key: cord-0903311-hws9a044
authors: Debisarun, P. A.; Struycken, P.; Dominguez-Andres, J.; Moorlag, S. J. C. F. M.; Taks, E.; Gössling, K. L.; Ostermann, P. N.; Müller, L.; Schaal, H.; ten Oever, J.; van Crevel, R.; Netea, M. G.
title: The effect of influenza vaccination on trained immunity: impact on COVID-19
date: 2020-10-16
journal: nan
DOI: 10.1101/2020.10.14.20212498
sha: 81cda81fbc84e071badd32b9de55cdf4ad0f8f1e
doc_id: 903311
cord_uid: hws9a044

Every year, influenza causes 290.000 to 650.000 deaths worldwide and vaccination is encouraged to prevent infection in high-risk individuals. Interestingly, cross-protective effects of vaccination against heterologous infections have been reported, and long-term boosting of innate immunity (also termed trained immunity) has been proposed as the underlying mechanism. Several epidemiological studies also suggested cross-protection between influenza vaccination and COVID-19 during the current pandemic. However, the mechanism behind such an effect is unknown. Using an established in-vitro model of trained immunity, we demonstrate that the quadrivalent inactivated influenza vaccine used in the Netherlands in the 2019-2020 influenza season can induce a trained immunity response, including an improvement of cytokine responses after stimulation of human immune cells with SARS-CoV-2. In addition, we found that SARS-CoV-2 infection was less common among Dutch hospital employees who had received influenza vaccination during the 2019/2020 winter season (RR = 0,61 (95% CI, 0.4585 - 0.8195, P = 0.001). In conclusion, a quadrivalent inactivated influenza vaccine can induce trained immunity responses against SARS-CoV-2, which may result in relative protection against COVID-19. These data, coupled with similar recent independent reports, argue for a beneficial effect of influenza vaccination against influenza as well as COVID-19, and suggests its effective deployment in the 2020-2021 influenza season to protect against both infections.

As of October 2020 there were over 37 million confirmed cases and one million deaths due to COVID-45 19 [1]. In many cases, SARS-CoV-2 infections only cause mild symptoms that resolve spontaneously. 46

However, in the elderly or in patients with underlying health conditions such as cardiovascular disease, 47 obesity, diabetes or pre-existing lung conditions, the disease is often more severe and potentially lethal. 48

Various complications can arise and include, pulmonary edema, severe pneumonia, acute respiratory 49 distress syndrome (ARDS) and thrombotic complications among others [2] . Due to the rapid spread and 50 the high clinical and socio-economic burden of COVID-19, the efforts to prevent and combat the disease 51 have been enormous. Despite the numerous ongoing developments and clinical trials to create specific 52 vaccines against the virus, the earliest expected vaccine is likely to be deployed at least 4-6 months 53 from now and is not expected to be readily available on a large scale [3, 4] . 54

In addition to COVID-19, we are still exposed to other threatening pathogens. In countries with 55 temperate climates, seasonal influenza outbreaks mainly occur in the winter, with the first cases starting 56 to appear as early as September. This leads to recurrent widespread mortality and morbidity causing 3 57 to 5 million cases of severe illness and 290.000 to 650.000 deaths per year [5] . During the previous 58 2017/2018 flu season in Europe, an estimated excess mortality of 125.000 deaths was measured [6] . 59

Because of the high morbidity and socioeconomic burden of recurrent influenza epidemics, vaccination 60 has become a key strategy in protecting high-risk individuals against the flu and is therefore a widely 61 promoted public health strategy [7] [8] [9] [10] . Despite the wide use of flu vaccines, there is contradictory 62 information on how the influenza vaccine might affect the outcome of other infections, including COVID-63 19 . The potential interaction between vaccines and infections other than their target disease, has 64 attracted a great deal of attention lately. It has been demonstrated that certain vaccines (such as bacillus 65

Calmette-Guérin (BCG), measles-containing vaccines, or oral polio vaccine) have strong beneficial 66 protective effects through long-term boosting of innate immunity, a process called trained immunity [11] . 67

In line with this, several recent studies suggested a potential beneficial effect of influenza vaccination 68 on susceptibility to COVID-19 [12, 13] . Despite earlier reports that have shown little or opposite effects 69 of influenza vaccines on heterologous infections in children [14] [15] [16] [17] [18] [19] . With the flu season on its way and 70 influenza vaccination campaigns starting off soon, it is paramount to clarify the exact effects of influenza 71 vaccination on the incidence and the disease course of COVID- 19. 72 . CC-BY-NC-ND 4.0 International license It is made available under a perpetuity.

is the author/funder, who has granted medRxiv a license to display the preprint in (which was not certified by peer review) preprint

The copyright holder for this this version posted October 16, 2020. . https://doi.org/10.1101/2020. 10.14.20212498 doi: medRxiv preprint Observational data healthcare workers 116

The Radboudumc hospital registration database of SARS-CoV-2 PCR-positive healthcare workers as 117 of June 1 st 2020 was consulted. The corresponding influenza vaccination status of healthcare workers 118 was retrieved from the database of the Department of Occupational Health and Safety of the hospital, 119 as well as the total influenza vaccination coverage rate (VCR) of the Radboudumc during the flu season 120 of 2019/2020. Additionally, SARS-CoV-2 positive employees were sent a questionnaire to assess 121 disease duration, severity and comorbidities. Disease duration was measured as the number of days 122 between the SARS-CoV-2 PCR test and the first day employees resumed their work. All hospital 123 employees are equally offered an influenza vaccination every year. However, SARS-CoV-2 testing in 124 the beginning of the pandemic was only available for employees who were indispensable for patientcare, 125 due to shortage of testing materials. Giving the observational nature of this study and the use of short 126 questionnaires only, no ethical approval was required. 127 . CC-BY-NC-ND 4.0 International license It is made available under a perpetuity.

is the author/funder, who has granted medRxiv a license to display the preprint in (which was not certified by peer review) preprint

The copyright holder for this this version posted October 16, 2020. . https://doi.org/10.1101/2020.10.14.20212498 doi: medRxiv preprint

Hospital database analysis was done using IBM SPSS statistics 25. To assess the association between 130 COVID-19 incidence and influenza vaccination status, a Chi-square test was used. No correction for 131 confounding was possible because no individual characteristics were available in the SARS-CoV-2 PCR 132 negative health care workers; only influenza vaccination status data for the entire group was known. 133

Missing values of other variables were left out of analysis. Cytokine and LDH concentrations were 134 analyzed using Wilcoxon matched-pairs signed rank test. Absolute cytokine concentrations were 135 determined after 24 hours in influenza vaccine stimulated conditions (with and without BCG) and 136 compared with unstimulated conditions. As a readout for trained immunity responses, influenza vaccine-137 primed (at day 6 LPS and SARS-CoV-2 restimulated) conditions were compared to RPMI conditions as 138 a negative control and BCG only conditions as a positive control. To assess synergistic effects on trained 139 immunity, combined influenza vaccine and BCG stimulated conditions were also compared. LDH values 140 were calculated to percentages of cell death, with lysed cells as a positive control. Data were analyzed 141 using Graphpad 8.02 (La Jolla, San Diego, CA, USA). A two-sided P-value below .05 (*) or below .01 142 (**) was considered statistically significant. Data are shown as means ± SEM. is the author/funder, who has granted medRxiv a license to display the preprint in (which was not certified by peer review) preprint

The copyright holder for this this version posted October 16, 2020. . https://doi.org/10.1101/2020.10.14.20212498 doi: medRxiv preprint collected after 24 h stimulation and showed no differences between conditions (Supplementary Figure  155 1), arguing against toxic effects of the vaccines. 156

In a separate set of experiments, 6 days after an initial 24h period of training of human PBMCs with the 159 quadrivalent inactivated influenza vaccine, the cells were restimulated with heat-inactivated SARS-CoV- is the author/funder, who has granted medRxiv a license to display the preprint in (which was not certified by peer review) preprint

The copyright holder for this this version posted October 16, 2020. . https://doi.org/10.1101/2020.10.14.20212498 doi: medRxiv preprint employees made up a slightly higher proportion of the SARS-CoV-2 positive group (79%) compared to 182 the SARS-CoV-2 negative group (70%) ( Table 1) In the present study, we demonstrate that the tetravalent inactivated influenza vaccine Vaxigrip Tetra® 205 induces trained immunity in an established in-vitro model, resulting in improved responsiveness of 206 immune cells to SARS-CoV-2 stimulation. In addition, Vaxigrip Tetra® amplifies the capacity of the BCG 207 vaccine to induce trained immunity [21] . These results are in line with a previous study from our group 208 in which 40 healthy volunteers received a trivalent influenza vaccine 14 days after receiving either BCG 209 or placebo [22] . Whilst BCG exhibited a broad increase of pro-inflammatory cytokines for a larger set of 210 . CC-BY-NC-ND 4.0 International license It is made available under a perpetuity.

is the author/funder, who has granted medRxiv a license to display the preprint in (which was not certified by peer review) preprint

The copyright holder for this this version posted October 16, 2020. . https://doi.org/10.1101/2020.10.14.20212498 doi: medRxiv preprint unrelated pathogens, the influenza vaccine showed more selective augmentation of cytokine responses 211 after ex-vivo restimulation of peripheral blood leukocytes, similar to our observations. In addition, we 212 complement these data with an epidemiological analysis that shows an inverse association between 213 influenza vaccination using a quadrivalent inactivated influenza vaccine, and COVID-19 incidence. This is the author/funder, who has granted medRxiv a license to display the preprint in (which was not certified by peer review) preprint

The copyright holder for this this version posted October 16, 2020. . https://doi.org/10.1101/2020.10.14.20212498 doi: medRxiv preprint

Overall, the majority of the ecological data are leaning to a possible protective effect of the influenza 242 vaccines. However, ecological studies also have several limitations. In this respect, there can be 243 systematic differences in how countries and areas report disease, mortality and exposures. For 244 example, in Italy any deceased person with a positive SARS-CoV-2 test is registered as a SARS-CoV-245 2 related death [28] . Besides this, information on confounding factors and effect modifiers can be 246 missing and correction for confounders sometimes is impossible, causing over-or underestimation of and healthcare facilities was performed [30] . 256

Many of these studies hypothesized that trained immunity may be the mechanism underlying these 258 observations [21, [31] [32] [33] . The most extensively studied vaccine that induces trained immunity is BCG, 259 which is currently being examined for its putative protective effects against COVID-19 duration and 260 severity in several clinical trials (NCT04328441, NCT04348370, NCT04327206, NL8609). Although this 261 property is usually assigned to live vaccines [34] , whether influenza vaccination can also induce trained 262 immunity was not known. In this study we also found that Vaxigrip Tetra® induces a trained immunity is the author/funder, who has granted medRxiv a license to display the preprint in (which was not certified by peer review) preprint

The copyright holder for this this version posted October 16, 2020. . https://doi.org/10.1101/2020.10.14.20212498 doi: medRxiv preprint fine-tune the inflammation and counteract excessive inflammation. In our experimental setup we 270 observed both, increased production of IL-6 paralleled with IL-1Ra, after stimulation with the influenza 271 vaccine and BCG. This confirms that these cytokines might contribute to keeping a balance in the 272 inflammatory status of the individual [36] . In addition, trained immunity is also known to be induced in 273 NK cells, which play an important role in containing viral infections, among others through their 274 production of IFN-γ. The increased IFN-γ produced after stimulation with the influenza vaccine and BCG, 275 as shown here, can indicate the long-term functional reprogramming of NK cells, which then activate 276 macrophages to further orchestrate the clearing of pathogens [37] . 277

Our study also has important limitations. The database analysis performed in this study did not allow to 279 correct for confounders, as we were not able to access data of individual influenza vaccination status in 280 SARS-CoV-2 negative employees. An important confounder is the difference in the rate of the direct 281 contact with patients between employees who developed SARS-CoV-2 or not, since this was the 282 variable unevenly distributed among the two groups, with less direct patient contact in SARS-CoV-2 283 negative employees. However, earlier studies have reported that most of the SARS-CoV-2 infections in 284 hospital personnel occur in society, rather than through patient contact in the hospitals [38] [39] [40] . In conclusion, we provide observational data suggesting a potentially protective role of the quadrivalent 293 inactivated influenza vaccine on COVID-19 incidence. In addition, we report first insights in the 294 immunological mechanisms underlying these observations. We show that a quadrivalent inactivated 295 influenza vaccine can induce trained immunity, and the plausible mechanisms through which an 296 enhanced antiviral state is acquired after vaccination. Considering these data, and with at least several 297 months more needed until a specific SARS-CoV-2 vaccine is available, influenza vaccination may 298 . CC-BY-NC-ND 4.0 International license It is made available under a perpetuity.

is the author/funder, who has granted medRxiv a license to display the preprint in (which was not certified by peer review) preprint

The copyright holder for this this version posted October 16, 2020. . https://doi.org/10.1101/2020.10.14.20212498 doi: medRxiv preprint contribute not only to reduction of influenza but also to the COVID-19-related burden on the healthcare 299 system. While our data show that earlier influenza vaccination is safe in relation to a later SARS-CoV-2 300 infection, we recommend vaccination in the absence of active COVID-19, because of the theoretical 301 possibility to induce a cytokine storm by an enhanced immune response if the vaccine is given during 302 an active infection. Additionally, our data suggest that the presence of a previous BCG vaccination prior 303 to the influenza vaccination could lead to enhanced responses and improved protection, raising the 304 possibility to conduct clinical trials to asses this hypothesis. is the author/funder, who has granted medRxiv a license to display the preprint in (which was not certified by peer review) preprint

The copyright holder for this this version posted October 16, 2020. . https://doi.org/10.1101/2020.10.14.20212498 doi: medRxiv preprint is the author/funder, who has granted medRxiv a license to display the preprint in (which was not certified by peer review) preprint

The copyright holder for this this version posted October 16, 2020. . https://doi.org/10.1101/2020.10.14.20212498 doi: medRxiv preprint is the author/funder, who has granted medRxiv a license to display the preprint in (which was not certified by peer review) preprint

The copyright holder for this this version posted October 16, 2020. . https://doi.org/10.1101/2020.10.14.20212498 doi: medRxiv preprint Positive correlation between influenza vaccination and COVID-19 incidence in Europe (r = 0.66 ± 0.13, P = .000017) and the USA (r = 0.50 ± 0.14, P <.05).

Positive correlation between influenza vaccination and COVID-19 attributable mortality in Europe (r = 0.68 ± 0.13, p = 0.000006) and the USA (r = 0.50 ± 0.14, P <.05).

Crosssectional study . CC-BY-NC-ND 4.0 International license It is made available under a perpetuity.

is the author/funder, who has granted medRxiv a license to display the preprint in (which was not certified by peer review) preprint

The copyright holder for this this version posted October 16, 2020. . https://doi.org/10.1101/2020.10.14.20212498 doi: medRxiv preprint to the cells. In other wells, the vaccine was combined with BCG (5µg/mL) and control wells only contained RPMI. After 24 h the supernatant was collected and cells were washed. Fresh RPMI was added to the cells and left to incubate for another 5 days. On day 6, the PBMCs were restimulated with LPS (10 ng/mL) or inactivated SARS-CoV-2 (40x diluted, TCID50/mL 6.67*10e4) for another 24 h. On day 7, the supernatants were collected and stored for cytokine measurements. is the author/funder, who has granted medRxiv a license to display the preprint in (which was not certified by peer review) preprint

The copyright holder for this this version posted October 16, 2020. . https://doi.org/10.1101/2020.10.14.20212498 doi: medRxiv preprint is the author/funder, who has granted medRxiv a license to display the preprint in (which was not certified by peer review) preprint

The copyright holder for this this version posted October 16, 2020. is the author/funder, who has granted medRxiv a license to display the preprint in (which was not certified by peer review) preprint

The copyright holder for this this version posted October 16, 2020. . https://doi.org/10.1101/2020.10.14.20212498 doi: medRxiv preprint

trained with several dilutions of the inactivated quadrivalent vaccine (10x, 50x, 100x, 400x) with or without BCG (5 μg/mL), restimulated at day 6 with LPS (10 ng/ml) or SARS-CoV-2 ((40x diluted, TCID50/mL 6.67*10e4) for 24 h and compared to RPMI conditions. Vaxigrip Tetra® amplified IL-1Ra (A, B) and IL-6 (C, D) responses after SARS-CoV-2 or LPS restimulation, compared to stimulation of naïve PBMCs. Training with BCG diminished the training effect of Vaxigrip Tetra® on IL-1Ra production, but increased IL-6 responses even more. A significant increase in IFN-γ production was seen after training of PBMCs with 10x dilutions of the influenza vaccine and restimulation with SARS-CoV-2. The combination of the influenza vaccine with BCG induced even more IFN-γ than BCG alone (E, F). No significant increase in TNF-α by Vaxigrip Tetra® was observed after restimulation with SARS-CoV-2 (G). Vaxigrip Tetra® enhanced TNF production upon LPS restimulation, but only in combination with BCG (H). (Wilcoxon matched pairs signed rank test, n =7-9, * = P < .05, ** = P <.01)

. CC-BY-NC-ND 4.0 International license It is made available under a perpetuity.

is the author/funder, who has granted medRxiv a license to display the preprint in (which was not certified by peer review) preprint

The copyright holder for this this version posted October 16, 2020. . https://doi.org/10.1101/2020.10.14.20212498 doi: medRxiv preprint is the author/funder, who has granted medRxiv a license to display the preprint in (which was not certified by peer review) preprint

The copyright holder for this this version posted October 16, 2020. .

Epidemiological and clinical characteristics of 99 cases of 2019 novel coronavirus pneumonia in Wuhan, China: a descriptive study

The SARS-CoV-2 Vaccine Pipeline: an Overview

The COVID-19 vaccine development landscape

European all-cause excess and influenza-attributable mortality in the 2017/18 season: should the burden of influenza B be reconsidered?

Human Influenza Virus Infections. Semin Respir Crit Care Med

World Health Organization Fact Sheet on Influenza

Estimated influenza illnesses and hospitalizations averted by vaccination--United States, 2013-14 influenza season

Prevention and Control of Influenza with Vaccines: Recommendations of the Advisory Committee on Immunization Practices

Defining trained immunity and its role in health and disease

Exploratory analysis of immunization records highlights decreased SARS-CoV-2 rates in individuals with recent non-COVID-19 vaccinations

Relationship between Influenza Vaccination Coverage Rate and COVID-19 Outbreak: An Italian Ecological Study. Vaccines (Basel)

Vaccine Effectiveness Against Laboratory-confirmed Influenza in Healthy Young Children: A Case-Control Study

Assessment of temporally-related acute respiratory illness following influenza vaccination. Vaccine

Increased risk of noninfluenza respiratory virus infections associated with receipt of inactivated influenza vaccine

Pilot comparative study on the health of vaccinated and unvaccinated 6-to 12-year old U.S. children

Epidemiology of respiratory viral infections in children enrolled in a study of influenza vaccine effectiveness. Influenza Other Respir Viruses

Influenza vaccination is not associated with detection of noninfluenza respiratory viruses in seasonal studies of influenza vaccine effectiveness

SARS-CoV-2 targets neurons of 3D human brain organoids

Bacille Calmette-Guerin induces NOD2-dependent nonspecific protection from reinfection via epigenetic reprogramming of monocytes

BCG Vaccination Enhances the Immunogenicity of Subsequent Influenza Vaccination in Healthy Volunteers: A Randomized, Placebo-Controlled Pilot Study

Influenza Vaccination and COVID19 Mortality in the USA. medRxiv, 2020. 25. MC., A., Correlation of influenza vaccination and the COVID-19 severity. SSRN, 2020. 26. Lisewski and A. Martin, Association between Influenza Vaccination Rates and SARS-CoV-2 Outbreak Infection Rates in OECD Countries

Case-Fatality Rate and Characteristics of Patients Dying in Relation to COVID-19 in Italy

Could seasonal influenza vaccination influence COVID-19 risk? medRxiv

Inactivated Trivalent influenza vaccien is associated with lower mortality among Covid-19 patients in Brazil. medRxiv

Trained immunity: A program of innate immune memory in health and disease

Glutaminolysis and Fumarate Accumulation Integrate Immunometabolic and Epigenetic Programs in Trained Immunity

Immunometabolic Pathways in BCG-Induced Trained Immunity

Trained innate immunity as underlying mechanism for the long-term, nonspecific effects of vaccines

IL-6 in inflammation, immunity, and disease

The role of the interleukin-1 family in trained immunity

BCG-induced trained immunity in NK cells: Role for non-specific protection to infection

Incidence of Nosocomial COVID-19 in Patients Hospitalized at a Large US Academic Medical Center

Nosocomial transmission of COVID-19: a retrospective study of 66 hospitalacquired cases in a London teaching hospital

Reducing nosocomial transmission of COVID-19: implementation of a COVID-19 triage system