Dimeric 2G12 as a potent protection against HIV-1

Abstract

We previously showed that broadly neutralizing anti-HIV-1 antibody 2G12 (human IgG1) naturally forms dimers that are more potent than monomeric 2G12 in in vitro neutralization of various strains of HIV-1. In this study, we have investigated the protective effects of monomeric versus dimeric 2G12 against HIV-1 infection in vivo using a humanized mouse model. Our results showed that passively transferred, purified 2G12 dimer is more potent than 2G12 monomer at preventing CD4 T cell loss and suppressing the increase of viral load following HIV-1 infection of humanized mice. Using humanized mice bearing IgG "backpack" tumors that provided 2G12 antibodies continuously, we found that a sustained dimer concentration of 5-25 µg/ml during the course of infection provides effective protection against HIV-1. Importantly, 2G12 dimer at this concentration does not favor mutations of the HIV-1 envelope that would cause the virus to completely escape 2G12 neutralization. We have therefore identified dimeric 2G12 as a potent prophylactic reagent against HIV-1 in vivo, which could be used as part of an antibody cocktail to prevent HIV-1 infection.

Figure 1. Protection against HIV-1 infection in humanized mice by passively transferred 2G12 dimer.

Humanized mice were injected (i.v.) with 0.5 mg of purified 2G12 monomer (n = 6) or 2G12 dimer (n = 5) at 4 months of age and challenged by HIV-1_JR-CSF (i.v.; 400 ng of p24) one day after passive transfer. (A) Concentrations of 2G12 in the mouse plasma. 2G12 was measured by Myc-specific ELISA and data are shown as average ± s.e.m. All data shown from now on are average ± s.e.m. unless a bar graph is shown, where the data are presented as average + s.e.m. (B) The ratios of human CD4:CD8 in the peripheral blood. CD4 and CD8 cell populations were measured weekly by flow cytometry. The ratios of CD4:CD8, after normalization to week-0 values (set as 1), were plotted. Mice with no antibody but HIV-1 challenge (n = 7) and mice with neither antibody nor HIV-1 (n = 6) were used as controls. The CD4:CD8 ratio in this humanized mouse model increased with the age of the mice as seen in the “No HIV; No antibody” control. Therefore, all mice involved in this study were age-matched. One-way analysis of variance (ANOVA) with Tukey's multiple comparison posttest showed that the group with 2G12 dimer was not different from the No HIV-1 control (p>0.05); however, it had significantly higher CD4:CD8 ratios than the group with 2G12 monomer (p<0.05). All data points in each treatment group, regardless of time, were included in the ANOVA. (C) Viral load in the mouse plasma. Viral RNA was extracted from mouse plasma and the viral load was measured. The detection limit of the assay was 20,000 HIV-1 copies/ml of mouse plasma. The differences were not statistically significant. (D) Total viral load from week 0 to week 4. Area under the curve (AUC) of the viral load during week 0-4 was calculated and plotted. The group with 2G12 dimer had significantly lower viral load than the other two groups based on individual Mann-Whitney tests (p = 0.0303 between 2G12 dimer and No antibody; p = 0.0159 between the 2G12 dimer and 2G12 monomer). (E) Percentages of human CD4 and CD8 T cells in the spleen after 4 weeks. Plots shown were pre-gated on CD3 T cells. (F) Immunohistochemical analysis of HIV-1 p24 in the spleen and mesenteric lymph node after 4 weeks. The images were taken using an Olympus BX51 microscope with 400× magnification (Scale bar, 25 µm).

Figure 2. Generation of humanized mice with IgG “backpack” tumors.

Rag2^−/−γ_c ^−/− mice were intrahepatically (i.h.) injected with 0.1-0.2×10⁶ human CD34⁺ hematopoietic stem and progenitor cells at 1 day of age. The humanized mice were screened for CD45⁺ human cells at 6 weeks of age, and mice with good reconstitution were selected for the following procedures. First, to achieve a sustained level of anti-HIV-1 neutralizing antibodies (NAb) in mice, we delivered the antibodies through subcutaneous (s.c.) injection of a cell line on the right side of the lower back when the mice were 3-months-old. The cell line, 293T/TK/NAb, formed controllable packs on the lower back of the mice. The backpack size was closely monitored biweekly and the drug ganciclovir was injected (i.p.) after HIV-1 challenge and when the backpacks exceeded the size limit of 1.5 cm². The concentrations of 2G12 produced in the blood were monitored by ELISA. Second, to establish HIV-1 infection in these mice, the JR-CSF strain of the virus was injected intravenously (i.v.) at a dose of 400 ng p24 when the backpacked mice were approximately 4-month-old. The infected mice were monitored weekly for the percentages of T cell populations, the HIV-1 viral load, and the concentrations of 2G12 in the blood. They were sacrificed 4 weeks after HIV-1 inoculation and the blood and tissues were analyzed.

Figure 3. Protection against HIV-1 infection in humanized mice by a sustained low level of dimeric 2G12.

Backpacked mice were generated as described in Figure 2. Backpacks expressing wild-type 2G12 were named “2G12 BP” (n = 8) whereas the ones expressing D2 mutant were named “D2 BP” (n = 7). Another group of mice (n = 7) were made to carry large wild-type 2G12 backpacks (“BP”) that generated >100 µg/ml of 2G12 monomer plus dimer in the blood before HIV-1 inoculation. The concentrations of total 2G12 and 2G12 dimer are shown in Table 1. (A) Fold differences of %CD4 T cells in the peripheral blood from week 0 to week 1, week 2, or week 4. CD4 T cells were measured weekly by flow cytometry. One-way ANOVA with Tukey's multiple comparison posttest showed that the group with D2 BP had a significantly higher percentage of CD4 T cells than the “HIV-1; No Ab” control (n = 7) at week 1 (p<0.05). The BP group had significantly higher percentages of CD4 T cells than the “HIV-1; No Ab” group for week 1, week 2, and week 4 (p<0.05). (B) Viral load in the mouse plasma. The detection limit of the assay was 20,000 HIV-1 copies/ml of mouse plasma. The differences were not statistically significant. The virus was not detectable in two different mice in the D2 BP group at week 1 and week 4, respectively. One mouse in the BP group also had undetectable viral load at week 4. The D2 BP mice had an averaged 90% reduction in viral load from the “HIV; No Ab” control at week 1 and 4. At week 2, the averaged reduction was 70%. (C) Number of p24⁺ cells in the mesenteric lymph node (mLN). Four weeks after HIV-1 challenge, mLNs were harvested, fixed, and sectioned for immunohistochemical analysis of HIV-1 p24. The numbers of p24⁺ cells were counted manually and presented as the number of cells per mm² area of the specimen. One-way ANOVA with Tukey's multiple comparison posttest showed that both D2 BP and BP groups had significantly lower numbers of p24⁺ cells in mLN than the HIV-1 groups with 2G12 BP or no antibody (p<0.05). (D) Percentage of CD4 T cells in CD3⁺ splenocytes. Mice were sacrificed after 4 weeks and human T cells in their spleens were analyzed by flow cytometry. The groups were not significantly different.

Figure 4. Sequence changes in HIV-1 envelopes and antibody escape.

(A) Mouse-derived HIV-1 envelope sequences (regions of interest). Viral RNA was extracted from mouse plasma 4 weeks after HIV-1 challenge, reverse transcribed, and sequenced as described in Materials and Methods. The HIV-1 envelope sequences cloned from mice were aligned to the wild-type JRCSF envelope sequence. The percentages of viral clones with mutations are shown in Table 2. Three sequences from two representative mice of each group are shown in this panel, with the mouse number shown in parentheses. The sequence numbers are also shown next to the sequences. For better viewing, all relevant Asn codons (N295, N332, N339, N386, N392, N448) and their adjacent Ser/Thr codons with mutations are shown in red and blue, respectively, except for the wild-type JRCSF sequence where the wild-type codons are colored. The mutated nucleotides are also bolded and underlined. Some nucleotides other than those that form the carbohydrate anchors were also mutated and they are in black color and underlined. (B) Antibody escape. Four representative envelope genes (labeled with ** in panel A) were subcloned into an expression plasmid for the in vitro neutralization assay. The viral envelope from a D2 BP mouse had a single mutation of N386T. This mutation also occurred in 2G12 BP mice. The viral envelope from a BP mouse had a mutation of N295S. The viral envelope from an HIV-1-only mouse contained no mutations. The envelope of the virus injected to mice (the input virus) was used as the control. Pseudoviruses were made from these envelopes and in vitro neutralization assay was performed. Neutralization of pseudoviruses by the 2G12 monomer is shown in this panel. (C) Pseudoviruses were made and in vitro neutralization assay was performed as described in panel B. Neutralization of pseudoviruses by the 2G12 dimer is shown in this panel.