Multiscale affinity maturation simulations to elicit broadly neutralizing antibodies against HIV

Abstract

The design of vaccines against highly mutable pathogens, such as HIV and influenza, requires a detailed understanding of how the adaptive immune system responds to encountering multiple variant antigens (Ags). Here, we describe a multiscale model of B cell receptor (BCR) affinity maturation that employs actual BCR nucleotide sequences and treats BCR/Ag interactions in atomistic detail. We apply the model to simulate the maturation of a broadly neutralizing Ab (bnAb) against HIV. Starting from a germline precursor sequence of the VRC01 anti-HIV Ab, we simulate BCR evolution in response to different vaccination protocols and different Ags, which were previously designed by us. The simulation results provide qualitative guidelines for future vaccine design and reveal unique insights into bnAb evolution against the CD4 binding site of HIV. Our model makes possible direct comparisons of simulated BCR populations with results of deep sequencing data, which will be explored in future applications.

Fig 1. Simulation framework for the AM model, highlighting new features.

Amino acid (aa) and nucleotide (nt) sequences of the Ag and germline BCR, respectively, are input into the model in step I. During B cell proliferation and mutation (step II), a mutation model [19] is used to determine where and which mutations will occur in the BCR sequences; nts that have a higher mutability score (χ_Mut)–which accounts for the effects of surrounding nts–have a greater chance of being selected for mutation. This is followed in step III by BCR/Ag binding free energy calculations, which utilize the program Modeller [25] and published statistical potentials [26]. Final steps in the model include a second selection step, recycling, and differentiation (details in text). BCR breadth is computed as the fraction of Ags in a panel for which the BCR binds above a chosen threshold (see Methods).

Fig 2. Results from AM simulations.

(A) mean BCR breadth of vaccination protocols with t = 1, 2, and 3 single-Ag sequential immunizations, (B) mean BCR breadth vs. mean number of GC cycles, (C) changes in mean BCR breadth vs. changes in frustration between current (t) and former (t-1) immunizations, and (D) mean BCR breadth vs. mean number of mutations towards conserved Ag residues. In (C), linear fits were used to calculate the R² values presented in the table below the graph. Error bars represent the standard deviation of two independent simulations of each vaccination protocol (except for the KR-KR-KR and HQ-EU-WT protocols, for which only one successful trial could be obtained), consisting of between one and three immunizations/GC reactions (error bars for mean BCR breadth are only shown in (A) for clarity).

Fig 3. Anti-CD4bs bnAbs employ interfacial composition and electrostatic pattern matching (ICM) during AM.

Interfacial amino acid compositions are shown in white (CH103), pink (VRC01), purple (3BNC60), and gray (NIH45-46), with solid and hatched bars representing the mature and germline forms of the Abs, respectively. For each amino acid type, the Abs’ target interfacial fraction in the CD4bs of HIV is indicated by a black bar (note that the Abs’ acidic interfacial compositions evolve to match the basic interfacial composition of the CD4bs, and vice versa).

Fig 4. Results from AM simulations: mean weighted degree of interfacial composition and electrostatic pattern matching (ICM) of all 1-, 2-, or 3-Ag protocols, sorted by the mean ICM score.

Error bars represent the standard deviation of two independent simulations of each vaccination protocol (except for the KR-KR-KR and HQ-EU-WT protocols, for which only one successful trial could be obtained), consisting of between one and three immunizations/GC reactions. Gray and black dotted lines indicate the respective values for VRC01GL and VRC01.

Fig 5. Binding free energy distributions of VRC01GL and VRC01 against a panel of 106 HIV antigens (Ags), using the scoring function developed in this work.

A black dashed line indicates the threshold of -9.9 kcal/mol used to determine the breadth (B) of the BCRs. Note that the intermediate-colored region arises from the overlap of the two distributions.

Fig 6. Definition of interfacial BCR residues used to characterize AM simulations.

(A) VMD snapshots of (left) VRC01GL and (right) VRC01, in complex with HIV-based Ags. Ags are colored in silver and Abs in gray. Residues included in the in silico CD4bs interfacial residue definition are colored orange if overlapping between the two Abs and colored green otherwise, and interfacial residues excluded from the definition due to contact or alignment with HIV glycans (red) are colored cyan. (B) Interfacial amino acid compositions for VRC01GL and VRC01 are shown by hatched and solid pink bars, respectively, using the residues defined in orange/green in (A), which mimic the biological trends shown in Fig 3. For each amino acid type, the Abs’ target interfacial fraction in the CD4bs of HIV is indicated by a black bar (note that the Abs’ acidic interfacial compositions evolve to match the basic interfacial composition of the CD4bs, and vice versa).