Elovanoids downregulate canonical SARS-CoV-2 cell-entry mediators and enhance protective signaling in human alveolar cells

The pro-homeostatic lipid mediators elovanoids (ELVs) attenuate cell binding and entrance of the SARS-CoV-2 receptor-binding domain (RBD) in human primary alveoli cells in culture. We uncovered that very-long-chain polyunsaturated fatty acid precursors (VLC-PUFA,n-3) activate ELV biosynthesis in lung cells. Both ELVs and their precursors reduce the binding to RBD. ELVs downregulate angiotensin-converting enzyme 2 (ACE2) and enhance the expression of a set of protective proteins hindering cell surface virus binding and upregulating defensive proteins against lung damage. These �ndings open avenues for potential preventive and disease-modiable therapeutic approaches for COVID-19.


Introduction
The high transmissibility of Severe Acute Respiratory Syndrome-coronavirus 2 (SARS-CoV-2) is due, at least in part, to infectivity in lung type II alveolar cells 1 SARS-CoV-2 triggers a wide range of disease phenotypes with severe acute respiratory distress syndrome (ARDS), including interstitial pneumonia 2 and viral sepsis 3 .Here, we tested if the pro-homeostatic lipid mediators, the elovanoids (ELVs) [4][5][6][7] , would block the entrance of the spike (S) protein receptor-binding domain (RBD) that would prompt a protective response against SARS-CoV-2 infection.
Lung alveoli viral attachment through the S protein RBD to ACE2 is followed by proteolytic activation for fusion and viral cell entry [8][9][10] .We use human alveolar primary cell cultures (Fig. 1a, and Extended Data Fig. 1).Most of the cells are type II (oil red, speci c marker, Fig. 1a, right panel, and Extended Data Fig. 1, bottom panels) and positive to Foxj1, HT2-280 antigen, and β-tubulin IV (Fig. 1a, right panel; Fig. 1g,h and Extended Data Fig. 1).Pneumocytes type II are also mobile, showing lamellae or lopodia positive to HT2-280, a speci c type II (Fig. 1b).We exposed these cells to RBD (from S protein)-Alexa 594 for 24 hours.In parallel, we used Nucleocapsid protein N as a speci city control of RBD internalization.In 3D reconstructions of Z-stack images (Imaris software, Bitplane, UK), the lipophilic staining (cell mask) shows a dense membrane above the nuclear zone that is localized close to oil red (Fig. 1a,c i-viii)..A below view (Fig. 1c vi) shows that the RBD protein signal (red) passes through the membrane (white) to the intracellular space surrounding the nucleus (blue), and can also be seen in the above view (Fig. 1c vii,viii).. Herein, we demonstrate for the rst time that RBD was shown to be internalized in SARS-CoV-2 since previous work have shown the same for SARS-CoV-1 11 .When IL1β or TNFα was added, the internalized RBD signal was increased (Fig. 1d).In addition, Nucleocapsid (N) protein, a structural viral protein not involved in ACE2 and SARS-CoV-2 interaction 12 , is at the same level as the control with no protein added that accounted for auto uorescence.RBD was internalized at higher rates than N.In digitalized images, plotted vs. Z-axis in a Z-stack shows the differential position of the N protein versus the RBD with respect to the membrane level (Fig. 1c i-v,ix-xi and Fig. 1d).This observation documents that the N protein remains on the membrane and the extracellular side while the RBD spans intracellularly, passing through to the cytoplasm and demonstrating that RBD internalization is speci c and dependent on IL1β and TNFα (Fig. 1d,e).RBD-Alexa 594 internalization was decreased +/-IL1β + TNFα when ELV-N32 and ELV-N34 were added (Fig. 1f, upper panel).In addition, acetylenic ELV-N32 or ELV-N34 (Extended Data Fig. 2) showed a steep decrease in RBD protein internalization +/-IL1β (Fig. 1f, lower panel).Moreover, the addition of the ELVs precursors 32:6 or 34:6 reduces RBD located below the membrane, suggesting that the pneumocytes convert these precursors into ELVs and thus prevent RBD internalization (Fig. 1f, upper panel).
Since ELVs stimulate protective proteins expression in cells confronted with uncompensated oxidative stress [5][6][7] , we next explored if these lipids under conditions that downregulate canonical SARS-CoV-2 cell-entry mediators in pneumocytes will also activate protective proteins synthesis.We found that ELV heightens the expression of Sirtuin 1 (Fig. 1j), RNF146 (Extended Data Fig. 3a,b), PHB, Bcl-Xl, and Bcl2 (Extended Data Fig. 3c-e).These proteins are involved in pro-homeostatic cellular functions.Sirtuin 1 (Silent information regulator factor 2-related enzyme 1) is a NAD(+)-dependent deacetylase of histone and non-histone proteins and transcription factors, and its regulatory functions target in ammation, aging, mitochondrial biogenesis, and cellular senescence 13 .RNF146 is an E3 ubiquitin-protein ligase that degrades parsylated proteins, thus protecting cells from Parthanatos cell death 14 .PHB (prohibitin type I) functions comprise scaffolding mitochondrial protein, adaptor in membrane signaling, transcriptional coregulator, and neuroprotection 6 .Bcl-XL and Bcll2 downregulate apoptosis and in ammasome formation 15 .Our data suggest that, in addition to halting the entrance of the RBD, ELVs in the lung curb cell-damaging/apoptotic events and thus sustains homeostasis by counteracting in ammation overactivation by the formation of protective proteins.
An evolving question prompted by our data is whether alveolar cells in culture can synthesize ELVs.Thus, we incubated human alveolar cells with the precursors VLC-PUFAs (32:6 or 34:6) and then analyzed the products by LC-MS/MS.Interestingly, we found that ELVs are in fact, formed.ELV-N32 was synthesized where the precursor 32:6 was added and not in cells exposed to 34:6.Inversely, ELV-N34 was found in the cultures were 34:6 was added and not in cells exposed to 32:6 (Fig. 2a,c).These results demonstrate that alveolar cells are endowed with pathways for the biosynthesis of ELV-N32 and ELV-N34 (Fig. 2b).We show MS fragmentation for stable derivatives of intermediaries (Fig. 2a-c) as well as of ELVs themselves (Fig. 2a).Moreover, we uncovered that ELVs were actively released from cells to the incubation media, indicating that they act both as autocrine and paracrine mediators.
Our ndings contribute to broadening our understanding of the duality of ACE2 in lung function and diseases.In health, ACE2 fosters lung homeostasis by generating Ang-(1-7) and enhancing host defense that would counteract ACE2 virus-induced downregulation of proin ammatory signaling.Herein, we show that ELVs uncover another participant when RBD of the S protein binds to ACE2 and enters alveolar cells in culture.The ELVs are likely part of a fast and coordinated pro-homeostatic in ammatory downregulatory response.To be tested in the future is the prediction that delayed ELV-mediated protective responses would lead to severe lung and systemic in ammation.So direct virus triggered cell damage is critical, but also the activation of the induction of protective proteins.Is diet engaged in building precursors of ELVs in the lung?Diet has been shown to affect ACE2 expression 16 and the supply to build ELV precursors 7,17 .This may contribute to explaining why some patients develop hyperin ammatory/immune responses and severe disease, but others experience mild or even asymptomatic COVID-19.Questions that remain to be addressed include whether the expression of the protective proteins identi ed here in the alveoli are activated all at once? Are they coordinated with adaptive immune responses to limit virus spread?Are enzymes for ELV synthesis under tight transcriptional control so that the mediators are expressed at appropriate times and/or levels?To our knowledge, ELVs are the rst protective mediators to be identi ed in the human alveoli confronted with the RBD of the S protein.
Additional research will be needed to elucidate the molecular mechanisms of ACE2 downregulation.Also, the use of the entire S protein instead of the RBD, as in our present study, will provide the connection between cell attachment and cell entrance, as affected by ELVs and VLC-PUFAs, since proteases expression is correlated with ACE2 downregulation.Moreover, the use of the intact virus would offer a direct demonstration of the signi cance of ELVs.Since the SARS-CoV-2 affects nasal mucosa, GI, the eye, and the nervous system exploring the protective potential of ELVs in other cell types would further expand the scope of our observations beyond the lung.Our results provide a foundation for future research and offer speci c mediators for interventions to modify disease risk, progression, and protection of the lung from COVID-19 or other pathologies.

Methods
Primary cultures of human alveolar cells and assessment of protein internalization.
We have used primary cell cultures of human alveoli, which consist of a mixture of ciliated cells, club cells, type I pneumocytes, and type II pneumocytes (PromoCell, HSAEpC).We have characterized the histology and immunocytochemistry in these primary cultures (Fig. 1a,b,g,h).We performed all the experiments in 48 wells with passage 4 cells seeded at 15000 cells per cm 2 density.The cells were incubated to con uency and maintained in the proprietary Medium provided by Promocell with the addition of Pen/Strep and exposed to 0.5 g of tagged protein per well for 24 hours in the presence or absence of 10 ng/ml IL1β (PeProTech Inc., Rocky Hill, NJ Cat# 200-01B) and/or 10 ng/ml TNFα (Cell Sciences Inc., Newburyport MA.Cat# CRH520B).After this period, cells were incubated 10 min with 1/1000 cell mask (Thermo Scienti c cat# C37608) medium and xed using PFA 4%.After xation, nuclei were stained with 10 g/ml Hoechst 33342 (Thermo Scienti c cat#H3570).To characterize the cell types in culture after xation, we performed immunocytochemistry using the following primary antibodies: HT1-53, a marker of pneumocytes type I, and HT2-280 (Terrace Biotech cat#HT1-53 and HT2-280); Foxj1 (Santa Cruz Biotech, sc-53139) and β-Tubulin IV marker of pneumocytes type II (Abcam cat# ab179509).ACE2 (Santa Cruz Biotech, sc-390851) and TMPRSS2 (Abcam cat #Ab109131) were used for Immunostain the two mentioned proteins in cell culture.ACE2 Protein abundance using Jess technology.
The western assay was performed using a Jess Protein Simple system (San Jose, CA, USA) following the manufacturer's protocol.Brie y, samples were lysed with RIPA buffer containing a protease inhibitor cocktail (Sigma, Cat.P8340).Soluble protein concentration was determined by BCA assay (Thermo Fisher Scienti c, Cat.23225) and 0.4 µg used/reaction.Samples were heated at 95 °C/5 min, and 3 µL of each sample were loaded.The 12-230 kDa cartridge (Protein Simple-#SM-W004) was used.Primary antibodies were diluted in antibody diluent 2 buffer (Protein Simple, #042-203), and the working solution of secondary antibodies was provided by the company (Protein Simple, #042-206).For data analysis, the area of spectra that matched the molecular weight of the target protein was used (Fig. 1j).We used the anti-ACE2 antibody from Abcam (cat# ab108252) in a concentration of 1 g/ml.The standardization was performed using total protein stain and using an anti-GAPDH antibody from Santa Cruz Biotech (cat# Sc-25778).

Quanti cation of RNF-146.
Western blot was performed from human primary alveolar cells, as described in Calandria et al.,, 2015  (18).Brie y, cell lysates were produced using RIPA buffer supplemented with protease inhibitor cocktail (Sigma, cat# P8340.St Louis MO).Total protein (30 mg) was mixed with Laemmli buffer containing DTT and loaded in Novex 4-12% precast gels and ran in X-Cell running system at 120V for 1.5 hours.The transference was performed using the Trans Blot Turbo dry transferring system (Bio-Rad, Hercules CA) on low uorescent background PVDF membranes (GE Healthcare, Piscataway NJ).Membranes were incubated with the corresponding primary antibodies overnight.Primary antibodies used RNF-146 (UC Davis/NIH-Neuromab Lab Facility, cat #75-233) and GAPDH (Satnta Cruz Biotech Cat# sc-47724).After this period, the membranes were incubated with uorescent-tagged secondary antibodies (GE healthcare, cat# PA45011) for 1 hour and imaged.Data was acquired using ChemiDoc MP (Biorad).Densitometric analysis was performed using ImageLab 6.0.1.(Biorad).
Preparation of tagged RBD and N nucleocapside proteins.
We have obtained the Recombinant SARS-CoV-2, S1 Subunit Protein (RBD), and Recombinant SARS-CoV-2 Nucleocapsid Protein from Raybiotech (Peachtree Corners GA, cat# 230-30162-1000 and 230-30164-500 correspondingly).The proteins were labeled using Alexa Fluor™ 546 Protein Labeling Kit from Thermo Scienti c (cat# A10237) following manufacturer directions except for the RBD that was puri ed from the dye with Amicon-Ultra 10K cutoff lters (Merck, Millipore cat#UFC201024) instead of the column provided by the kit has a restrictive MW of 50KD (the recombinant RBD protein was 25KDa).The recovery of the protein and labeling e ciency was measured using nanodrop and was about 80% recovery and 0.02 dye molecules per aminoacid.We added 0.5 g per well of protein.
Quantitation of cell surface binding and internalization of tag RBD of the viral S protein.
0.5 g of Alexa 594-conjugated RBD domain belonging to the SARS-CoV-2 virus Spike protein (Raybiotech.Cat.230-30162-1000) was incubated with human alveolar cells for 24 hours.After this period, Cell Mask (Thermo sher Scienti c, cat#C37608) was added to a nal concentration of 1 in 1000 and left in the incubator for 10 minutes.The cells were then xed with 4% Paraformaldehyde in PBS, washed three times, and Nucblue (Thermo sher Scienti c cat# R37606) was added for nuclear staining.The images were taken using an Olympus FluoView 3000 laser confocal microscope as z-stacks with a xed step size of 1.6 m.The cero was registered for each well specifying -30 and 30 m as the lower and upper limit in the Z-axes.Using the Z-drift compensation system, 9 blind positions were set up per well, and the image acquisition was performed automatically.The pictures were processed using Imaris 9.5.1 software (Bitplane) to render the 3D image and assess the position of the different surfaces (elements rendered with the surface function) along the Z-axes.The portion of the tagged protein that was internalized was calculated from the total sum of intensity that crossed the membrane inner limit to the intracellular space in the Z-axes.The proportion was calculated in percentage of internalization.
Real-Time PCR using Taqman probes and SYBR green assay.
cDNA produced using 1 g of total RNA extracted by RNAeasy (Qiagen, Hilden Germany, cat# 74104).The rst strand of cDNA was produced using iScript™ Reverse Transcription Supermix for RT-qPCR (BioRad cat # 1708840).The quanti cation of Sirt 1, RNF-146, Bcl2, BcL-xl was performed using SYBRgreen assay with primers designed in house (Table 1) using SsoAdvanced Universal SYBR Green Supermix (Biorad cat#1725270).The quanti cation of ACE2 and TMPRSS2 mRNA was performed using Taqman probes (Biorad cat# qHsaCEP0051563 & qHsaCIP0028919 respectively) labeled with FAM and standardized using PGK1 probe labeled with HEX (Biorad cat# qHsaCEP0050174).Boxes color code all the histograms in the gure.The inserts depict the structures of 27-monohydroxy-32:6 and 29-monohydroxy-34:6 along with the product ions as they are cleaved at a given bond.The complete structures of ELV-N32 and ELV-N34 were