Programming cytomegalovirus as an HIV vaccine

Abstract

The initial development of cytomegalovirus (CMV) as a vaccine vector for HIV/simian immunodeficiency virus (SIV) was predicated on its potential to pre-position high-frequency, effector-differentiated, CD8⁺ T cells in tissues for immediate immune interception of nascent primary infection. This goal was achieved and also led to the unexpected discoveries that non-human primate (NHP) CMVs can be programmed to differentially elicit CD8⁺ T cell responses that recognize viral peptides via classical MHC-Ia, and/or MHC-II, and/or MHC-E, and that MHC-E-restricted CD8⁺ T cell responses can uniquely mediate stringent arrest and subsequent clearance of highly pathogenic SIV, an unprecedented type of vaccine-mediated protection. These discoveries delineate CMV vector-elicited MHC-E-restricted CD8⁺ T cells as a functionally distinct T cell response with the potential for superior efficacy against HIV-1, and possibly other infectious agents or cancers.

Keywords: HIV vaccine; SIV replication arrest; cytomegalovirus; immune programming.

Figure 1.. Immune:viral infection intercept dynamics determine HIV/SIV vaccine efficacy.

Although primary HIV/SIV infection is a continuous process, it can be divided into sequential steps that are relevant to vaccine efficacy [4]. The figure summarizes these key steps, the characteristics of the viral infection at each step, and the potential for different vaccine-elicited immune responses to intercept infection at each step in this dynamic process. Optimized efficacy refers to a vaccine-elicited response with all the necessary attributes needed for the optimal anti-viral activity of that particular response. For Ab-targeted vaccines, this would include stable high titers in plasma and tissues with documented anti-viral activity (neutralization or Fc-mediated anti-viral function). For effector memory T cell targeted vaccines (CMV vectors), this would include pre-positioned MHC-E-restricted CD8⁺ T cell responses and the protection-associated innate signaling. For central memory T cell targeted vaccines (conventional prime-boost approaches), this would include polyfunctional, broadly targeted CD8⁺ T cell responses with high anamnestic expansion potential. Note that the Ab and effector memory T cell responses are prepositioned and thus able to intercept infection when viral mass and diversity are low, prior to the potential for viral adaptations leading to immune escape. However, even the most potent anamnestic responses intercept infection too late to avoid viral adaptation and infection-associated immune dysfunction, and thus can manifest long-term viral suppression only if responses precisely target functionally constrained parts of the viral proteome (Box 1). Even then, infections are not cleared, and remain subject to immune escape (Box 2).

Figure 2.. Characteristics of SIV replication arrest efficacy in non-human primates.

The figure illustrates the general characteristics of 68–1 RhCMV/SIV vector-associated efficacy against highly pathogenic SIV. The key features of this efficacy are 1) unequivocal take of infection as demonstrated by the induction of SIV-specific T cell responses to SIV Ags not included in the vaccine, direct demonstration of viral (v) SIV DNA/RNA by PCR/RT-PCR, and adoptive transfer of typical SIV infection to naïve Rhesus Macaques (RMs) by injection of cells from the vaccine protected RMs; 2) arrest of infection in portals of entry and sites of early spread prior to exponential expansion; and 3) contraction/clearance of infection over time until protected RMs can no longer be distinguished by either immunologic or virologic criteria from vaccinated, never-challenged RMs [29,31,33]. Bone marrow biopsy was the most sensitive method to sequentially evaluate SIV infection in protected RMs [33], but it should be noted that necropsy analysis of protected RMs during the early period of protection also revealed SIV in portals of entry (rectal/colonic mucosa), draining lymph nodes (mesenteric and iliosacral) and both liver and spleen, in addition to bone marrow [31]. At late timepoints, tissue analysis was extensive at necropsy with no detectable SIV RNA or DNA in most specimens, and with no infectious virus demonstrable by adoptive transfer [31,33].

Key figure, Figure 3.. 68–1 RhCMV/SIV vector tropism restriction with cell type-specific microRNAs (miRs) reveals independent priming of MHC-E- vs. MHC-II-restricted CD8⁺ T cells in different cell types.

To inhibit vector infection of specific cell types, thereby restricting vector tropism, target sequences for cell type-selective miRs are inserted in the 3’ untranslated region of two essential vector genes. In cell types lacking expression of the targeting miR, vector growth is unimpeded. However, vector infection of cells expressing the targeting miR results in inhibition of expression of the targeted essential viral genes and abrogation of productive infection in that cell type [34,35,45]. Based on reference [35], the figure illustrates the effect of 68–1 RhCMV/SIV vector restriction by miR-142, which is selectively expressed by myeloid-derived cells (including macrophages and dendritic cells), versus miR-126, which is selectively expressed by endothelial cells. Mir-142 inhibition of vector infection of myeloid-derived cells abrogates priming of MHC-E-restricted CD8⁺ T cells, leaving a uniformly MHC-II-restricted response. In contrast, miR-126 restriction of vector infection in endothelial cells abrogates MHC-II-restricted CD8⁺ T cell priming, leaving a uniformly MHC-E-restricted CD8⁺ T cell response. Thus, MHC-E- and MHC-II-restricted CD8⁺ T cells are directly and specifically primed by myeloid-derived and endothelial cells, respectively. Of note, when both of these direct presentation pathways are blocked by combining miR-142- and miR-126-mediated inhibition, the ensuing response is MHC-Ia-restricted -- the default response that is likely cross-primed [46]. Despite the fact that CD8⁺ T cell responses elicited by these different miR-restricted 68–1 RhCMV/SIV vectors were similar in magnitude and differentiation, vaccine efficacy was only observed with the miR-126-restricted vector, confirming the requirement for the MHC-E-restricted CD8⁺ T cell response for replication arrest efficacy [35].

Figure 4.. 68–1 RhCMV/SIV vaccine protection correlates with a whole blood gene expression signature that includes an IL-15-induced transcription response in non-human primates.

A. The figure illustrates the effect of treating RMs with three ascending doses of an IL-15 agonist (Rhesus heterodimeric IL-15/IL15 receptor alpha) using whole blood transcriptome analyses at pre-treatment baseline and at the indicated days post-treatment to monitor the evolution of IL-15 responsiveness [47]. Linear modeling of the whole blood transcriptomic response revealed that two major co-expression clusters (Cluster A; acutely induced by IL-15, and Cluster B, acutely suppressed by IL-15) of genes regulated by IL-15 represented within the RhCMV/SIV vaccine-induced whole blood gene expression signature linked to protection against SIV infection [47]. The graph illustrates the time course average gene expression of each cluster to monitor the response through 29-days post-IL-15 treatment, revealing a “counter-regulation period” of homeostatic re-equilibration of gene expression. B. These figures illustrates the average log-fold change showing the differential expression of cluster A and B genes in 68–1 RhCMV/SIV vaccinated RMs that were protected (red) or nonprotected (gray/black) against SIV challenge [47]. Time points from within week (W) 0 (first vaccination) to W18 (boost vaccination) to W88 (start of SIV challenge) and corresponding day (D) are shown for RM cohorts who received RhCMV/SIV vaccine via an oral or subcutaneous (Subq) route. The magnitude of gene expression induction (Cluster A) and suppression (Cluster B) separate protected, from non-protected vaccinated RMs. C. Functional categories of gene networks within Cluster A (upper) and Cluster B (lower) genes of the whole blood response in RhCMV/SIV vaccinated RMs shown in Panels A and B are shown [47]. The colors summarize the log2-fold change in average gene expression compared to pre-vaccination baseline for the indicated networks.