B cell immunofocusing and repriming in two HIV-1 Env immunization regimens

Diverse and rapidly mutating viruses pose challenges to immunogen and vaccine design. In this study, we evaluated the ability of memory B-cells obtained from two independent NHP trials to cross-react with individual HIV-1 vaccine components of two different multivalent immunization strategies. We demonstrated that while an HIV-1 Env multiclade, multivalent immunization regimen resulted in a dominant memory B-cell response that converged toward shared epitopes, in a sequential immunization with clonally-related non-stabilized gp140 HIV-1 Envs followed by SOSIP-stabilized gp140 trimers, the change in immunogen format resulted in repriming of the B-cell response.


Main text
Successful vaccines against highly mutating viruses, such as influenza virus and SARS-CoV-2, rely on sequential immunizations with mono-or multivalent formulations to counter viral diversity and broaden the spectrum of primed B-cells at the time of exposure. 1 For HIV-1, multivalent, multiclade vaccination schemes -a method historically used to counter viral diversity -and sequential immunization strategies with clonally related HIV-1 envelope (Env) glycoproteins have been pursued. 2Despite intensive studies, the relative frequency of immunogen monoversus cross-reactive B-cells in repeated immunizations is not fully characterized.Here, we analyzed the cross-reactivity of B-cells isolated from non-human primates (NHPs) immunized with either a multiclade vaccine formulation or sequential immunizations with clonally related non-stabilized gp140 Envs followed by SOSIP stabilized gp140 trimers.
In summary, this study exemplifies how the choice of immunogen designs can affect the engagement of different B-cell pools in the context of sequential immunizations: In particular, how they can alter the balance between B-cell convergence toward shared epitopes (i.e., immunofocusing) and B-cell repriming.Specifically, despite the genetic distance among immunogens in the multiclade immunization scheme (Figure S2), the observed cross-clade breadth was predominantly mediated by B-cells immunofocused on shared epitopes, rather than by an additive engagement of multiple clade-specific B-cells.Immunofocusing was unsurprisingly more pronounced in sequential immunizations with clonally related non-stabilized gp140 Envs.Non-stabilized gp140 Env immunization expanded a pool of B-cells for which the cognate epitopes were accessible on the surface of SOSIP-stabilized gp140 trimers.However, despite these cross-reactive B-cells being present, the change in immunogen from non-stabilized gp140 to SOSIP-stabilized gp140 Env largely bypassed their re-engagement and induced substantial B-cell repriming of a different B-cell pool that did not cross-react with the previously administered gp140 Envs.
The minimal engagement of these pre-existing cross-reactive B-cells implies that the epitope immunodominance hierarchy of the SOSIP trimer differed from that of the non-stabilized gp140 Envs.The stabilized SOSIP trimer design used in this study was templated on the BG505 SOSIP.664trimer backbone, which displays a dominant neo-epitope cluster at the base region of the trimer. 5,6This dominant neo-epitope may have contributed to the observed B-cell repriming by SOSIP immunization in our study, supporting the development of strategies to silence undesired responses against the SOSIP base. 7,8Critically, and regardless of the specific epitopes involved, we demonstrate that B-cell priming with non-stabilized Env was insufficient to modify the effect of a different epitope immunodominance hierarchy in a SOSIP trimer.
Immunogen engineering is a powerful tool to alter immunodominance hierarchies. 9These results suggest that caution is needed to avoid off-track B-cell engagement when changing immunogen designs within sequential immunizations.In particular, immunodominance hierarchies of epitopes presented on different designs may not be similar.This consideration may also be relevant to hybrid immunogens targeting multiple viruses (e.g., influenza and SARS-CoV-2) and polyvalent preparations aimed at inducing balanced responses against multiple virus serotypes (e.g., Dengue virus).Conversely, these results also suggest that, should B-cell repriming be desirable (e.g., to target immunodominant epitopes introduced in highly mutating viruses through natural evolution) pre-existing immunity against other epitopes should not constitute an insurmountable obstacle.
Finally, highly mutating viruses are known to escape humoral immunity by either shielding or directly mutating neutralizing epitopes (e.g., HIV-1, SARS-CoV-2, influenza antigenic drift), or by large envelope rearrangements (e.g., influenza antigenic shift).Our findings raise the intriguing, albeit speculative, hypothesis that the introduction of new immunodominant epitopes on their Env could suffice to redirect the immune response away from potentially protective epitopes, despite them being expressed on the virion surface, and act as an additional mechanism of immune evasion.Plasma viral load determination.Quantitative real-time reverse-transcriptase (RT)-PCR assay to determine SHIV.CH505.375H.dCTviral load was performed as previously described. 10The sensitivity of the assay is 60 copies/ml of plasma.
NHP sequential immunization with IDLV expressing CH505 Envs.The study design has been previously reported in detail. 4Briefly, NHPs were immunized at 6-month intervals with non-stabilized gp140 Envs isolated from individual CH505, including the transmitted founder virus and sequentially evolved Envs isolated from weeks 53, 76 and 100 selected based on their binding profile to monoclonal antibodies of the CH103 bnAb lineage at progressive maturation stages (CH505.T/F, CH505.w53,CH505.w76,CH505.w100,respectively).Two additional immunizations were performed using stabilized CH505.w136SOSIP Env trimer.Immunogens were delivered either through IDLV alone (3x10 8 TU i.m.) or in combination with the respective protein (100 µg s.c.) in GLE-SE adjuvant.Six weeks after the last immunization, NHPs were challenged intrarectally with the autologous SHIV.CH505.375H.dCTstrain.PBMCs were collected 14 days after the last immunization with CH505.w76gp140 (week 75) and CH505.w100SOSIP (week 117).The rhesus macaques used in this study were housed at BIOQUAL, Inc. and the study was approved by the BIOQUAL Institutional Animal Use and Care Committee (Study # 18-001).
Memory B-cell cultures.We have previously described the memory B-cell culture method adapted to induce proliferation and differentiation of Rhesus memory B-cells into antibodysecreting cells. 3Briefly, memory B-cells were sorted as described above in bulk into wells containing 5000 MS40L feeder cells, RPMI-1640 supplemented with 15% FBS, 1 mM sodium pyruvate, 1% non-essential amino acids, 25 mM HEPES buffer, 2.5 μg mL − ELISA culture supernatant screening.High-binding 384-well plates were pre-coated with streptavidin for biotinylated antigens and coated with 2 µg mL −1 of Env proteins or 0.5 µg mL −1 of polyvalent goat anti-human Ig Ab (Life Technologies, Cat# H17000) to measure IgG, IgA, and IgM levels, diluted in 0.1 M NaHCO3 solution.Culture supernatants were tested at 1:3 dilution in blocking buffer.After two washes with washing solution, secondary HRP-conjugated antibodies (Jackson ImmunoResearch, Cat.no.109-035-098, 109-035-129, and 109-035-011) were added at lot-specific optimal concentrations for 1 h.After 4 washes, plates were developed for 10 min using 15 µL per well SureBlue Reserve TMB microwell peroxidase substrate before adding 15 µL per well of 0.1 M HCl.All of the following criteria had to be met to define culture supernatant positivity for binding to Env immunogens: measurable IgG, IgA, or IgM levels; OD450 > 0.1 and > 2 × OD450 reads from blank wells; OD450 > 120% OD650.
HIV-1 gp120 Env phylogenetic analysis.The unrooted phylogeny of the four HIV-1 gp120 protein sequences was calculated using the maximum likelihood algorithm implemented in IQtree-2.The same topology with similar branch lengths was also calculated using the Bayesian phylogenetic analysis program MrBayes.

Figure 1 .
Figure 1.Tetravalent gp120 Env immunization.a NHP immunization scheme with multivalent HIV-1 gp120 Env cocktails delivered as DNA and monomeric proteins.b Viremia time-course of the four NHPs included in this study post-seroconversion or after the fifteenth challenge for RVv15.c Immunogen binding profiles of 734 B-cells isolated from all four NHPs.Reactivity with individual and multiple gp120 Env immunogens is color coded as indicated.d Binding to individual immunogens of monoreactive B-cells, expressed as percentage of the total number of monoreactive B-cells.Data are shown as aggregate ("All NHPs") and for each NHP.e Immunogen cross-reactivity of multireactive B-cells, expressed as percentage of the total number of multireactive B-cells, shown as aggregate ("All NHPs") and for each NHP.

Figure 2 .
Figure 2. Sequential immunization with clonally related Envs. a NHP immunization scheme (see methods and ref 4 ).b,c Cross-reactivity of B-cells isolated (b) before and (c) after SOSIP trimer immunizations with non-stabilized gp140 Envs immunogens (outer ring) and stabilized CH505.w136SOSIP trimer (inner ring).Data are shown as aggregate ("All NHPs") and for each NHP.Number of B-cells analyzed is shown for each chart.Reactivity with one or more nonstabilized gp140 Env immunogens is color coded as indicated.Absence of binding to any of the four non-stabilized gp140 Env immunogens is shown in gray.