Trained immunity is induced in humans after immunization with an adenoviral vector COVID-19 vaccine

Background Heterologous effects of vaccines are mediated by “trained immunity,” whereby myeloid cells are metabolically and epigenetically reprogrammed, resulting in heightened responses to subsequent insults. Adenovirus vaccine vector has been reported to induce trained immunity in mice. Therefore, we sought to determine whether the ChAdOx1 nCoV-19 vaccine (AZD1222), which uses an adenoviral vector, could induce trained immunity in vivo in humans. Methods Ten healthy volunteers donated blood on the day before receiving the ChAdOx1 nCoV-19 vaccine and on days 14, 56, and 83 after vaccination. Monocytes were purified from PBMCs, cell phenotype was determined by flow cytometry, expression of metabolic enzymes was quantified by RT-qPCR, and production of cytokines and chemokines in response to stimulation ex vivo was analyzed by multiplex ELISA. Results Monocyte frequency and count were increased in peripheral blood up to 3 months after vaccination compared with their own prevaccine controls. Expression of HLA-DR, CD40, and CD80 was enhanced on monocytes for up to 3 months following vaccination. Moreover, monocytes had increased expression of glycolysis-associated enzymes 2 months after vaccination. Upon stimulation ex vivo with unrelated antigens, monocytes produced increased IL-1β, IL-6, IL-10, CXCL1, and MIP-1α and decreased TNF, compared with prevaccine controls. Resting monocytes produced more IFN-γ, IL-18, and MCP-1 up to 3 months after vaccination compared with prevaccine controls. Conclusion These data provide evidence for the induction of trained immunity following a single dose of the ChAdOx1 nCoV-19 vaccine. Funding This work was funded by the Health Research Board (EIA-2019-010) and Science Foundation Ireland Strategic Partnership Programme (proposal ID 20/SPP/3685).


Background
Heterologous effects of vaccines are mediated by 'trained immunity' whereby myeloid cells are metabolically and epigenetically reprogrammed resulting in heightened responses to subsequent insults. Adenovirus vaccine vector has been reported to induce trained immunity in mice. Therefore, we sought to determine if the ChAdOx1 nCoV-19 vaccine (AZD1222), which uses an adenoviral vector, could induce trained immunity in vivo in humans.

Methods
Ten healthy volunteers donated blood on the day before receiving the ChAdOx1 nCoV-19 vaccine and on day 14, 56 and 90 post vaccination. Monocytes were purified from PBMC; cell phenotype was determined by flow cytometry, expression of metabolic enzymes were quantified by RT-qPCR and production of cytokines and chemokine in response to stimulation ex vivo were analyzed by multiplex ELISA.

Results
Monocyte frequency and count were increased in peripheral blood up to 3 months post vaccination compared with their own pre-vaccine control. Expression of HLA-DR, CD40 and CD80 was enhanced on monocytes for up to 3 months following vaccination. Moreover, monocytes had increased expression of glycolysis-associated enzymes 2 months post vaccination. Upon stimulation ex vivo with unrelated antigens, monocytes produced increased IL-1β, IL-6, IL-10, CXCL1, and MIP-1α, and decreased TNF, compared with pre-vaccine controls. Resting monocytes produced more IFN-γ, IL-18, and MCP-1 up to 3 months post vaccination compared with pre-vaccine controls.

Conclusion
These data provide evidence for the induction of trained immunity following a single dose of the ChAdOx1 nCoV-19 vaccine.

Introduction
The vaccine effort against SARS-CoV-2 was unprecedented and highly successful, largely due to vast body of pre-existing R&D in the field. Vaccine design focused on developing adaptive immune responses against SARS-CoV-2 through the production of neutralizing antibodies from B cells and through the generation of memory T cells which elicit protection against SARS-CoV-2 severity and mortality. The innate immune response is critical to the orchestration of protective adaptive immune responses post vaccination. A growing body of evidence now indicates that the mammalian innate immune response has a capacity for a type of memory termed "trained immunity" (1). The role that trained immunity plays in vaccine efficacy remains unknown. However, trained immunity mediates the non-specific protective effects of live attenuated vaccines, such as the BCG vaccine against tuberculosis (2)(3)(4)(5) which is known to reduce all-cause mortality in infants (6)(7)(8)(9). This occurs because the vaccine induces epigenetic and metabolic rewiring of monocytes which leaves them primed to respond in a heightened manner when they are subsequently stimulated (2,3,10). Importantly, this effect outlives the short timeframe of immune activation subsequent to vaccination and has been attributed to changes in the bone marrow hematopoietic stem cell niche which result in enhanced myelopoiesis and the egression of myeloid cells that are epigenetically and metabolically reprogrammed due to trained immunity (5,11). Comparisons between randomized clinical trials of COVID-19 vaccines suggest that adenovirus vector based vaccines may have non-specific protective effects resulting in significantly reduced all-cause mortality and non-COVID, non-accident mortality compared to mRNA based vaccines (12).
Notably, an intranasal adenoviral vaccine vector has been shown to induce trained immunity in the airways of mice, and reduced disease burden when the mice were subsequently infected with S. pneumoniae or M. tuberculosis (Mtb) (13,14). Furthermore, preliminary studies showed that vaccination of humans with an aerosolized adenovirus-vectored tuberculosis vaccine induced persisting transcriptional changes in alveolar macrophages (15). We sought to determine if the ChAdOx1 nCoV-19 vaccine (produced by AstraZeneca), which uses a replication-deficient simian adenovirus vector could induce trained immunity in vivo in humans. We drew venous blood from healthy adults the day prior to vaccination and 14-, 56-and 90-days post vaccination with a single dose of the ChAdOx1 nCoV-19 vaccine. Monocyte phenotype and function was assessed at these time points post vaccination and compared with their own pre-vaccine baseline control sample.
Our results indicate that the ChAdOx1 nCoV-19 vaccine may enhance myelopoiesis up to 3 months post vaccination. Our data indicates that the vaccine induced metabolic reprogramming in monocytes, which was sustained at 2 months post vaccination. Moreover, monocytes exhibited enhanced antigen presentation functions and had increased capacity to produce key cytokines and chemokines in response to subsequent unrelated stimuli. Taken together, these data indicate that the ChAdOx1 nCoV-19 vaccine can induce prolonged innate immune activation, with many features of trained immunity.

Results
Blood from healthy donors who received the ChAdOx1 nCoV-19 vaccine was collected the day before (day -1) and 14-, 56-and 90-days post vaccination ( Figure 1A). Donors were excluded from participating at any time point if they had received another vaccine in the last 3 months or if they knowingly got infected with COVID-19 or another infectious agent ( Figure   1B). The age range of the participants was 23-35 years old ( Figure 1C) and included 6 female and 4 male donors ( Figure 1D). PBMC were isolated and monocytes were enriched using a hyperosmotic percoll gradient. Cells were examined by flow cytometry and real time RT-qPCR. Monocytes stimulated ex vivo to assess their ability to respond to unrelated stimuli pre and post vaccination. Concentrations of soluble inflammatory mediators present in the supernatants 24 h post stimulation were determined by multiplex ELISA. Increased myelopoiesis and modulation of myeloid progenitors is an essential mechanism underpinning trained immunity (5,11,16,17) and mechanistically accounts for the longevity of the effects since training outlasts the short lifespan of the original activated monocytes.
Reprogramming of hematopoietic stem cells in the bone marrow results in monocytes with enhanced function against related and unrelated infections (5,11,(16)(17)(18). To assess the effect of ChAdOx1 nCoV-19 vaccination on myelopoiesis, the absolute number and relative frequency of monocytes in the peripheral blood was assessed. Cells were Fc blocked and stained fluorochrome-conjugated antibodies specific for CD14, CD16 and CD68 prior to acquisition by flow cytometry. Monocytes were gated as shown in the gating strategy ( Figure   2A).
We examined the effect of the ChAdOx1 nCoV-19 vaccine on the frequencies of the total monocyte (CD14 + ) population and non-classical/intermediate monocytes that co-express CD16 (CD14 + CD16 + ), which exert effector function in humans in vivo during infection and inflammation (19,20). In addition to being a marker for monocytes, CD14 is as a coreceptor for TLR-4 and facilitates cellular responses to LPS or gram-negative bacteria (21). Cell surface bound CD14 is cleaved and released as a scavenger receptor when monocytes are activated (22). The median fluorescence intensity (MFI), a surrogate for relative protein expression, of CD14 was significantly decreased on day 14 post vaccination but returns to baseline thereafter ( Figure 2B). This may be indicative of monocyte activation, suggesting that there is prolonged innate immune activation 14 days post vaccination, which returns to homeostasis by day 56.
The absolute number of monocytes was determined by multiplying the cell count obtained from the percoll monocyte enrichment process by the frequency of CD14 + cells present in the enriched population, as determined by flow cytometry ( Figure 2C). The monocyte frequency in the blood was calculated by dividing the absolute number of monocytes by the absolute number of PBMC ( Figure 2D). The absolute numbers ( Figure 2C) and the frequencies ( Figure   2D) of monocytes in the peripheral blood were significantly increased at all time points post vaccination compared with pre-vaccine controls. The absolute numbers and frequencies of monocytes co-expressing CD14 and CD16 was significantly increased at all time points post vaccination compared with pre-vaccine controls (Figure 2E &F). Cumulatively, these data suggest that myelopoiesis is enhanced post vaccination, and maintained for up to 3 months. Antigen presentation by innate cells to activate the adaptive immune system is a crucial event in the vaccination process. In animal models, trained monocytes show an increased expression of MHC-II and the cell costimulatory molecules such as CD80 and CD86 (13,23). Human dendritic cells matured with β-glucan, a fungal cell wall component also known to induce trained immunity, were also shown to have enhanced HLA-DR, CD40, CD80 and CD86 expression (24). Furthermore, human monocytes trained with C. albicans had increased RNA levels of HLA-DRB1, CD40, CD80 and CD86 (25). There is evidence to suggest that trained cells may enhance T cells responses (26), therefore, we sought to determine the effect of the Both in vitro and in vivo studies have shown that metabolic reprogramming is an essential event in the induction of trained immunity, which is mechanistically associated with enhanced protection against unrelated pathogens (10,(27)(28)(29)(30)(31). We sought to determine if the ChAdOx1 nCoV-19 vaccine effected the cellular metabolism of monocytes over time. Therefore, we examined the transcript expression levels of key glycolytic enzymes; GPI, PFKFB3, GAPDH and PKM2 by real time RT-qPCR. These genes encode enzymatic proteins throughout the glycolytic pathway ( Figure 4A).
Vaccination did not significantly alter the expression of GPI in monocytes ( Figure 4B), however, expression of PFKFB3 was significantly increased on day 14, 56 and 90 post vaccination compared with pre-vaccine controls ( Figure 4C). GAPDH and PKM2 were significantly increased on day 14 and 56 post vaccination but had returned to pre-vaccination levels by day 90 ( Figure 4D&E).
Since the production of IL-1β is associated in the literature with enhanced glycolysis (32)(33)(34), we sought to determine if the enhanced expression of glycolytic enzymes in monocytes post vaccination, resulted in enhanced production of IL-1β. In addition, trained immunity has been shown to increase the ability of myeloid cells to produce IL-1β. For example, human monocytes trained with β-glucan in vitro increased the concentration of intracellular IL-1β in response to Mtb stimulation and murine bone marrow-derived macrophages from BCG vaccinated mice showed increased expression of IL1B (5,17). This is likely due to the increase in glycolysis induced by trained immunity (10,30). Furthermore, airway macrophages from mice vaccinated intranasally with an adenovirus-vectored Mtb vaccine produced significantly more IL-1β in response to LPS and Mtb whole cell lysate compared with unvaccinated mice (14).
Monocytes were isolated and stimulated ex vivo with irradiated Mtb (iH37Rv) for 24 hours.
The concentration of IL-1β was determined by multiplex ELISA ( Figure 4F&G). There was no difference in the detection of IL-1β in unstimulated cells over time ( Figure 4F). Monocytes stimulated with irradiated Mtb produced significantly more IL-1β on day 56 post vaccination compared with their pre-vaccine controls ( Figure 4G).
To determine if metabolic reprogramming was restricted to glycolysis, or if vaccination could affect oxidative phosphorylation, we quantified the expression of ATB5B (a gene that encodes ATP synthase and used as a marker of oxidative phosphorylation; Figure 4H), in monocytes pre-and post-vaccination. Expression of ATB5B was significantly increased on day 14 and significantly decreased by day 90 post vaccination compared with pre-vaccine controls ( Figure   4H).
Taken together, these data indicate that monocytes are metabolically reprogrammed towards enhanced glycolysis for 2 months in people vaccinated with a single dose of the ChAdOx1 nCoV-19 vaccine. Moreover, these reprogrammed monocytes have increased production of the key proinflammatory cytokine IL-1β at day 56 post vaccination compared with pre-vaccine controls, demonstrating that these cells have increased functional outputs which have previously been associated with enhanced capacity for glycolysis.

vaccination.
The effects of trained immunity on exerting non-specific protection against infection are mediated by functional production of cytokines and chemokines that expedite and amplify the subsequent immune response (2,4,13,17). Monocytes were enriched from PBMC using a percoll gradient and further purified by adherence to plastic for 1 h. Cells were stimulated ex IL-6 is a pleiotropic cytokine with a critical role in the induction of inflammation, including acting as a pyrogen. As such, it is an early mediator of the inflammatory response to a diverse range of insults including bacterial and viral infections. The production of IL-6 was previously shown to be enhanced in monocytes that exhibit trained immunity (2,4,17), therefore we examined the production of IL-6 in response to stimulation in monocytes from people who had undergone vaccination compared with their pre-vaccine control monocytes ( Figure 5A).
Monocytes stimulated with irradiated Mtb had significantly increased production of IL-6 at all time points up to 3 months (90 days) post vaccination compared with pre-vaccine controls.
Similarly, monocytes stimulated with TLR-agonists LPS or Pam3Csk4 produced significantly more IL-6 than their pre-vaccine controls at day 14 and day 56 post vaccination ( Figure 5A).
Enhanced TNF is also associated in the literature with trained immunity induced by BCG or βglucan (2,4,17) and is known to be seminal in the response to Mtb infection (35). Interestingly, our data indicated that production of TNF in response to stimulation with Mtb was significantly decreased post vaccination ( Figure 5B). In addition, cells stimulated ex vivo with LPS

Discussion
These data indicate that the ChAdOx1 nCoV-19 vaccine induced trained immunity in humans in vivo. Together with evidence indicating that adenovirus vector based, but not mRNA based, COVID-19 vaccines may exert non-specific protective effects in humans (12) and that empty adenovirus vector can induce trained immunity in mice (13), we postulate that the effects observed may be caused, at least in part, by the vector rather than the payload. In support of this, intranasal empty adenovirus vector vaccine and adenovirus vectors loaded with Mtb or SARS-CoV-2 antigens induced trained immunity in airway macrophages identified by their high expression of MHC-II and increased glycolytic metabolism (13,36,37), similar to the phenotype we observed in peripheral blood monocytes post intramuscular vaccination. Our data demonstrates that monocyte absolute numbers and frequencies were increased in the blood up to 90 days post vaccination, indicative of a preferential skewing of hematopoiesis towards myelopoiesis, which is associated with the induction of trained immunity in animal models (11,16,38) and in humans (18). Notably, the fluorescence intensity, which is directly  (13,36). In contrast, gene expression of HLA-DR was downregulated in people 2 weeks and 3 months post BCG vaccination compared with the pre-vaccine control (39), suggesting that the features of trained innate immunity may differ according to the inducing stimulus. We observed increased expression of co-stimulatory molecules CD40, CD80 and CD86 post vaccination. However, CD80 and CD86 were not enhanced in macrophages from the lungs of mice exposed to the adenovirus vector (13). This divergence may be due to human versus mouse variation or due to tissue-specific versus peripherally induced trained immunity. When we segregated the data into monocyte Furthermore, production of IL-1β in response to stimulation with Mtb was significantly increased at day 56 but not at day 14, suggesting that trained immunity may not yet be induced at this earlier time point. By two months (56 days), the cells exhibit features consistent with trained immunity and then by 3 months (90 days) post vaccination, this effect was waning in some of the datasets. However, many of the phenotypes are maintained at day 90, so it is plausible that the effects of trained innate immunity are longer lasting but could not be analyzed in this study without the confounding factor of the booster dose. Longitudinal studies using single dose adenoviral vector vaccines are therefore warranted to assess the longevity of these effects on the innate immune responses.
Strikingly, resting monocytes taken from donors 14-and 56-days post vaccination produced IFN-γ, IL-18 and MCP-1 in culture without stimulation. It is difficult to conclude whether this is in keeping with monocytes that have undergone trained immunity since many of the published data present changes in cytokine production as fold-change or do not show the complete data set with unstimulated controls (2)(3)(4)17). However, trained immunity induced in human monocytes in vitro by β-glucan resulted in elevated gene expression of MCP-1 and CCL18 in resting cells on day 6 post training, which was then further elevated upon restimulation with Mtb (17). In addition, the significantly elevated GM-CSF observed in resting monocytes on day 14 post vaccination may also play a key role in the enhancement of myelopoiesis and reprogramming towards trained immunity, in keeping with its role in inducing trained immunity in the bone marrow of mice exposed to β-glucan (11). The increased production of IL-18 observed in resting monocytes, in addition to the increased IFN-γ, suggests that NK cell function and activation may be enhanced in this setting (46,47). Moreover,

features of trained immunity can be induced in NK cells by BCG vaccination, viral infection
or exposure to cytokines in vitro including IL-18 (48)(49)(50)(51). We therefore postulate that NK cells may play a key role in trained immunity induce by ChAdOx1 nCoV-19 vaccination and further studies are warranted to determine this.
Whilst these elevated proinflammatory cytokines may be beneficial in mediating non-specific immunity against other infections, it is also plausible that this inflammation may be deleterious in certain contexts. For example, MCP-1 (also known as CCL2) regulates the migration of monocytes, and other immune cells, from the blood into the tissue. It has previously been associated with inflammatory atherosclerosis which is mediated, at least in part, by trained immunity (29,52). However, our data indicates that the ability to produce regulatory IL-10 increases over time, which may protect against excessive inflammation.
Our data indicates that intramuscular delivery of the ChAdOx1 nCoV-19 vaccine can induce trained immunity in the peripheral blood. We do not know whether these monocytes with enhanced function would exert expedited and elevated responses to subsequent infections in tissues such as the lung. Work in animal models shows that intranasally delivered adenoviral vaccine vector can induce trained immunity in tissue resident alveolar macrophages which results in enhanced protection against a subsequent challenge (13,14,37). This tissue specific trained immunity was mediated by IFN-γ (13). Further evidence in BCG treated mice indicates that IFN-γ is required to induce trained immunity in the bone marrow hematopoietic stem cell niche (5). In keeping, IFN-γ is required for the induction of trained immunity in human monocytes stimulated with BCG (53). Taken together, these data provide strong rationale suggesting that the elevated IFN-γ observed ex vivo in human monocytes post ChAdOx1 nCoV-19 vaccination may mechanistically mediate the increased myelopoiesis and the effects of trained immunity observed herein. In addition, metabolic reprogramming induced by ChAdOx1 nCoV-19 vaccination is likely to result in epigenetic changes that mediate the enhanced monocyte function in response to subsequent unrelated stimuli (54).
Trained immunity induced by BCG is dependent on NOD-2 in vivo in humans (2). IL-6 and MCP-1 produced in response to adenovirus vectors are significantly reduced in NOD-2 deficient mice (55). Furthermore, adenovirus synergizes with the NOD-2 agonist muramyl dipeptide resulting in increased IL-1β and TNF production (56). Therefore, we postulate that NOD-2 may have a role in mediating trained innate immunity induced by the ChAdOx1 nCoV-19 vaccine.
Since downstream responses to both TLR2 and TLR4 agonists are enhanced post ChAdOx1 nCoV-19 vaccination, we speculate that common signaling molecules in these pathways may be mechanistically involved in the induction of trained immunity such as MyD88, NFκB and activation of NLRP3 (required for the caspase-dependent processing of IL-1β and IL-18, both of which are increased in our data). In summary, we suggest that the mechanisms underpinning the trained innate immunity elicited by ChAdOx1 nCoV-19 vaccination may include a role for IFN-γ, epigenetic reprogramming, NOD-2 and innate signaling transducers.
Our data may aid in the design of novel vaccination strategies that combine traditional intramuscular routes with intranasal/airway delivery of the vaccine. This may result in trained immunity in circulating monocytes and other myeloid cells which may have enhanced efficacy against respiratory infection in combination with trained tissue resident alveolar macrophages.
Since both the alveolar macrophage and the infiltrating monocyte derived macrophage have distinct but important roles to play during infection, supporting the functions of both populations may have increased benefit in host defense. In addition, understanding the kinetics of enhanced innate immune function after the priming dose of vaccination may be crucial to optimize the timing of subsequent booster doses. Real world evidence from the clinical trials for ChAdOx1 nCoV-19 indicate that the booster dose induced better efficacy when administered more than 8 weeks post priming compared to boosters given less than 6 weeks after the initial immunization (57). The molecular mechanism behind this observation is unknown, however, there is evidence indicating that kinetics of germinal center formation and increased selection of B cells with higher antigen affinity occurs when the boosting time frame is delayed (58). The increased antigen presenting function and cytokine/chemokine profile we observed is likely to enhance adaptive memory responses to booster vaccination. Our data indicates that prolonged innate immune activation occurs at day 14 (2 weeks) and then changes to a phenotype more consistent with trained innate immunity by day 56 (8 weeks). Therefore, we postulate that the kinetics of innate immune function may be a significant contributor to the improved vaccine booster efficacy observed after 8 weeks compared to boosters given earlier than 6 weeks post priming. We, and others (59), propose that tracking innate immune responses in addition to traditional B and T cell responses after the priming vaccination may therefore help to identify the optimal vaccine regimen.
Our data suggests that trained immunity may be induced in humans by other vaccines using similar adenovirus vector-based platforms being developed in murine models for TB and nextgeneration COVID-19 vaccines (14,37). However, we postulate that trained innate immunity induced by distinct adenoviral vectors with different target antigens will likely induce differential innate immune profiles. Therefore, further investigations are warranted to allow us to specifically understand the types of trained innate immunity induced by discrete immune stimuli and the subsequently elicited protective versus potential pathogenic effects to allow us to harness the potential of trained innate immunity towards clinical benefit.

Study limitations
Whilst our study provides evidence that the ChAdOx1 nCoV-19 vaccine induced phenotypic and functional changes in myeloid cells, consistent with those previously reported in the literature in other settings of trained immunity, we did not have the capacity to undertake epigenetic profiling of monocytes pre and post vaccination.
We cannot definitively conclude if this altered monocyte function in response to vaccination was caused by the adenovirus vector or the SARS-CoV2 spike protein payload or a combination of both. However, there is evidence in the literature to suggest that empty adenovirus vectors can induce trained immunity (13), which allows us to postulate that the vector likely plays a key role in the induction of trained immunity in our study. Moreover, since we examined the effects on monocytes in a naïve population up to 3 months after the single, initial vaccine dose (and prior to the booster), it is unlikely that our results are confounded by an adaptive memory response.
Since exposure to microbes induces trained immunity, the study of these effects in a human population is confounded due to continuous environmental exposure to pathogens. Despite our small cohort, our data illustrated statistically significant effects, indicating that this study was appropriately powered. In addition, our study design also benefits from the longitudinal nature of the sample collection whereby every donor is their own control (pre and post vaccine). We began our study in Dublin, Ireland, in early March 2021 (day -1) during a period of extended social restrictions (from late December 2020 to June 2021). All other blood draws also occurred within this period (day 90 was drawn at the end of May 2021). Therefore, the probability of our donors being exposed to SARS-CoV2 (or other infections) is likely to be lower than that expected normally during longitudinal studies, however, we cannot definitively rule out the occurrence of an asymptomatic infection in our donors. The Health Protection Surveillance Centre indicates that during that period, from available data, there were very low rates of influenza and RSV suggesting that the social restrictions may have reduced the background confounding variables of exposure to other infections in a human population (60).
We cannot rule out that seasonal variation may contribute somewhat to the changing immune phenotypes observed over time in our study. However, 5 out of 10 of our volunteers took regular multi-vitamin supplements, with 3 of these 5 donors specifically taking vitamin D supplements. We did not observe substantial spread in any of our data that may indicate that seasonal variation was mitigated with vitamin D supplementation.
Trained innate immunity and heterologous effects of vaccines have a sex-differential (61,62).
Although 4 males and 6 females were included in the study, when we segregated the data based on sex, we did not find any statistically significant differences (data not shown), however, we acknowledge that our study was not sufficiently powered to determine a sex differential. For transparency we have color coded each data point to differentiate between male and female donors throughout the datasets.

Conclusion
The ChAdOx1 nCoV-19 vaccination induced prolonged innate immune activation with evidence to support the hypothesis that adenoviral vector-based vaccines induce trained immunity in humans. Our study is the first, to our knowledge to show that monocyte phenotype and proinflammatory function both at baseline and in response to subsequent unrelated insults, is enhanced up to 3 months post vaccination with a single dose of the ChAdOx1 nCoV-19 vaccine. These data improve our understanding of the contributions of innate immune responses to vaccine efficacy and to heterologous vaccine effects, and may aid in the design of future vaccines or innovative vaccine strategies.

Monocyte Isolation
This study used venous blood from healthy volunteers (n=10) who were all aged between 23

Statistical Analysis
Data were analysed using GraphPad Prism software (version 9). One-way repeated measures ANOVA was used to statistically analyse differences in the frequency and absolute number of CD14 + cells, the frequency of CD14 + CD16 + cells, and in the median fluorescence intensity of CD14, HLA-DR, CD40, CD80 and CD86 following vaccination with the ChAdOx1 nCoV-19 vaccine. Statistically significant differences between the expression of mRNA transcripts of GPI, PFKFB3, GAPDH, PKM2, and ATP5B were determined by a mixed-effects model (REML) ANOVA with Šídák's multiple comparisons test. For the production of cytokines and chemokines using multiplex ELISA, statistically significant differences were assessed using repeated measures one-way ANOVA with appropriate post-tests. Statistically significant differences were denoted as ****P<0.0001, ***P<0.001, **P<0.01, *P<0.05.

Study Approval
This study was approved by the Faculty of Health Science Research Ethics level 2 Committee, Trinity College Dublin. Written informed consent was received prior to participation, in accordance with GDPR and Health Research Regulations.

Author contributions
Authorship was assigned based on meeting ICMJE criteria.   (A) Cells were Fc blocked and stained with fluorochrome-conjugated antibodies specific for CD14, CD68 and CD16. Total monocytes were identified as CD14 + CD68and CD14 + CD16 + monocytes were also examined. The median fluorescent intensity of CD14 in the total ex vivo CD14 + population was assessed over time (B). The absolute number of CD14 + cells was calculated by multiplying the total cell yield from the hyperosmotic percoll enrichment by the percent CD14 + cells (C). Monocyte frequency was calculated by dividing the total number of CD14 + cells by the total number of PBMC (D). The absolute number of CD14 + CD16 + cells was calculated by multiplying the total cell yield from the hyperosmotic percoll enrichment by the percent CD14 + CD16 + cells (E). CD14 + CD16 + monocyte frequency was calculated by dividing the total number of CD14 + CD16 + cells by the total number of PBMC (F). Each dot represents an individual donor (n=10), with blue dots denoting male donors and pink dots denoting female donors. Data is graphed as the mean value ± SD. Statistically significant differences between the groups were determined by repeated measures one-way ANOVA using Dunnett's multiple comparisons test; **P<0.01, *P<0.05.

Figure 3: Vaccination enhanced the expression of cell markers associated with antigen presentation and T cell activation.
Monocytes were enriched from the PBMC of healthy donors on day -1 (pre-vaccine), day 14, day 56 and day 90 after vaccination using a hyperosmotic percoll gradient. The cell surface expression of the antigen presentation marker HLA-DR (A) and the T cell co-stimulatory molecules CD40 (B), CD80 (C) and CD86 (D) on ex vivo monocytes was assessed by flow cytometry. Graphs show collated data with each dot representing an individual donor (n=10); blue dots denote male donors and pink dots denote female donors. Representative histograms illustrate the difference in the median fluorescence intensity of each marker in stained (full outline) and unstained (dotted outline) samples on day -1 (blue) and day 90 (red). Data is graphed as the mean value ± SD. Statistically significant differences between the groups were determined by a repeated measures one-way ANOVA using Tukey's multiple comparisons test; ****P<0.0001, ***P<0.001, **P<0.01, *P<0.05. (A) This diagram shows the breakdown of glucose to pyruvate via the glycolytic pathway. Pathway intermediates are in black, enzymes are in green, and the enzymes analyzed in this study are in red. Monocytes were enriched from the PBMC of healthy donors on the day before (day-1) and day 14, 56 and 90 after vaccination using a hyperosmotic percoll gradient. Relative expression of transcript levels of GPI (B), PFKFB3 (C), GAPDH (D), and PKM2 (E) are shown. Isolated monocytes were stimulated ex vivo with medium (F) or irradiated M. tuberculosis (10 μg/ml iH37Rv; G) and the concentration of IL-1β in the supernatant was measured by multiplex ELISA. Relative expression of transcript levels ATP5B (H), a gene marker of oxidative phosphorylation, was also determined. Gene expression was determined using RT-qPCR. Each dot represents an individual donor (n=7-8) with blue dots denoting male donors and pink dots denoting female donors. Statistically significant differences between the groups were determined by a mixed-effects model (REML) ANOVA with Šídák's multiple comparisons test (B-E, H) and a repeated measures one-way ANOVA using Dunnett's multiple comparisons test (G); **P<0.01, *P<0.05. Monocytes were isolated from the PBMC of healthy donors on the day before (day-1) and day 14, 56 and 90 after vaccination using a hyperosmotic percoll gradient. Monocytes were further purified using plastic adherence and were routinely over 90% pure. Monocytes were left to rest overnight and stimulation ex vivo with medium (unstimulated), irradiated M. tuberculosis (iH37Rv; 10 μg/ml), LPS (10ng/ml), or Pam3Csk4 (10 μg/ml) for 24 hours. Each dot represents an individual donor (n=6) with blue dots denoting male donors and pink dots denoting female donors. Data is graphed as the mean value ± SD. Statistically significant differences between the groups were determined by a repeated measures one-way ANOVA using Dunnett's multiple comparisons test; ****P<0.0001, ***P<0.001, **P<0.01, *P<0.05. Figure 6: Vaccination results in altered chemokine production in response to unrelated stimuli. Monocytes were enriched from the PBMC of healthy donors on the day before (day-1) and day 14, 56 and 90 after vaccination using a hyperosmotic percoll gradient. Monocytes were further purified using plastic adherence and were routinely over 90% pure. Monocytes were left to rest overnight and stimulation ex vivo with medium (unstimulated), irradiated M. tuberculosis (iH37Rv; 10 μg/ml), LPS (10ng/ml), or Pam3Csk4 (10 μg/ml) for 24 hours. The concentrations of MCP-1 (A), CXCL1 (B), CXCL2 (C), and MIP-1α (D) in [ng/ml] in the supernatants were assessed using a multiplex ELISA. Graphs show collated data with each dot representing an individual donor (n=6) with blue dots denoting male donors and pink dots denoting female donors. Data is graphed as the mean value ± SD. Statistically significant differences between the groups were determined by a repeated measures one-way ANOVA using Dunnett's multiple comparisons test; ****P<0.0001, ***P<0.001, **P<0.01, *P<0.05.