Abstract
Mucosal-associated invariant T (MAIT) cells can be activated by viruses through a cytokine-dependent mechanism, and thereby protect from lethal infection. Given this, we reasoned MAIT cells may have a critical role in the immunogenicity of replication-incompetent adenovirus vectors, which are novel and highly potent vaccine platforms. In vitro, ChAdOx1 (Chimpanzee Adenovirus Ox1) induced potent activation of MAIT cells. Activation required transduction of monocytes and plasmacytoid dendritic cells to produce IL-18 and IFN-α, respectively. IFN-α-induced monocyte-derived TNF-α was identified as a novel intermediate in this activation pathway, and activation required combinatorial signaling of all three cytokines. Furthermore, ChAdOx1-induced in vivo MAIT cell activation in both mice and human volunteers. Strikingly, MAIT cell activation was necessary in vivo for development of ChAdOx1-induced HCV-specific CD8 T cell responses. These findings define a novel role for MAIT cells in the immunogenicity of viral vector vaccines, with potential implications for future design.
One sentence summary Robust immunogenicity of candidate adenovirus vaccine vectors requires the activation of unconventional T cells.
Body
Mucosal-associated invariant T (MAIT) cells, an abundant T cell population in humans, bridge innate and adaptive immunity due to their ability to execute effector functions following cytokine stimulation in the absence of TCR signals(1). In vivo, MAIT cells can respond to viruses in this TCR-independent manner, and mediate protection against lethal infection via early amplification of local effector mechanisms(2–4). We reasoned that such focused activity could play a critical role in viral vaccine immunogenicity. Replication-incompetent adenovirus (Ad) vectors are novel and highly potent vaccine platforms for many human diseases(5). We therefore sought to determine if such vectors activate MAIT cells and if this activation impacts on vaccine immunogenicity.
Firstly, to determine if MAIT cells respond to Ad vectors, we stimulated human PBMCs for 24 h with increasing MOIs of Ad5 and ChAdOx1, two clinically-relevant vectors(6, 7). ChAdOx1 induced robust dose-dependent upregulation of IFN-γ, CD69, and granzyme B by MAIT cells (Fig. 1A-C; Fig. S1A-D). In contrast, Ad5 only weakly activated MAIT cells even at the maximum dose (Fig. 1A-C). Activation in response to Ad vectors was confirmed using the MR1/5-OP-RU tetramer to identify MAIT cells (Fig. S1E). Vδ2+ T cells share many characteristics with MAIT cells(8, 9), and showed analogous Ad vector-induced activation (Fig. S1A, S1F).
We tested a wider range of Ad vectors including three species C-derived vectors (weak innate inducers(10–12)): Ad5(13), Ad6(13), and ChAdN13 (unpublished), and five non-species C vectors (strong innate inducers(10–12)): Ad35 (B)(13), Ad24 (D)(13), ChAdOx1 (E) (13, 14), ChAd63 (E)(13), and ChAd68 (AdC68; E)(15). In response to stimulation with the various vectors there was a gradient of IFN-γ, CD69, and granzyme B production by MAIT and Vδ2+ T cells, which resulted in greater average activation by non-species C as compared to C vectors (Fig. 1D-F; Fig. S1G-I), consistent with the above reports of differential innate immune activation by these families of vectors.
We next determined if MAIT and Vδ2+ T cells are activated following administration of Ad vectors to humans. We analyzed the activation of MAIT and Vδ2+ T cells and plasma cytokine levels on day -1 and day 1 following immunization of humans with 5×1010 vp of a novel ChAdOx1-MenB.1 vaccine (Fig. S2A, S2B). We observed modest, but statistically-significant upregulation of CD69 on MAIT and Vδ2+ T cells one day following ChAdOx1 immunization (Fig. 1G, 1H), with no changes in cell frequency (Fig. S2C). The degree of MAIT and Vδ2+ T cell activation was highly correlated within individuals (Fig. 1I). Plasma cytokines/chemokines IFN-γ, IL-6, CCL-2, and TNF-α were induced following vaccination (Fig. S2D), consistent with data from non-human primates(11), and the degree of MAIT and Vδ2+ T cell activation was correlated with changes in these innate cytokines/chemokines (Fig. 1J; Fig. S2E).
The mechanism of Ad vector-induced activation of MAIT cells was next investigated. Ad5 and ChAdOx1 displayed similar abilities to transduce PBMCs (Fig. S3A), and HLA-DR+CD11c+CD19-CD3-monocytes/cDCs were the major transduced population (83-98% of GFP+ cells) (Fig. S3B, S3C). While Ad5 and ChAdOx1 both efficiently transduced monocytes/cDCs (Fig. 2A), Ad5 transduced only 1.5% of CD123+ pDCs compared with 17.4% of CD123+ pDCs transduced by ChAdOx1 (MOI=103 vp) (Fig. 2A), consistent with a prior report of poor pDC transduction by Ad5(12).
Given their efficient transduction, we sought to determine the role of monocytes in Ad vector-induced activation of MAIT cells. Depletion of monocytes significantly reduced expression of IFN-γ, CD69, and granzyme B by MAIT cells following ChAdOx1 stimulation (Fig. 2B; Fig. S4A). Consistent with prior studies on viruses(2, 3), MAIT cell activation by Ad vectors was independent of TCR signaling (Fig. S4B) -- suggesting a cytokine-mediated activation process. Depletion of monocytes abolished IL-18 secretion following vector stimulation (Fig. 2C), and blockade of IL-18 signaling reduced MAIT cell IFN-γ, CD69, and granzyme B production (Fig. 2D; Fig. S4C). Blocking IL-12 reduced only IFN-γ production by MAIT cells (Fig. S4C), and blocking IL-15 had no effect. In contrast with ChAdOx1, Ad5 stimulation did not induce detectable levels of IL-18 or IL-12p70 (Fig S4D, S4E), consistent with the non-stimulatory nature of this vector. Direct inhibition of the Cathepsin B-NLRP3 inflammasome pathway(16) using four different pharmacologic approaches (Ca-074 Me, MCC950, elevated extracellular [K+], and Z-YVAD-FMK), significantly reduced expression of IFN-γ, CD69, and granzyme B by MAIT cells (Fig. 2E; Fig. S5A-C), and production of IL-18 following ChAdOx1 stimulation (Fig. 2F; Fig. S5D), similar to prior data examining IL-1β(17, 18). This effect was not due to altered transduction of PBMCs by ChAdOx1 (Fig. S5E).
Given the differential transduction of pDCs, the role of these cells in Ad vector-mediated activation of MAIT cells was investigated. Depletion of CD123+ pDCs resulted in a significant 67% reduction in IFN-γ production by MAIT cells (Fig. 2G), and reduced IFN-α levels by >99% following ChAdOx1 stimulation (Fig. 2H). Inhibition of type I interferon signaling reduced IFN-γ production by MAIT cells by 56-58% (Fig. 2I). Compared with ChAdOx1, Ad5 induced negligible amounts of IFN-α (Fig. S6A, S6B), consistent with previous reports(11, 12).
We envisaged a model where monocyte-derived IL-18 and pDC-derived IFN-α were the minimal factors required to activate MAIT cells in response to ChAdOx1 stimulation. However, while IFN-α/β + IL-18 induced MAIT cell IFN-γ in a PBMC culture, this was not seen using isolated MAIT cells (Fig. 3A), despite these cytokines upregulating CD69 on isolated MAIT cells (Fig. S7A). Depletion of monocytes from PBMCs reduced MAIT cell IFN-γ production following IFN-α + IL-18 stimulation (Fig. S7B), and addition of monocytes rescued this (Fig. 3B), indicating a monocyte-derived, IFN-α-dependent factor. The stimulatory factor was secreted, as either conditioned supernatant from IFN-α-treated monocytes (combined with IL-18), or provision of PBMCs across a transwell, significantly rescued IFN-γ production by isolated MAIT cells (Fig. 3C; Fig. S7C). IFN-α-stimulated monocytes secreted multiple interferon-responsive chemokines (e.g. MCP-1/CCL2), as well as TNF-α (Fig. 3D; Fig. S7D). Addition of recombinant TNF-α or an anti-TNFR2 agonist to IFN-α + IL-18-stimulated isolated MAIT cells increased IFN-γ production by >300% (from 4% to 16.5% and 17.6%, respectively; Fig. 3E; Fig. S7E). We confirmed the critical role of TNF-α as the presence of anti-TNF-α antibody (adalimumab) during IFN-α + IL-18 stimulation of PBMCs inhibited IFN-γ production by MAIT cells (Fig. S7F). The stimulatory capacity of supernatant from IFN-α-conditioned monocytes was also inhibited by the presence of adalimumab (Fig. S7G). TNF-α blockae using either adalimumab or recombinant TNFR2-Fc fusion protein (etanercept), but not a control anti-α4β7 antibody (vedolizumab), inhibited IFN-γ production by MAIT cells in response to ChAdOx1 (Fig. 3F; Fig. S7H). Depletion of monocytes reduced ChAdOx1-induced TNF-α production by 94% (Fig. 3G). Furthermore, Ad5 induced minimal TNF-α as compared with ChAdOx1 (Fig. S7I), consistent with the differential capacity of these two vectors to stimulate IFN-α production by pDCs (Fig. S6A).
Vδ2+ T cells were activated by Ad vectors through similar mechanisms (Fig. S8A-G). Compiling the data, the activation of innate-like T cells in response to Ad vectors requires the concerted action of IFN-α, TNF-α, and IL-18 (Fig. S9). These data extend prior reports of IFN-α-dependent activation of MAIT and Vδ2+ T cells by viruses(3, 19), by identifying a novel role for TNF-α as a necessary critical intermediary in this signaling pathway.
We next sought to determine the impact of MAIT cell activation on the induction of conventional T cell responses by ChAdOx1 immunization. C57BL/6J mice were immunized intramuscularly with ChAdOx1 or Ad5 at 108 IU, and MAIT cell activation in the spleen, liver, and inguinal LNs was measured on day 1 (Fig. S10A-C). ChAdOx1 induced substantial upregulation of CD69 and granzyme B on MAIT cells in the inguinal LNs, and to a lesser degree in the liver (Fig. 4A, 4B). Ad5 induced significantly less expression of CD69 and granzyme B. In mice, iNKT cells are the most abundant innate-like T cell population(20), and CD69 and granzyme B were also significantly upregulated on iNKT cells following ChAdOx1 immunization, with Ad5 inducing less activation (Fig. S10D, S10E). These findings validate the use of a mouse model, as these data recapitulate the (differential) activation of MAIT cells by Ad vectors.
Having validated the model, we next addressed the role of MAIT cells in immunogenicity of Ad vector vaccines. Wildtype (WT) C57BL/6J and MR1 KO mice (Fig. S10F-H)(21) were immunized intramuscularly with 108 IU of ChAdOx1 expressing an optimized invariant chain-linked HCV antigen(22, 23), and HCV-specific immune responses were measured on day 16 post-immunization. Following vaccination, MR1 KO mice had significantly reduced frequencies of CD8 T cells that produced IFN-γ, TNF-α, or both IFN-γ and TNF-α in response to HCV peptides, as compared with WT mice (Fig. 4C-F). This functional defect appeared specific to the CD8 T cell compartment, as there was no significant reduction in the frequency of HCV-specific CD4 T cells following vaccination of MR1 KO mice (Fig. S10I). HCV-specific CD8 T cells from MR1 KO mice also showed reduced degranulation, as measured by CD107a (Fig. 4G), and these cells displayed impaired differentiation towards KLRG1+ effector cells (Fig. S10J).
In summary, MAIT cells are capable of sensing the diversity of the Ad vector-induced innate immune activation landscape (e.g. IFN-α, TNF-α, IL-18) and can integrate these signals to augment vaccine-induced adaptive immune responses. The blend of signals required to maximally trigger MAIT cells uncovered here includes a novel and critical pathway via IFN-dependent TNF-α release, relying on cross-talk between two distinct populations of transduced cells, and varying between adenovirus serotypes. This full integration process is required for robust IFN-γ production, which has been shown to be critical for MAIT cell-mediated protection from viral infection(4).
This non-redundant role for MAIT cells places them in a critical bridging position between innate and adaptive immunity, despite many potentially shared functions with other innate-like populations(9). These data, coupled with studies in the lung(4, 24, 25), support an emerging model that MAIT cells can function to orchestrate early events in T cell-mediated immunity. It is striking that activation of MAIT cells – an abundant human innate-like population – is tightly and mechanistically linked to the immunogenicity of adenovirus vectors, which have emerged as a potent platform for T cell immunogenicity in human clinical trials(26, 27). This knowledge can be harnessed to further improve the design and development of these – and potentially other – vaccines against infections and cancer.
Funding
NMP is supported by an Oxford-UCB Postdoctoral Fellowship.
AA is supported by a Wellcome Clinical Training Fellowship [216417/Z/19/Z].
LCG is supported by a Wellcome PhD Studentship [109028/Z/15/Z].
MEBF is supported by an Oxford-Celgene Doctoral Fellowship.
CR is supported by the NIHR Biomedical Research Centre and is a Jenner Institute Investigator.
AJP is supported by the NIHR Oxford Biomedical Research Centre and is an NIHR Senior Investigator.
PK is supported by the Wellcome Trust [WT109965MA], the Medical Research Council (STOP-HCV), an NIHR Senior Fellowship, and the NIHR Biomedical Research Centre (Oxford).
The ChAdOx1-MenB.1 clinical trial is funded by the MRC (DPFS).
The views expressed are those of the authors and not necessarily those of the NHS, the NIHR or the Department of Health.
Author contributions
NMP and PK designed the project.
NMP, CD, CSR, EB, AJP, and PK designed the experiments.
NMP, AA, LCG, CD, CH, MEBF, LSR performed the experiments.
LSR, SC, BO, MR, SC, AF, CR, EB, and AJP provided samples and reagents.
All authors contributed to the writing and editing of the manuscript.
Competing interests
Authors declare no competing interests.
Data Availability
All primary data available upon request.
Acknowledgements
We would like to thank: Stephanie Slevin and Helen Ferry for assistance with flow cytometry and panel design, Carl-Philipp Hackstein and Christian Willberg for critical discussions, Mariolina Salio and Vincenzo Cerundolo for provision of MR1 KO mice, Marilù Esposito, Hussein Al-Mossawi, Lian Ni Lee, and Timothy Donnison for provision of reagents, the NIH Tetramer Facility for provision of MR1 and CD1d tetramers, and all of the volunteers for participation in the trial and donation of blood samples.