Epigenetic Adjuvants: Durable reprogramming of the innate immune response with adjuvants

Abstract

Development of effective vaccines is a critical global health priority. Stimulating antigen-specific B and T cells to elicit long-lasting protection remains the central paradigm of vaccinology. Adjuvants are components that enhance vaccine immunogenicity by targeting specific innate immune receptors and pathways. Recent data highlight the capacity of adjuvants to induce durable epigenetic reprogramming of the innate immune system to engender heightened resistance against pathogens. This raises the prospect of developing Epigenetic Adjuvants that, in addition to stimulating robust T and B cell responses, convey broad protection against diverse pathogens by training the innate immune system. In this review, we discuss our emerging understanding of the various vaccines and adjuvants and their effects on durable reprogramming of the innate immune response, their putative mechanisms of action, and the promise and challenges of developing Epigenetic Adjuvants as a universal vaccine strategy.

Introduction

Vaccines are crucial in the protection against infections and represent one of the greatest life-saving devices in history. The recent COVID-19 pandemic highlights the need for rapid development and production of effective vaccines in combating the spread of pathogens. The development of new vaccine strategies, such as mRNA vaccination [1–3] thus represent a major public health imperative. Adjuvants are molecular compounds that enhance the immunogenicity of vaccines by stimulating the innate immune system to induce robust T and B cell responses [4]. Several adjuvants, for example the TLR9 agonist CpG1018, TLR4 agonist-containing AS01b and AS04 [5–7], or squalene-based adjuvants such as AS03 [8] and MF59 [9] have been included for use in currently licensed vaccines, while others have demonstrated great promise in clinical and pre-clinical trials and are being extensively studied [10].

For several decades, vaccination strategies were predicated on the concept of eliciting protection through antigen-specific antibody and T cell responses. However, various epidemiological studies have suggested that immunization with vaccines such as the BCG [11] and measles [12,13] and Shingrix [14] vaccines can provide heterologous protection in an antigen non-specific fashion against a range of unrelated infections. In this context, the plausibility of repurposing BCG as a preventive measure against COVID-19 is currently being explored in clinical trials [15,16]. The recent notion of trained immunity developed by Mihai Netea and colleagues raises the possibility that innate immune cells could also acquire a form of immune memory through metabolic and epigenetic reprogramming [17,18]. Our recent work provides evidence that vaccine adjuvants also induce epigenetic imprinting of innate immune cells leading to a durable state of antiviral immunity [19,20]. This raises the prospect of Epigenetic Adjuvants, that epigenetically reprogram the innate immune system, leading to a durable and heightened state of resistance to a broad array of different pathogens. However, the precise immunological parameters and underlying mechanisms of innate reprogramming, and its impact on protection against diverse pathogens remain poorly understood. In this review, we summarize the current knowledge in the field, and suggest the following key challenges that need to be met in order to realize the dream of Epigenetic Adjuvants: (i) a deeper mechanistic exploration of the critical parameters (adjuvant types and their delivery routes and modalities; innate receptors and pathways triggered; durability of the effect) that regulate heterologous protection by the innate immune system; (ii) rigorous assessment of the impact of durable epigenetic reprogramming in protecting against diverse pathogens in animal models; (iii) rigorous assessment of the mechanisms and impact of trained immunity in randomized controlled vaccine trials or experimental medicine trials in humans, involving controlled human infection models with pathogens.

Innate stimuli in eliciting trained immunity

Trained immunity has been demonstrated using various stimuli in murine and human models through transcriptomic, epigenetic, and metabolic profiling and/or functional analysis of cytokine production. This has been observed in in vitro re-stimulation models, samples from human vaccination trials, and experimental mouse models.

Training in myeloid cells has been assessed in vitro using various stimuli. TLR ligands, such as R848, poly(I:C), Tri-DAP, MDP were found to enhance pro-inflammatory cytokine response upon re-stimulation at day 6 post-prime, whereas Pam3CSK4-, LPS-, and flagellin induced a more tolerogenic response, whereby monocytes were more refractory to secondary stimuli and produced a reduced level of pro-inflammatory cytokines upon re-stimulation [21]. Similarly, priming of human monocytes with β-glucan in vitro resulted in increased production of pro-inflammatory cytokines upon secondary stimulation and this effect was observed up till 2 weeks post-prime [17].

Using in vivo mouse models, BCG vaccine was previously found to induce long-term reprogramming in murine bone marrow haematopoietic stem and progenitor cells (HSPCs) related to IFN-related pathway and inflammatory response [22,23]. Adoptive transfer of the bone marrow HSPCs conferred protection against M. tuberculosis in recipient mice. Reprogramming at the progenitor level was recapitulated in human studies where BCG vaccination led to sustained epigenetic changes in bone marrow HSCs and blood monocytes [24]. Moreover, BCG-trained monocytes exhibited a metabolic shift towards glycolysis and increased H3K4 trimethylation at promoters of cytokine genes (IL-6, TNF-a) that persisted till 3 months post-vaccination in human [25–27]. In addition to monocytes, BCG vaccination in human also mediated reprogramming of blood neutrophils in pro-inflammatory cytokine genes, such as IL-1β and IL-8 [28]. BCG triggers multiple pathogen recognition receptors (PRRs) TLR2, TLR4, C-type lectins and NOD2 receptor, and the training effect of BCG was previously attributed to the NOD2 receptor [23,26,29].

β-glucan and lipopolysaccharide (LPS) were similarly found to induce myeloid reprogramming on a progenitor level. β-glucan is a fungal cell wall component that binds Dectin-1, and was demonstrated to modulate myelopoiesis in murine bone marrow HSPCs, persisting up to 28 days post-immunization [30,31]. Training by β-glucan was further associated with metabolic rewiring, with induction of transcription factors associated with aerobic glycolysis and activation of cholesterol synthesis pathway [18,32–34]. Additionally, LPS, a TLR4 ligand, was found to induce persistent epigenetic modifications in bone marrow HSPCs and myeloid-specific progenitor cells, with increased chromatin accessibility of myeloid enhancers observed up till 12 weeks post-immunization [35].

Apart from bacterial and fungal components, adenovirus vector was also demonstrated to elicit training of tissue resident macrophages in the lungs in a mouse model [36]. Yao et al. demonstrated that alveolar macrophages retained activation marker, specifically MHCII, and an increased chemokine production up till 28 days post-exposure. In addition, other non-PAMPs stimuli such as a western diet has also been associated with induction of trained immunity through systemic inflammation dependent on NLRP3 inflammasome [37]. These findings provide evidence that training could be induced by engagement of various PRRs, constituting those that recognize viral or bacterial components.

Functional relevance of trained immunity in the context of vaccination

There has been mounting evidence that prior training of the innate immune system could confer heterologous protection against secondary infections. Several epidemiological studies on BCG measles, Shingrix^®, smallpox, oral polio vaccines indicated non-specific protective effects [11–14,38–40] (Figure 1). In mouse models, BCG resulted in reduced mortality against PR8 influenza virus and C. albicans, and the protective effect was non-T or -B cell-dependent, as demonstrated in RAG1−/− mice [26,41]. Most recently, BCG immunization was found to protect against SARS-CoV2 challenge in mice [42]. This heterologous protective effect of BCG has also extended into human, where BCG-immunized individuals exposed to the live attenuated yellow fever vaccine (YF-17D) a month later had reduced yellow fever viremia [25]. And the reduced viremia correlated with increased IL-1β production. Additionally, rising evidence suggests a protective effect of BCG against respiratory infections in elderly and COVID-19 symptoms [43,44].

Figure 1.

Non-specific protection induced by vaccines in human studies

β-glucan and zymosan were found to protect against a broad range of bacteria in mouse models. β-glucan protected mice against L. braziliensis footpad infection, when infection was administered 7 days after priming [32]. This was attributed to increased induction of IL-32γ, which was also present in bone marrow progenitor cells a week after immunization. Zymosan administered as early as 5 weeks prior protected mice against various systemic and localized bacterial infections, including E. coli and L. monocytogenes [45]. This protection was possibly contributed by increased IL-1β production and higher stimulation of myelopoiesis in the bone marrow. Several studies have associated the protective capacity of training to central training of bone marrow HSPCs [46]. In a study with LPS, de Laval et al. demonstrated that transplanted BM HSPCs from LPS-exposed mice protected recipient mice from systemic P. aeruginosa infection, with reduced bacterial burden in the spleen and liver [35].

In addition to the “classic” stimuli known to induce myeloid training, adenoviral vectors were also found to afford protection against localized bacterial infection. This was mediated by so-called “memory alveolar macrophages” that promoted increased chemokine production and neutrophil recruitment [47]. TLR agonists, including TLR2/3/7/9, were also demonstrated to protect against secondary viral infection in mice, with protective effect observed up to one or several weeks after immunization [48–52]. For example, poly(IC:LC) protected mice against lethal SARS-CoV when delivered intranasally and the protective effect was observed when the agonist was given as early as 21 days prior to virus exposure [48] (Figure 2).

Figure 2.

Non-specific protection induced by vaccines in mouse models

Impact of trained immunity on T cell and antibody responses

Despite the current paucity of supporting data, trained innate cells have also been proposed to enhance T cell responses. Following BCG vaccination in human, pre-exposed PBMCs isolated as late as 1 year post-vaccination and re-stimulated with S. aureus led to increased production of Th1 and Th17 cytokines, collectively with increased activation markers on circulating monocytes, such as PRR expression [53]. However, the direct role of innate contribution to enhanced T cell cytokine production was not formally demonstrated. It was also proposed that enhanced MHCII and costimulatory molecules on APCs could promote T cell responses [54]. Additionally, data suggests that BCG vaccination could influence antibody response to secondary vaccination. Infants who had received the BCG vaccine have a higher antibody titer against other vaccine antigens, such as tetanus and pneumococcus [55]. Therefore, it would be relevant to understand how prior vaccination could impact secondary vaccine response in the context of adjuvanted vaccines.

Are epigenetic adjuvants feasible?

Adjuvants are known to enhance vaccine immunogenicity through specific targeting of the innate immune system [4]. Considering that bacterial LPS and fungal β-glucan have been demonstrated to elicit trained immunity, it could be perceived that an adjuvant, which mediates innate pathways, could exert similar effects [56]. Our recent study in human receiving flu vaccine adjuvanted with the squalene-based AS03 showed prolonged histone modifications and epigenetic changes in monocytes, up till 21 days post-vaccination [19]. In particular, classical monocytes were found to retain increased chromatin accessibility of antiviral IRF loci, including key antiviral immune regulators such as RIG-1 (DDX58), IRF1, IRF8, and a range of interferon stimulated genes. Importantly, PBMCs from vaccinated individuals controlled in vitro infection with Dengue and Zika virus more efficiently as evidenced by reduced viral load in post-vaccination cultures and viral titers correlated with IRF expression levels. In addition to an increase in IRF accessibility, we observed a decreased chromatin accessibility of AP-1 loci. Amongst the genomic regions with decreased chromatin accessibility were also many loci encoding cytokines (e.g., IL-1B, CCL2, CCL7, IL-10) and regions associated with TLR-associated cytokine production and MAP kinase activation [19], indicating a state of immune refractoriness, possibly protecting the host from hyperinflammation during a viral infection as seen in many patients with severe COVID-19.

In addition, we found that a TLR7/8 agonist, 3M-052, could induce long-lasting transcriptomic and epigenomic signatures in myeloid cells in mice. Specifically, myeloid cells, in particular monocytes, were found to retain increased gene expression and accessibility of interferon-stimulated genes (e.g. Cxcl10, Ifit2, Ddx58) and IRF and STAT loci accessibility, with these changes persisting up till 28 days after immunization [20]. We also found that a similar monocyte population with increased IRF/STAT and reduced AP-1 transcription factor gene expression was observed post-boost following BNT162b2 mRNA vaccination in human [57]. These findings suggest the induction of a distinct monocyte subset that could potentially be epigenetically reprogrammed with a heightened antiviral response following vaccination (Figure 3).

Figure 3.

Epigenetic Adjuvants [19,20,57]. Vaccination of humans with the AS03 adjuvanted H5N1 vaccine induces enhanced and durable accessibility of chromatic loci targeted by IRF transcription factors IRF 1, 2 and 7 (and durable expression of interferon stimulated genes [ISGs]), and reduced accessibility of loci targeted by AP-1 transcription factors in myeloid cells [18]. Similar results are observed in mice with vaccines adjuvanted the TLR7/8 ligand 3M-052/alum [19]. In the case of the seasonal influenza vaccine (TIV), which does not contain any exogenous adjuvants, vaccination of humans induced reduced accessibility of AP-1 targeted loci [18]. Enhanced accessibility of IRF loci was associated with enhanced resistance against infection with heterologous viruses, such as dengue and Zika. Diminished accessibility of AP-1 targeted loci resulted in a state of refractoriness, as evidenced by reduced cytokine production in response to TLR stimulation. The Pfizer-BioNTech BNT162b2 vaccine induces the development of a subset of myeloid cells with a very similar transcriptional profiles (high ISGs and IRF targeted genes, low AP-1 transcription factor) as the aforementioned myeloid cells induced by H5N1+AS03 and 3M-052/alum. However the epigenetic profile of these cells induced by BNT162b2 are not yet known.

Besides AS03 and 3M-052, other adjuvants, including R848 and poly(I:C), demonstrated innate training potential, where monocytes stimulated in vitro had enhanced pro-inflammatory cytokine production upon secondary stimulation [58]. The TLR3 agonist poly(I:C) when administered into mice upregulated IFN-β, IFN-γ, and pro-inflammatory cytokine expression and protected mice against secondary viral infection up to 5–7 days after priming [49]. In addition, cytokines such as IL-1β, type I and II IFN were also found to modulate metabolic and epigenetic changes in myeloid cells [18,31,59,60]. Adjuvants that promote the production of these cytokines could therefore potentiate training under certain conditions.

Most recently, the Shingrix vaccine, a recombinant zoster vaccine containing the AS01b adjuvant, has been suggested in epidemiological studies to convey some non-specific effect against COVID-19 [14]. Whether AS01b also induces lasting epigenomic and functional changes, such as those induced by AS03, in classical monocytes or any other immune cells is currently unknown.

Considerations and challenges in developing epigenetic adjuvants

Epigenetic adjuvants hold promise as a new class of antigen-agnostic vaccines that stimulate a durable and heightened state of innate defense against a diverse array of pathogens. Such “Universal Vaccines,” could have a significant impact during a pandemic, when conventional vaccines that confer antigen-specific immune responses are not available. The deployment of epigenetic adjuvants as a prophylactic (perhaps as a nasal spray or oral pill) in a population at the beginning of a pandemic, could engender enhanced innate resistance for a limited period. This could thus provide a stopgap solution, until conventional vaccines are developed. Yet, major challenges need to be addressed to realize this dream. These include defining the critical immunological parameters (e.g. innate receptors and pathways, route of delivery, durability of the effect), and rigorous assessment of the mechanisms and impact of epigenetic adjuvants in humans.

A key question relates to the duration of innate memory and length of protection afforded by vaccines capable of inducing epigenetic reprogramming in innate cells. It has been proposed that training could last up to 1 year in human [61]. Current studies have pointed to trained immunity being relatively short-lived, as compared to memory in adaptive immune cells. Regardless, an epigenetic adjuvant that could be administered every few weeks could still have great utility in stopping the pandemic dead in its tracks.

Highlights.

• Recent studies have shown that adjuvants induce epigenetic reprogramming of myeloid cells leading to a durable state of heightened antiviral resistance

• These studies provide proof of concept for Epigenetic Adjuvants

• Experimental medicine studies in humans are needed to rigorously translate these promising results into a new class of universal vaccines that work in an antigen-agnostic manner