Western diet increases COVID-19 disease severity in the Syrian hamster

Summary Pre-existing comorbidities such as obesity or metabolic diseases can adversely affect the clinical outcome of COVID-19. Chronic metabolic disorders are globally on the rise and often a consequence of an unhealthy diet, referred to as a Western Diet. For the first time in the Syrian hamster model, we demonstrate the detrimental impact of a continuous high-fat high-sugar diet on COVID-19 outcome. We observed increased weight loss and lung pathology, such as exudate, vasculitis, hemorrhage, fibrin, and edema, delayed viral clearance and functional lung recovery, and prolonged viral shedding. This was accompanied by an increased trend of systemic IL-10 and IL-6, as well as a dysregulated serum lipid response dominated by polyunsaturated fatty acid-containing phosphatidylethanolamine, recapitulating cytokine and lipid responses associated with severe human COVID-19. Our data support the hamster model for testing restrictive or targeted diets and immunomodulatory therapies to mediate the adverse effects of metabolic disease on COVID-19.


Introduction
response after application of oral glucose load and found no difference between the diet regimens. 91 The insulin resistance index (fasting glucose level (mmol/L) x fasting insulin level (mIU/L) showed 92 no significant differences (Fig 1 C assess diet induced pathology. There was no difference in body fat-to-weight ratio (Fig 1 D, N

High-fat and high-sugar diet induces liver damage and systemic hyperlipidemia 99
We investigated the changes in lipid metabolism through a blood lipid biochemistry panel (Sup 100 Table 1). Due to increased levels of fat in the samples collected from HFHS animals, HDL and 101 At 14 DPI, thickened septa, presumably from interstitial fibrosis with alveolar bronchiolization, 194 were observed in lungs from RD animals (N = 2) (Sup Fig 4 A-D). In contrast, HFHS animals at 195 14 DPI had less septal thickening and more septal, alveolar, and perivascular inflammation (N = 196 2). At 21 DPI four out of five of the RD animals and three out of three of the HFHS animals had 197 thickened alveolar septa with alveolar bronchiolization (Sup Fig 4 E-H). 198 Immunohistochemistry staining for SARS-CoV-2 antigen was increased at 7 DPI in lungs of HFHS 199 animals compared to RD animals (median = 2.71 (RD) / 5.043 (HFHS), N = 10 / 4) (Fig 4 E.

J.L). 200
To confirm this finding, we compared genomic RNA, subgenomic (sg)RNA (surrogate for 201 replicating virus) and infectious viral particles isolated from lungs at 7 DPI. Levels of gRNA and 202 sgRNA in the lungs of HFHS animals at 7 DPI were significantly increased as compared to RD 203 animals. Additionally, no infectious virus could be isolated from a subset of RD animals and 204 overall, significantly more infectious virus could be isolated in HFHS animals (Fig 4 M.N (which was not certified by peer review) is the author/funder. This article is a US Government work. It is not subject to copyright under 17 USC The copyright holder for this preprint this version posted June 17, 2021. ; https://doi.org/10.1101/2021.06.17.448814 doi: bioRxiv preprint

Immune infiltration in the lung during the acute-phase of infection and humoral immunity 219
are not significantly affected by high-fat high-sugar diet 220 Using immunohistochemistry, we investigated the infiltration of macrophages (IBA 1 staining), T-221 cells (CD3 staining), and B-cells (Pax 5 staining) over the course of infection (Fig 5). Macrophages 222 were detected throughout all sections but were increased in 7 and 14 DPI samples in pneumonic 223 areas irrespective of diet regimen. In addition, T lymphocytes were increased in 7 and 14 DPI 224 samples in pneumonic areas. No increase in B cells was observed. To quantify the influx of 225 macrophages and T cells we used morphometric analysis (Sup Fig 5). No significant difference 226 was seen between the RD and HFHS groups. Both macrophages and T cells increased in 227 numbers at 7 DPI as compared to pre-challenge conditions for both groups. (Fig 6 A (3.075 / 3.530 (pre-challenge)) / (13.630 / 10.480 (7 DPI)) % reactivity and median T cells = (4.515 230 / 4.125 (pre-challenge)) / (11.340 / 11.255 (7 DPI)) % reactivity, ordinary two-way ANOVA, 231 followed by Sidak's multiple comparisons test, p = 0.1007 / 0.3564 and p = 0.0001 / 0.0001, 232 respectively). 233 The humoral response to SARS-CoV-2 was not significantly impacted by diet regimen. Animals 234 seroconverted at 7 DPI, as measured by anti-spike IgG ELISA (Fig 6 C respectively). Neutralization of virus by sera collected at 14 and 21 DPI was compared to assess 238 potential differences in affinity maturation and no significant difference was found (Fig 6 D DPI median = 120/120 reciprocal titer, ordinary two-way ANOVA, followed by Tukey's multiple 241 comparisons test, p = 0.5535 and p = 0.4688, respectively). The cytokine kinetics were analyzed in serum throughout the course of infection by ELISA. Serum 246 samples were collected pre-challenge (0 DPI), on 7 DPI, 14 DPI and 21 DPI (Fig 6 E). Pro-247 inflammatory tumor necrosis factor (TNF)-a, interleukin (IL)-6, antiviral interferon (IFN)-g, and (IL)-248 10 did not significantly differ between diet regimens pre-challenge. After infection, RD animals 249 mounted a significant IFN-g response which lasted into recovery (14 and 21 DPI), while no 250 response was seen in HFHS animals (RD: N = 5/10, HFHS: N = 4, pre-challenge median = 629 / 251 To examine compositional changes in the circulating lipidome over the course of infection, the 263 lipidome was analyzed between 0 DPI and 7 DPI of infection. This analysis revealed distinct lipid 264 dynamics in response to SARS-CoV-2 infection (Fig 6 F). RD  diet-induced morbidity, led to increased weight gain during adolescence, and ultimately led to in 282 increased glucose tolerance, systemic hyperlipidemia, increased total cholesterol and a liver 283 pathology reminiscent of a NAFLD-like phenotype. The lack of net weight gain in this model may 284 present a means of decoupling liver associated pathologies such as NAFLD from obesity-285 associated disease more broadly. In humans NAFLD is predominantly a consequence of obesity 286 and frequently associated also with other comorbidities as well (Sanyal, 2019). In the context of 287 COVID-19, NAFLD is associated with increased hospitalization and disease severity (Bramante CoV-2, which was accompanied by more severe disease presentation. Our data is in agreement 297 with findings in mice, which have reported enhanced morbidity in aged and diabetic obese mice 298 in a mouse-adapted SARS-CoV-2 model (Rathnasinghe et al., 2021). Conversely, we also 299 observed increased weight loss, pathology, delayed lung recovery and influx of immune cells into 300 the lung in a subset of hamsters fed a regular diet as compared to what has been shown in 301 younger animals (Chan et al., 2020; Rosenke et al., 2020). This is likely due to the increased age 302 of the animals used in this study (Osterrieder et al., 2020). Previously, lung function analysis after

High-fat high-sugar diet 388
Four to six-week-old male Syrian Golden hamsters (ENVIGO) were randomly assigned to either 389 regular rodent chow (Teklad Global 16% Protein Rodent Diet, Envigo) or a HFHS diet for 16 390 weeks (Purina Chow #5001 with 11.5% Corn Oil, 11.5% Coconut Oil, 0.5% Cholesterol, 0.25% 391 Deoxycholic Acid, and 10% Fructose: Dyets Inc., Dyet#615088). Pre-challenge oral glucose tests 392 were performed on all animals. Five animals from each diet group were euthanized after the 16 393 wks for collection of pre-challenge tissue samples and weights. For each diet group, 5 animals 394 were randomly designated for flexiVent calibration and excluded from further analysis. Three 395 animals in the HFHS regimen were euthanized throughout the 16-week diet regimen due to 396 secondary morbidities and were not included in analyses. Pre-challenge, an additional 5 animals 397 in the RD group and additional 8 animals in the HFHS group were excluded from the study due 398 to experimental reasons, and one animal in the HFHS group due to secondary morbidities. glucose load (2 g/kg glucose) was administered. Blood samples were collected from the 404 retroorbital sinus using capillary tube at 0-, 30-, 60-, and 120-min post glucose administration. 405 Blood glucose was measured using the AlphaTRAK blood glucose monitoring system (Zoetis), 406 calibrated for cats. Serum was separated and used for measurement of insulin. Insulin was 407 measured using the rat/mouse insulin ELISA kit from Millipore (EZRMI-13K), according to the 408 manufacturer's instructions (Wang et al., 2001).              105 and is also made available for use under a CC0 license.