Vectored delivery of anti-SIV envelope targeting mAb via AAV8 protects rhesus macaques from repeated limiting dose intrarectal swarm SIVsmE660 challenge

Abstract

Gene based delivery of immunoglobulins promises to safely and durably provide protective immunity to individuals at risk of acquiring infectious diseases such as HIV. We used a rhesus macaque animal model to optimize delivery of naturally-arising, autologous anti-SIV neutralizing antibodies expressed by Adeno-Associated Virus 8 (AAV8) vectors. Vectored transgene expression was confirmed by quantitation of target antibody abundance in serum and mucosal surfaces. We tested the expression achieved at varying doses and numbers of injections. Expression of the transgene reached a saturation at about 2 x 10(12) AAV8 genome copies (gc) per needle-injection, a physical limitation that may not scale clinically into human trials. In contrast, expression increased proportionately with the number of injections. In terms of anti-drug immunity, anti-vector antibody responses were universally strong, while those directed against the natural transgene mAb were detected in only 20% of animals. An anti-transgene antibody response was invariably associated with loss of detectable plasma expression of the antibody. Despite having atypical glycosylation profiles, transgenes derived from AAV-directed muscle cell expression retained full functional activity, including mucosal accumulation, in vitro neutralization, and protection against repeated limiting dose SIVsmE660 swarm challenge. Our findings demonstrate feasibility of a gene therapy-based passive immunization strategy against infectious disease, and illustrate the potential for the nonhuman primate model to inform clinical AAV-based approaches to passive immunization.

Fig 1. Stable transgene mAb expression following AAV8 administration in NHP.

Rhesus macaques were injected with both AAV8-ITS01 and AAV8-ITS06.02 at 1 x 10¹³ gc/mAb/animal in the right and left quadriceps, respectively. (A) Serum ITS01 (blue lines) and ITS06.02 (red lines) and their sum (black lines) are shown individually for each animal over time following injection. (B) The average and standard deviation of expression levels are shown across all animals. Dashed lines indicate the lower limit of detection by binding to SIVmac239 gp140 (2 standard deviations above negative control sera) for ITS01 (blue), ITS06.02 (red). Asterisks denote statistically significant difference between mean ITS01 and ITS06.02 expression at week 4 (p ≤ 0.05).

Fig 2. Impact of anti-vector immunity on transgene expression.

(A) Anti-vector reactivity and (B) expression of ITS01 is shown for the six animals of Fig 1. Animals with low (A10V030; orange) and intermediate (A10V026; purple) pre-existing anti-vector immunity are highlighted. After 6 months, two of these animals were re-injected with 1 x 10¹³ gc AAV8-ITS01 in the right quadriceps. (C) Anti-AAV8 reactivity was boosted, however (D) no increase in serum ITS01 was observed. Anti-AAV8 measurements are in arbitrary optical density (OD_450nm)/ml units.

Fig 3. Dose and injection site number titration of AAV administration.

(A) Absolute (upper panels) or fractional (lower panels) serum antibody levels of ITS01, ITS06.02 or both are shown as a smoothed fit with 95% confidence interval shading for varying doses (left panels) or shots (right panels). At the lowest dose, data for a single animal is shown (the other one animal showed no expression of transgene). Animals received 2 injections into the quadriceps with AAV8-ITS01 and/or AAV8-ITS06.02 at low, medium, high doses (2 x 10¹¹ (n = 1; green), 2 x 10¹² (n = 24; cyan), 2 x 10¹³ (n = 6; violet) gc/animal, respectively). Controls received no injections (n = 16; gray). For varying shots, animals received 2 x 10¹² gc/animal spread across 1, 2, 4, 8 injection sites (1: R quadriceps (n = 2; orange), 2: both quadriceps (n = 24; cyan), 4: both quadriceps and gastrocnemii (n = 2; magenta), 8: both quadriceps, gastrocnemii, deltoids and biceps (n = 2; blue), respectively). (B) Fractional mucosal transgene IgG is shown for individual macaques in mucosal compartments (n = 28). A scale adjustment is an x-axis break to emphasize early mucosal accumulation denoted by diagonal hashes between 4 and 5 weeks post AAV. (C) The correlation between serum versus rectal transgene mAb is shown for weeks 2, 4, and 5–24 post AAV. Points are colored by number of shots and point shapes denote dose, low, medium or high as noted in the legend.

Fig 4. Infrequent ADA responses accompany loss of serum transgene mAb.

(A) Serum mAb measurements in injected animals that did (+) or did not (-) develop anti-ITS01 (left) and/or anti-ITS06.02 (right). Hashed lines denote the limit of ITS01 or ITS06.02 positivity (2 SD above mean of negative control serum) at 1.3 μg/ml, blue and 0.8 μg/ml, red, respectively). Asterisks indicate statistical significance (p ≤ 0.05). (B) Polyclonal serum ADA responses were quantified in relative μg/ml equivalents to monoclonal control anti-idiotype antibodies. Hashed lines denote anti-ITS01 or anti-ITS06.02 threshold of positivity set at 1 RU. Dose and AAV treatments are coded by color (Control, low (2 x 10¹¹ gc/animal; blue), medium (2 x 10¹² gc/animal; red), intermediate (5 x 10¹² gc/animal; green), and high (2 x 10¹³ gc/animal; purple)) and shape denotes antibody delivered: control (diamonds), ITS01 (squares), ITS06.02 (circles), ITS01 + ITS06.02 (upward triangles), ITS01 + ITS06.02 + ITS52 (downward triangles). (C) Serum ADA response estimates (hashed lines) and AAV mAb estimates (solid lines) are shown for ITS01 (blue) and ITS06.02 (red) for 3 representative examples of animals that developed no ADA (top panel), ADA against both mAbs (middle panel), or ADA against only one mAb (bottom panel).

Fig 5. AAV-expressed mAb exhibit a distinct glycosylation profile from in vitro purified or naturally secreted mAb.

Glycans were eluted from serum anti-SIV env reactive mAbs from passively immunized or SIV-infected animals and analyzed via capillary electrophoresis. (A-D) Analysis of total grouped glycan species are displayed as a percent of total glycans (E) Individual glycan species of are quantified as a percent of all glycans and compared by derivation. (F-I) Analysis of total grouped glycan species of AAV expressed NAbs are compared across time. (J) Individual glycan species of are quantified as a percent of all glycans and compared across time for AAV derived antibodies. (A-D, F-I) Differences between groups reaching p ≤ 0.001 are denoted by black lines. (E, J) Differences between groups of individual glycans reaching p ≤ 0.001 are described in text.

Fig 6. AAV-expressed serum mAb retains full neutralization activity.

(A) Standard neutralization regression analyses are shown for ITS01 (upper) and ITS06.02 (lower) from purified mAb (red, open circles) or sera from AAV injected rhesus macaques (blue, open squares) (n = 6). Neutralization of SIVsmE660 clones 2A5 (left) and A8 (right) for mAbs and sera (ITS01, week 24; ITS06.02, weeks 5–8) are overlaid. (B) EC₅₀ (ng/ml) and (C) maximum neutralization (%) are listed for ITS01 and ITS06.02 purified or from transduced NHP sera. Antibodies in serum are derived from AAV injected animals at peak (week 4), set point (weeks 5–8) or long term (week 24) post AAV. EC₅₀ are displayed in ng/ml with the following color coding: 10–90 Light Green; 1–9 Yellow. Maximum neutralization values are displayed as percent with the following color coding: 20–49.9% Yellow; 50–74.9% Light Orange; 75–94.9% Orange; 95–100% Red.

Fig 7. Vectored immunoprophylaxis mediates a protective effect against limiting dose SIVsmE660 swarm challenge in rhesus macaques.

(A) Limiting dose intrarectal challenges were administered weekly as denoted in the Kaplan-Meier curve on the x-axis to animals receiving mAb (ITS01+ITS06.02) via AAV in our initial (blue, n = 6) and repeat (green, n = 4) groups or no mAb controls (gray, n = 24). Two animals were censored from analysis after challenge 2 due to false positive viral loads (vertical ticks) (B) Serum anti-SIV env transgene mAb was quantified and plotted (y-axis) versus time (x-axis) for mAb pretreated animals and four contemporaneous control animals. Diamonds indicate timepoint of infection corresponding to animals’ expression level line. Challenges at AID_0.33 are denoted as grey downward facing triangles above. (C) Plasma viral loads (y-axis) are plotted versus time post infection (x-axis).