key: cord-0839402-njivdxqr
authors: Kilic, Gizem; Bulut, Ozlem; Jaeger, Martin; Horst, Rob ter; Koeken, Valerie A. C. M.; Moorlag, Simone; Mourits, Vera P.; de Bree, Charlotte; Domínguez-Andrés, Jorge; Joosten, Leo A. B.; Netea, Mihai G.
title: The immunological factors predisposing to severe COVID-19 are already present in healthy elderly and men
date: 2021-05-03
journal: bioRxiv
DOI: 10.1101/2021.04.30.442229
sha: b7354822f32b4f205d3114659ca1b98a98081659
doc_id: 839402
cord_uid: njivdxqr

Background Male sex and old age are risk factors for COVID-19 severity, but the underlying causes are unknown. A possible explanation for this might be the differences in immunological profiles in males and the elderly before the infection. Given the seasonal profile of COVID-19, the seasonal response against SARS-CoV-2 could also be different in these groups. Methods The abundance of circulating proteins and immune populations associated with severe COVID-19 was analyzed in 2 healthy cohorts. PBMCs of female, male, young, and old subjects in different seasons of the year were stimulated with heat-inactivated SARS-CoV-2. Result Several T cell subsets, which are known to be depleted in severe COVID-19 patients, were intrinsically less abundant in men and older individuals. Plasma proteins increasing with disease severity, including HGF, IL-8, and MCP-1, were more abundant in the elderly and males. The elderly produced significantly more IL-1RA and had a dysregulated IFNγ response with lower production in the summer compared with young individuals. Conclusions The immune characteristics of severe COVID-19, described by a differential abundance of immune cells and circulating inflammatory proteins, are intrinsically present in healthy men and the elderly. This might explain the susceptibility of men and the elderly to SARS-CoV-2 infection. Summary Immunological profile of severe COVID-19, characterized by altered immune cell populations and inflammatory plasma proteins is intrinsically present in healthy men and the elderly. Different age and sex groups show distinct seasonal responses to SARS-CoV-2.

Having emerged in China in December 2019, the coronavirus disease (COVID-19) caused by severe 37 acute respiratory syndrome coronavirus-2 (SARS-CoV-2) has become a major health crisis. As of April 38 2021, SARS-CoV-2 has led to over 135 million infections and more than 3 million deaths worldwide 39

[1]. 40

The most vulnerable groups are people older than 70 years old and adults with underlying health 41 conditions such as chronic respiratory problems and diabetes [2] . While age is the strongest 42 predictor of death from COVID-19, the sharp increase in fatality after 50 years of age is more critical 43 in men [3, 4] . Association of male sex with higher mortality has been consistently reported in 44 different populations [5] [6] [7] . 45

The factors underlying the impact of age and sex on the susceptibility to severe COVID-19 are, 46 however, incompletely understood. Most studies to date have focused on the differences between 47 7 responses. From those 50, we selected 20 individuals considering the optimal age and sex matching 108 of young-old and male-female comparisons. Therefore, cells isolated from 5 young males, 5 old 109 males, 5 young females, and 5 old females were included for an in vitro study assessing seasonality's 110 impact on the cytokine responses to SARS-CoV-2. Cohort demographics are shown in Supp. Table 3 . 111

Upon thawing, PBMCs were stimulated with heat-inactivated SARS-CoV-2 at a concentration of 112 3.3×10 3 TCID50/mL and heat-inactivated Influenza A H1N1 (California strain) at a concentration of 113 3.3×10 5 /mL for 24 hours and 5 days. Cytokine concentrations in the supernatants were measured 114 with DuoSet® ELISA kits (R&D Systems, MN, USA) according to the manufacturer's protocols. 115

Statistical analyses were performed unsing R 3.6.1 (www.R-project.org) and GraphPad Prism 8 117 (GraphPad Software Inc., CA, USA). After adjusting the data for the covariate sex using linear 118 regression, correlation of cell numbers or protein levels with age was done using Spearman's rank-119 order correlation (Figures 2 and 4) . Correction for sex was not applied in the heatmaps where two 120 sexes were analyzed separately. After adjusting the data for the covariate age, differential protein 121 expression or cell numbers between males and females was tested using the Mann 

Next, we investigated the differential abundance of the same immune cell populations influencing 147 COVID-19 severity in females vs. males. Only CD8 + effector memory T cells were significantly more 148 abundant in males ( Figure 3 ). Almost all cell types, including neutrophils, naïve CD4 + and CD8 + T cells, 149 memory T cells, class-switched memory B cell, and CD56 bright NK cell counts, were significantly higher 150 in females. The T cell types and CD56 bright NK cells, which are depleted in severe COVID-19 and 151 elderly healthy people, are also apparently less abundant in males. 152 153

We selected 28 proteins whose plasma concentrations have been associated with severe COVID-19 155 and correlated them with age in healthy cohorts. Circulating IL-6 concentrations increased with age 156 in healthy women, while IL-18 concentrations were higher in healthy men with advancing age in 157

Cohort 1 (Figure 4 ). IL-8 concentrations positively correlated with old age in both females and males; 158 however, it was only significant when sexes were combined. Among investigated chemokines, only 159 MCP-1 was positively correlated with age in females in Cohort 2. 160 TNF-family proteins in plasma also changed with increasing age: TNF and TNFB concentrations were 161 significantly lower in the circulation of older men in Cohort 1 and 2, respectively. TRANCE sharply 162 declined in older individuals, more strikingly in males, while TWEAK was positively correlated only in 163 females with advancing age. Moreover, aging in males was associated with elevated osteoprotegerin 164 (OPG) concentrations, and HGF concentrations exhibited a considerable age-dependent increase in 165 females in both cohorts. 166

In summary, several proteins in plasma that are increased in severe COVID-19 patients, such as IL-6, 167 IL-8, IL-18, MCP-1, OPG, and HGF, are more abundant in healthy elderly compared to young 168 individuals. Furthermore, proteins that are lower in severe COVID-19, e.g., TRANCE, decline with age. 169 170

We also compared plasma protein concentrations between sexes. It must be noted that Koeken et 172 al. previously reported that males and females of Cohort 2 exhibit differences in baseline levels of 173 many inflammatory markers [17] . Here we provide a more detailed analysis of COVID-19-related 174 proteins among those in Cohorts 1 and 2. We observed a similar sex-dependent trend in both 175 cohorts ( Figure 5 ). Only OPG and colony-stimulating factor-1 (CSF-1), related to severe COVID-19, 176

were significantly higher in women. On the other hand, plasma concentrations of other severity 177 markers such as IL-8, IL-18, MCP-1, MCP-2, CCL3, and CCL4 were all higher in men. Furthermore, 10 TRAIL, TWEAK, and TRANCE, which are all lower in COVID-19 patients in ICU, were more abundant in 179 males [9] . Males exhibited more anti-inflammatory proteins, e.g., PD-L1 and IL-10. Growth factors 180 HGF and SCF were also more abundant in male plasma. 181

These analyses show that most of the inflammatory mediators playing a role in infection severity are 182 already higher in the circulation of healthy men. 183

We hypothesized that circulating inflammatory proteins could be related to impaired cytokine 184 response against the virus. Therefore, we correlated SARS-CoV-2-induced in vitro cytokine 185 productions with baseline circulating protein concentrations in a sub-cohort of Cohort 1. Indeed, 186

PBMCs of the individuals with higher baseline plasma levels of MCP-2 and IL-8 produced more IL-1RA 187 against SARS-CoV-2 in vitro (Supp. Figure 1) . Furthermore, MCP-1 was negatively correlated with 188

IFNγ production. The data indicate that higher baseline plasma concentrations of MCP-2, IL-8, and 189 MCP-1 are associated with the inability to produce an optimal defense against SARS-CoV-2 infection. 190

Next, we investigated the impact of seasonality on the SARS-CoV-2-induced immune response and 193 assessed the contribution of age and sex. To this end, we selected 20 individuals from Cohort 1, for 194 which cryo-preserved cells collected at 4 roughly equidistant time points in one year were available. 195

We stimulated their PBMCs with heat-inactivated SARS-CoV-2 and influenza A H1N1. 196

We found that cytokine production upon SARS-CoV-2 stimulation did not substantially vary during 197

the year, considering all 20 individuals (Supp. Figure 2) . However, SARS-CoV-2-induced cytokine 198 production did differ for different age groups and sexes throughout the year. Cytokines of the IL-1 199 biological pathway were higher in the elderly: IL-1β production tended to be greater in the elderly 200 than in young individuals ( Figure 6A ), while SARS-CoV-2 induced more IL-1RA all-year-round in the 201 old individuals( Figure 6B ). Interestingly, IFNγ production upon stimulation with SARS-CoV-2 had a 202 11 different seasonal profile in the young and elderly: young individuals produced more IFNγ in the 203 summer and fall ( Figure 6C) . Remarkably, the elderly did not display this seasonal effect, with low 204 IFNγ production throughout the year. TNFα and IL-6 productions upon stimulation were similar in 205 the young and the elderly (Supp. Figure 3A-B) . In addition, IL-1β production in response to CoV-2 in spring and summer was higher in females; however, the average yearly response failed to 207 reach statistical significance ( Figure 6D ). IL-1RA and IFNγ production in males and females were 208 comparable ( Figure 6E-F) . 209

Stimulating PBMCs using another RNA virus, influenza H1N1, resulted in a similar pattern to SARS-210

CoV-2 stimulation regarding the age and sex effects. The elderly tended to produce more IL-1β and 211 IL-1RA, while young individuals could produce higher IFNγ amounts upon influenza stimulation 212 (Supp. Figure 4) . Similar amounts of IL-1β, IL-1RA, and IFNγ were induced in males and females on 213 average, with few exceptions (Supp. Figure 4D-F) . Of note, concentrations of circulating IL-7 and IFNγ, which are increased in severe COVID-19, are 267 similar in men and women. Therefore, T cell numbers being higher in women is unlikely due to IL-7-268 induced lymphopoiesis, whereas higher T cell numbers do not necessarily lead to more circulating 269

IFNγ. An overview of the age-and sex-dependent immune profiles in healthy individuals potentially 270 predisposing to severe COVID-19 upon infection is provided in Figure 7 . anti-viral response. The deficiency of IL-1β or its receptor causes higher viral load and mortality in 294 murine models [38] . Furthermore, genetic variants in IL1B contribute to influenza susceptibility in 295 humans [39] . We observed higher IL-1β production in women in response to SARS-CoV-2 in vitro, 296

arguing that an initial potent anti-viral defense is essential to prevent severe disease. Moreover, IL-297

1RA is an antagonist of IL-1 bioactivity, helping prevent excessive inflammation [40] . However, early 298 IL-1RA production in patients is associated with COVID-19 severity [31], while our finding of higher 299 IL-1RA production in the elderly suggests that IL-1RA might be hindering their ability to mount an 300 15 optimal immune response against SARS-CoV-2. Alternatively, high IL-1RA production with increasing 301 age might mirror the general inflammatory profile of elderly individuals. 302

In conclusion, our findings shed light on the immunological factors that might explain why men and 303 the elderly have a higher risk of developing severe COVID-19. These results also emphasize the 304 importance of the IL1β/IL-1RA axis and IFNγ in anti-SARS-CoV-2 response. We propose that plots of plasma levels of selected proteins in individuals aged less than 50, more than 50, and more 433 than 60. *p ≤ 0.05, **p ≤ 0.01, ***p ≤ 0.001, ****p ≤ 0.0001. ⚥ whole cohort, ♀ females (n Cohort1 = 434 229, n Cohort2 = 183), ♂ males (n Cohort1 = 223, n Cohort2 = 141). 

European Center for Disease Prevention and Control. COVID-19 situation update worldwide

Available at: www.ecdc.europa.eu/en/geographical-distribution-2019-ncov-cases

Risk factors for Covid-19 severity and fatality: a 333 structured literature review

Infection fatality risk for SARS-CoV-2 in 335 community dwelling population of Spain: nationwide seroepidemiological study

Sex-specific clinical characteristics and prognosis of coronavirus 338 disease-19 infection in Wuhan, China: A retrospective study of 168 severe patients

With Coronavirus Disease 2019 in a Series of Patients in Metropolitan Detroit

Clinical characteristics of 113 deceased patients with 344 coronavirus disease 2019: retrospective study

Factors Associated With Death in Critically Ill Patients 346

With Coronavirus Disease 2019 in the US

Sex differences in immune responses that underlie 348 COVID-19 disease outcomes

Dysregulated innate and adaptive immune 350 responses discriminate disease severity in COVID-19

Dysregulation of Immune Response in Patients With Coronavirus 352 2019 (COVID-19) in Wuhan, China

A single-cell atlas of the peripheral immune response in 354 patients with severe COVID-19

Cell type-specific immune dysregulation in severely ill 356 COVID-19 patients

Longitudinal immune profiling reveals key myeloid 358 signatures associated with COVID-19

Weather Variability and COVID-19 Transmission: A Review of Recent 360 Research

Rayan RA Seasonal variation and COVID-19 infection pattern: A gap from evidence to reality

Genetic Factors on T and B Cell Immune Traits

BCG vaccination in humans inhibits systemic 366 inflammation in a sex-dependent manner

Considering how biological sex 368 impacts immune responses and COVID-19 outcomes

Aging in COVID-19: Vulnerability, immunity and 370 intervention

Lymphocyte Subset Counts in COVID-19 Patients: A 372

Clinical and immunological features of severe and moderate 374 coronavirus disease 2019

B cell memory: building two walls of protection against 376 pathogens

Overcoming immune dysfunction in the 378 elderly: trained immunity as a novel approach

Adaptive immunity to SARS-CoV-2 and COVID-19

An inflammatory cytokine signature predicts 382

COVID-19 severity and survival

Hepatocyte Growth Factor Receptor c-Met Instructs T 384

Cell Cardiotropism and Promotes T Cell Migration to the Heart via Autocrine Chemokine 385

Plasma IP-10 and MCP-3 levels are highly associated with disease 387 severity and predict the progression of COVID-19

Severe COVID-19 Is Marked by a Dysregulated 390

IL-10, T cell exhaustion and viral persistence

Longitudinal COVID-19 profiling associates IL-1RA and IL-10 with 396 disease severity and RANTES with mild disease

Complex Immune Dysregulation in 398

COVID-19 Patients with Severe Respiratory Failure

The effects of estrogen on osteoprotegerin

RANKL, and estrogen receptor expression in human osteoblasts

RANKL biology: bone metabolism, the immune 403 system, and beyond

Influenza seasonality: 405 underlying causes and modeling theories

Among Young Men With Severe COVID-19

Cytokine Storm in COVID-19: The Current Evidence 409 and Treatment Strategies

Inflammasomes and viruses: cellular defence versus viral 411 offence

Genetic variants in IL1A and IL1B contribute to the susceptibility to 413 2009 pandemic H1N1 influenza A virus

Overview of the IL-1 family in innate inflammation and acquired immunity