Vaccine priming of rare HIV broadly neutralizing antibody precursors in nonhuman primates

Abstract

Germline-targeting immunogens hold promise for initiating the induction of broadly neutralizing antibodies (bnAbs) to HIV and other pathogens. However, antibody-antigen recognition is typically dominated by heavy chain complementarity determining region 3 (HCDR3) interactions, and vaccine priming of HCDR3-dominant bnAbs by germline-targeting immunogens has not been demonstrated in humans or outbred animals. In this work, immunization with N332-GT5, an HIV envelope trimer designed to target precursors of the HCDR3-dominant bnAb BG18, primed bnAb-precursor B cells in eight of eight rhesus macaques to substantial frequencies and with diverse lineages in germinal center and memory B cells. We confirmed bnAb-mimicking, HCDR3-dominant, trimer-binding interactions with cryo-electron microscopy. Our results demonstrate proof of principle for HCDR3-dominant bnAb-precursor priming in outbred animals and suggest that N332-GT5 holds promise for the induction of similar responses in humans.

Fig. 1.. Detection of BG18-like responses in rhesus macaques.

(A) The definition of a BG18 type I HC precursor that was used for searching naïve B cells plotted in (B) for humans (HuPre1) and macaques (RmPre1 and RmPre2), where pre1 and pre2 refer to different alleles. The position of the FG(V/L) was allowed to start at position 7, 8, or 9 of the HCDR3. The alignment shows the FG(V/L) starting at position 8. The BG18-inferred germline (BG18_GL₀) HCDR3 sequence is shown for comparison. Black indicates templated regions of the HCDR3; blue and red indicate junction regions. (B) Frequency of BG18 type I HC sequences in 14 human donors and 60 rhesus macaques. (C) Immunization schedule of NHPs (n=8) with N332-GT5 + SMNP and sampling time points for which sorting and BCR sequencing were carried out. ED, escalating dose; LN FNA, lymph node fine needle aspirate; PBMC, peripheral blood mononuclear cells. (D) Frequency of Env⁺ GC B cells as a percentage of total CD20⁺ B cells. The gray area is set from 0.001% to the median frequency of Env⁺ GC B cells observed in the pre-immunization samples. Lines indicate the median +/− interquartile range. The dotted vertical line indicates that a bolus boost occurred at week 10. (E and H) Number of BG18 type I sequences identified in (E) GC or (H) memory B cells at different time points, with each circle representing a different animal. (F and I) Frequency of BG18-like sequences among Env⁺ B cells from (F) GC or (I) memory B cells. (G and J) Frequency of BG18-like sequences among (G) GC or (J) memory B cells. (E – J) Bars indicate the median +/− interquartile range for the responders. For (E and F) "All" indicates the result of combining all sequences from weeks 3 to 13. In (G) "All" indicates the median frequency among the responders from weeks 3 to 13 at the time points for which responses were detected. In E-G, MD39 group (n=4) indicates Env+ sequences isolated from GC B cells at weeks 3, 4, 7 and 10 after priming with BG505 SOSIP MD39 trimer. ND, not detected.

Fig. 2.. BG18 type I lineages and structural analysis.

(A) Gene segment assignment and representative HCDR3 sequence of 38 BG18 type I lineages. The D3-41 gene is colored blue and two critical contact residues are colored red. D3-41 in alternate reading frame is colored green. (B)-(E) upper panels show cryo-EM models of (B) BG18_GL0 Fab in complex with N332-GT2 (PDB ID: 6DFH), (C) canonical BG18 type I macaque Fab RM_N332_03 in complex with N332-GT5 (K_D, 16 pM), (D) BG18 type I Fab RM_N332_36 with D3-41 in alternate reading frame in complex with N332-GT5 (K_D, 38 pM), (E) BG18 type I Fab RM_N332_32 with kappa light chain in complex with N332-GT5 (K_D, 81 pM). (B – E) lower panels show magnified view of HCDR3 interacting with gp120 for each structure; gp120 is colored gray and the N332 epitope (N332-GT2 residues Gly324, Asp325, Val326, Arg327, Met328, Ala329, His330, Ile415, Leu416 and Pro417 or the equivalent position in N332-GT5) colored red; HCDR3s are colored blue, orange, purple, green for BG18_GL₀, RM_N332_03, RM_N332_36, and RM_N332-32, respectively.

Fig. 3.. Affinity maturation of BG18 type I and other N332-GT5-elicited antibodies.

(A) SHM of VH genes in BG18 type I and other Env⁺ sequences from GC B cells. (B) SHM of VH genes in BG18 type I and other Env⁺ sequences from memory B cells. (C) SHM of VL genes in BG18 type I and other Env⁺ sequences from GC B cells. (D) SHM of VL genes in BG18 type I and other Env⁺ sequences from memory B cells. For (A-D), each dot represents the median SHM for one animal, and lines indicate the median of the medians +/− interquartile range for all animals. (E) SPR K_Ds for BG18 type I Fabs isolated at weeks 3, 4, 7 and 10 post prime from GC B cells binding to N332-GT5 or N332-GT5-KO. The dotted line at 2 x10⁻⁵ M indicates no binding at the highest concentration tested. The dotted line at 1.6x10⁻¹¹ M represents the approximate upper affinity limit of the instrument. Lines indicate the median with interquartile range. Each data point is representative of 1 to 4 technical replicates. (F) SPR K_Ds for BG18 type I Fabs compared to other epitope-specific Fabs binding to a panel of boost candidates that are more similar to a native trimer than N332-GT5. Data points represent 1 technical replicate. The dotted line at 2 x10⁻⁵ M indicates no binding at the highest concentration tested. The dotted line at 1.6x10⁻¹¹ M represents the approximate upper affinity limit of the instrument.

Fig. 4.. Identification of a broader class of BG18-like antibodies.

(A) HCDR3 length distribution of Env⁺ GC B cells and epitope-specific memory B cells isolated at wk 12 after filtering out the BG18 type I sequences. (B) SPR K_Ds of Fabs derived from activated B cell supernatants that were positive for binding to BG505_B23 by ELISA. Each data point represents 1 technical replicate. Lines indicate median +/− interquartile range. The dotted line at 2 x10⁻⁵ M indicates no binding at the highest concentration tested. The dotted line at 1.6x10⁻¹¹ M represents the approximate upper affinity limit of the instrument. (C) HCDR3 sequence and gene segment assignment of two Fabs that bind with high affinity to the BG505_B23 trimer for which cryo-EM structures were determined. (D) Cryo-EM model of N332-GT2 in complex with BG18_GL₀ (PDB ID: 6DFH). (E) Cryo-EM model of Fab RM_N332_07 in complex with N332-GT5. (F) Cryo-EM model of Fab RM_N332_08 in complex with N332-GT5. (G) Diagram showing the latitudinal, longitudinal, and HC-LC twist angles for BG18 (PDB ID 6DFG), BG18_iGL0 (6DFH), HMP1 (6NF5), HMP42 (6NFC), RM_N332_03, RM_N332_32, RM_N332_36, RM_N332_07, and RM_N332_08 in yellow bars and seven other N332-dependent bnAbs [PGT122 (4TVP), DH270.6 (6UM6), BF520.1 (6MN7), PGT128 (5ACO), QA013.2 (7N65), PGT135 (4JM2), and 438-B11 (6UTK)] in cyan bars. (H) 3D scatter plot showing latitudinal, longitudinal, and HC-LC twist angles for the antibodies in (G).