How human dendritic cells respond to the vaccine vector MVA: elucidating mechanisms at the intersection of viral replication and innate immune sensing

• Applicant: Dr. Marius Döring, Ph.D.
• Subject Area: Immunology
• Term: from 2017 to 2020
• Project identifier: Deutsche Forschungsgemeinschaft (DFG) - Project number 390979210

Modified vaccinia virus Ankara (MVA) a DNA encoded poxvirus is a promising candidate live attenuated vaccine vector for HIV infection/AIDS, tuberculosis and some types of cancer. It induces potent antigen-specific T cell responses and replicates in human cells but does not produce viral progeny. Multiple innate sensors including TLRs, cytosolic DNA and RNA sensors as well as the inflammasome have been implicated in MVA sensing by myeloid cells. Among myeloid cells, dendritic cells (DCs) are critical for the induction of protective immune responses. In DCs, MVA infection leads to vaccine antigen expression, but paradoxically also to cell death and down regulation of costimulatory molecules. How DCs sense and respond to MVA infection remains incompletely understood. We are studying the immune response of human monocyte-derived DCs (MDDCs) to MVA, compared it to HIV-1 with Vpx a well-established model of cGAS-dependent sensing in MDDCs and investigate the role of uninfected bystander DCs in immune activation. To this end, MDDCs were generated from healthy donors. We overexpress a dominant-negative IRF3 (IRF3DN) and silenced cGAS using shRNA containing lentivectors. DCs were infected with MVAGFP or HIVGFP and CD86, Siglec-1, as well as GFP and cytokine expression were analyzed by FACS. We found that CD86 was induced on uninfected bystander DCs, while it was downregulated on MVA-infected DCs. IRF3DN inhibited DC activation by MVA and HIV. In contrast, cGAS depletion only modestly decreased DC response to MVA, while it abrogated the DC response to HIV. Cross-talk between MVA-infected and bystander DCs was mediated by soluble factors. Using blocking antibody cocktails against IFN I and TNF we showed that transactivation partially required type I IFN, TNF, and also additional JAK/STAT pathways. When pre-activating DCs before MVA-infection, we noticed an enhanced susceptibility of DCs to MVA without increased cell death. Unexpectedly, we found that MVA did not replicate in WT DCs, while it did in control HeLa cells. We identified SAMHD1 as restriction factor responsible for blocking MVA replication in DCs. Alleviating this restriction increased AraC-sensitive MVA antigen expression in DCs. Our results highlight the interplay between infected and bystander human DCs in response to MVA. However, the pre-activation of DCs using a TLR ligand induced high frequencies to DCs expressing both target antigen and CD86. Interestingly, our results show that MVA antigen production in DCs is largely compromised by the activity of the restriction factor SAMHD1. Altogether, these insights may help to improve MVA as a putative life attenuated vaccine vector.

• DFG Programme: Research Fellowships
• International Connection: France
• Host: Nicolas Manel, Ph.D.