in the revised Discussion, technologies such as single-cell analyses would help answer this key question, 2. Transcriptome analyses did not intend to compare the transcriptomes from SCAT vs. EWAT (it is known that they differed (Reference 22)) but to compare the transcriptome from SCAT mock vs. SCAT IAV; and the transcriptome from EWAT mock vs. EWAT IAV. We apologize if this was unclear. This part has also been re-written and shortened, as follows: To gain insights into the molecular basis of the depot-specific effects of influenza infection on fat tissues, we performed transcriptomic analyses on the SCAT and EWAT isolated from mock-treated and IAV-infected mice (p≤0.05 and fold change (FC) cutoff of ≥2). In SCAT, 1,214 genes were found differentially expressed between mock-treated and IAV-infected mice.
In EWAT, 660 differentially expressed genes (DEGs) were identified. Venn diagrams were used to display comparison of the lists of upregulated or downregulated DEGs (Fig. 3a). Upon infection, 296 genes were upregulated in both SCAT and EWAT, 268 genes were upregulated only in SCAT, and 177 genes were upregulated only in EWAT.

Nota bene:
We agree that the throughout analysis of the transcriptomic data should have been done at the first place. Yet, we first used free programs (GO Resources) to analyze our transcriptomic data, we only had access to Ingenuity Pathway Analysis (IPA) later.