HIV-Host Interactions: Implications for Vaccine Design Barton F. Haynes, George M. Shaw, Bette Korber, Garnett Kelsoe, Joseph Sodroski, Beatrice H. Hahn, Persephone Borrow, Andrew J. McMichael  Cell Host & Microbe  Volume 19, Issue 3, Pages 292-303 (March 2016) DOI: 10.1016/j.chom.2016.02.002 Copyright © 2016 Elsevier Inc. Terms and Conditions

Figure 1 Model of the HIV-1 Transmission Bottleneck Mucosal transmission reduces the genetic and phenotypic diversity of the donor HIV-1 quasi-species to only one or very few variants that seed infection in the recipient. Viruses that traverse the mucosa, but are defective or fail to initiate a productive infection (i.e., have a basic reproductive ratio Ro of lower than 1), will be extinguished. In contrast, the mucosal bottleneck selects for viruses with a high transmission fitness. Although the biological properties that comprise this phenotype remain to be fully elucidated, a high replicative capacity, increased infectivity, enhanced dendritic cell interaction, and greater resistance to the antiviral effects of type 1 interferons (IFNs) are likely to contribute (Parrish et al., 2013). Cell Host & Microbe 2016 19, 292-303DOI: (10.1016/j.chom.2016.02.002) Copyright © 2016 Elsevier Inc. Terms and Conditions

Figure 2 Co-Evolution of HIV Transmitted-Founder Virus and Evolving Neutralizing Antibodies The initial transmission event of sexually transmitted HIV-1 is mediated by one transmitted founder (TF) virus. The TF virus induces an initial antibody response, called the autologous neutralizing antibody, that is specific for the TF virus. The autologous neutralizing antibody neutralizes the TF but rapidly selects virus escape mutants, which in turn induces new antibody specificities. This process is repeated throughout virus evolution such that after years of infection, a spectrum of cross-reactive neutralizing antibodies are induced, with ∼20% of chronically infected individuals making high levels of very broadly reactive neutralizing antibodies. Cell Host & Microbe 2016 19, 292-303DOI: (10.1016/j.chom.2016.02.002) Copyright © 2016 Elsevier Inc. Terms and Conditions

Figure 3 HIV-1 Trimer and Broadly Neutralizing Antibody Binding Sites Co-crystal structure of the HIV-1 trimer (Pancera et al., 2014) with gp120 in blue and gp41 in gray. The five areas targeted by broadly neutralizing antibodies are the CD4 binding site (orange), V1V2 glycans (red), V3 glycans (green), gp120-gp41 bridging site (purple), and the MPER (dark red). The area of insertion of the envelope trimer into the membrane is noted by the transmembrane domain and the gp160 cytoplasmic domain is noted. Cell Host & Microbe 2016 19, 292-303DOI: (10.1016/j.chom.2016.02.002) Copyright © 2016 Elsevier Inc. Terms and Conditions

Figure 4 Cooperation of‘ B Cell Lineages in Induction of HIV-1 Broadly Neutralizing Antibodies The TF virus induces both a broadly neutralizing antibody (bnAb) lineage (CH103 lineage in red) as well as a second lineage (the CH235 cooperating lineage in blue). The TF directly drives the bnAb lineage while the cooperating antibody lineage selects virus escape mutants that bind to and are neutralized by the bnAb lineage. Thus, in this case, the bnAb lineage is initiated by the TF virus and is driven by escape mutants from other cooperating lineages. Cell Host & Microbe 2016 19, 292-303DOI: (10.1016/j.chom.2016.02.002) Copyright © 2016 Elsevier Inc. Terms and Conditions

Figure 5 Degrees of HIV-1 Control by Vaccines that Elicit CD8 T Cell Responses Each vaccine was designed to elicit only or primarily T cell responses. These were the STEP Adenovirus-5-Gag-Pol-Nef vaccine (McElrath et al., 2008), DNA-Adenovirus-5 SIV Gag (Casimiro et al., 2005), DNA-Adenovirus-5-mosaic Gag (Roederer et al., 2014), Adenovirus-26 gag + Adenovirus 5 gag (Liu et al., 2009), RhCMV68-1 − SIV (Hansen et al., 2011), and ChAd63-HIVconsv − MVA-HIVconsv (Borthwick et al., 2014) vaccines. The peak responses after vaccination are shown. On the y axis is the magnitude as virus-specific T cells per million PBMC (either directly from Elispot values or converted from % CD8 T cells in flow cytometry assays). On the x axis is shown the number of epitopes reported (breadth) corrected for the degree of matching between vaccine and challenge/infecting virus. For perfect matches between the vaccine and challenge, the correction value is 1.0, for the STEP vaccine it is estimated to be 0.7. The left hand plot shows values for the STEP trial compared to similar vaccines in the SIV model. In the right panel these are compared to the values for the RhCMV-68.1-SIV vaccine. Also shown here are the values for the HIVconsv conserved region vaccine in a phase I trial in humans where there was no HIV-1 exposure. Where known, the outcomes of SIV challenge or HIV-1 exposure are shown. Cell Host & Microbe 2016 19, 292-303DOI: (10.1016/j.chom.2016.02.002) Copyright © 2016 Elsevier Inc. Terms and Conditions