Viruses | Free Full-Text | Using an In Vivo Mouse Model to Determine the Exclusion Criteria of Preexisting Anti-AAV9 Neutralizing Antibody Titer of Pompe Disease Patients in Clinical Trials

By inergency On Mar 5, 2024

1. Introduction

Adeno-associated virus (AAV)-based gene therapy is a promising treatment for inherited genetic diseases. Due to exposure to wild-type AAV or acquisition from maternal immunity, a relatively large proportion of humans carry circulating antibodies against the AAV capsid. Anti-AAV Neutralizing Abs (NAbs) could block viral infection and reduce the efficiency of tissue transduction with AAV vectors, even at a low titer [1,2]. Hence, patients with preexisting anti-AAV antibodies were excluded in most gene therapy clinical trials. However, in hemophilia B gene therapy, successful AAV5 transduction was achieved in patients with low levels of preexisting NAbs [3], which indicated that anti-AAV NAb-positive patients below a certain level might also benefit from gene therapy. A quantitative assessment of the role of preexisting NAbs in the efficacy and safety of the drug is required for the successful clinical development of AAV gene therapy.

Most data on liver-directed AAV8 or AAV5 capsid-based vectors have been documented and the transduction in the liver assessed using report genes or secreted proteins such as FIX [1,4,5]. Little is known about the impact of preexisting NAbs on the AAV9 vector, which has a broad tropism and transduces many cell types, including cardiomyocytes, hepatocytes, muscles, and the central nervous system [6]. AAV9-SMN1 (named Zolgensma^®) has been approved for intravenous injection for spinal muscular atrophy (SMA) patients; nevertheless, its efficacy and safety data from anti-AAV9 antibody-positive patients are absent.

Pompe disease (PD) is a lysosomal glycogen storage disease caused by acid α-glucosidase (GAA) deficiency, which is characterized by glycogen accumulation in the heart, muscles, and central nervous system. The disease poses a range of clinical symptoms and outcomes, such as infantile-onset PD (IOPD), manifesting with hypertrophic cardiomyopathy, hypotonia, and respiratory insufficiency and progressing to death at early life, or late-onset PD (LOPD), presenting with a progressive myopathy in childhood or later [7,8]. AAV9 vector-mediated gene therapy is expected to introduce the functional GAA gene and produce the GAA enzyme continuously for PD patients. We developed a recombinant vector rAAV9-coGAA for the treatment of Pompe disease, named GC301 in brief. A single intravenous injection of GC301 resulted in an elevation of GAA enzyme activity, reduction of glycogen accumulation, and improvement of pathological changes in Pompe model mice [9]. These results support further clinical evaluation of GC301.

Here, we infused a mouse anti-AAV9 monoclonal antibody (MoAb) in naïve mouse serum into Balb/C mice 2 h before receiving 1.2 × 10¹⁴ or 3 × 10¹³ vg/kg GC301 and explored the role of NAbs on biodistribution and transduction of GC301 in the liver, heart, and muscles. The pharmacokinetics, pharmacodynamics, and cellular responses in combination with in vitro NAb assay validated the effects of preexisting NAbs. These parameters from in vivo models are used for establishing a threshold of preexisting anti-AAV9 NAb level from which acceptable efficacy and safety can be achieved in these NAb-positive patients in future clinical trials.

2. Materials and Methods

2.1. Vectors

rAAV9-coGAA, named GC301 in brief, consists of a codon-optimized human GAA that was transcribed and driven by a constitutive promoter [9]. GC301 used in the study was produced by a scaled production method based on the Bac-to-AAV package system and manufactured as described under GMP condition in Beijing Fiveplus Gene Technology Company (Beijing, China) [10]. Genome titers (i.e., vector genome copies [vg/mL]) of AAV vectors were determined by real-time PCR and with plasmid DNA as standards, as reported previously [9]. The vectors were qualified and characterized the same as batches used for clinical trials.

For the in vitro neutralization assay, an AAV9 vector encoding Gaussia luciferase (rAAV9-EGFP-2A-Gluc) was used as a reporter.

2.2. Cell Culture

HEK293 cell line (ATCC CRL-1573) was maintained under 37 °C, 5% CO₂ condition in DMEM supplemented with 10% FBS (D10) and used for the AAV neutralization assay.

2.3. Animal Studies

Animal protocols were approved by GeneCradle’s Ethical Committee and conducted by certified operators according to the guidelines and regulations on animal experiments. In all experiments, female Balb/C mice aged 4~6 weeks were used and randomly assigned to each group, with 5~6 animals in each group. Animals were passively immunized by tail vein injection with 0~42 μg/mL of AAV9 specific mouse monoclonal antibody (anti-AAV9 MoAb) diluted with naïve mouse serum in a volume of 100 μL, and mice injected with PBS were used as negative control (NC). The anti-AAV9 MoAb tested in our study was generated by us using hybridoma technique from empty AAV9 particle-immunized Balb/C mice. The MoAb is a specifically neutralized AAV9 serotype with an IC50 of 20 ng/mL in vitro. After 2 h, mice received 1.2 × 10¹⁴ (namely high dose, H) or 3 × 10¹³ vg/kg (namely low dose, L) of GC301 by the tail vein route, which is the dose for IOPD (ClinicalTrials.gov Identifier: NCT05793307) and LOPD in clinical trials, respectively. Blood samples were collected from the retro-orbital plexus using capillary tubes before and after vector administration. Mice were euthanized at 4 or 7 weeks post-GC301 administration, and tissues were collected and snap-frozen for additional studies. Mouse spleens were harvested at 4 weeks post-GC301 injection, and cells were isolated freshly for ELISPOT assay.

2.4. Anti-AAV9 Neutralizing Antibody Assay In Vitro

Mouse serum samples were heat-inactivated at 56 °C for 30 min. Neutralization was measured as the reduction in Gluc reporter gene expression from rAAV9-EGFP-2A-Gluc when inhibited by the NAbs present in the serum samples in a 96-well plate. Serum was 2x serial diluted with D10 from 1:20 for a total of 6 to 8 dilutions and mixed with 2 × 10⁸ vg rAAV9-EGFP-2A-Gluc in 50 μL of DMEM containing 0.1% BSA (Sigma, St. Louis, MO, USA). One hour after incubation at 37 °C, a trypsinized single-cell suspension of 20,000 HEK-293 cells in 100 μL of D10 was added. The plates were then incubated at 37 °C under 5% CO₂ for 48 to 72 h. After incubation, 20 μL of supernatant was aspirated and 50 μL of coelenterazine native substrate (Nanolight, Tempe, AZ, USA) was added to each well while protected from light. The relative luciferase unit (RLU) was measured using a GLOMAX 96 microplate luminometer (Promega, Madison, WI, USA). For each AAV NAb assay, negative serum with 2 μg/mL and 0.2 μg/mL of anti-AAV9 MoAb was used as the positive reference to establish the in-house standard, whereas pooled serum samples from naive mouse sera were used as the negative controls.

The neutralizing antibody titer was defined as the reciprocal of the serum dilution causing a 50% reduction of RLU compared to the rAAV9-EGFP-2A-Glu control. The transduction inhibition titer was analyzed by Prism software 8.0 (GraphPad Software, San Diego, CA, USA), and the titer at which a serum inhibited 50% of the transduction (IC50) was calculated using a nonlinear regression.

2.5. Vector Genome Copy Number Detected by Droplet Digital (dd) PCR

DNA was extracted from whole-organ homogenization or EDTA-Na2 anti-coagulate blood using TIANamp genomic DNA Kit (Tiangen Biotech, Beijing, China). The GC301 genome copy numbers in DNA of tissues and blood were determined using a digital droplet PCR assay with primers and probes, the same as that used for GC301 quantification. The primers and probe were coGAA-specific (forward primer 5′-GGCACCTGGTCGTGGAACTC-3′, reverse primer 5′-CCAGCCTCTGATCGGCAAGG-3′, probe [FAM] 5′-TGGCCAGGCTCCACCGCCTT-3′ [TAMRA]) and were synthesized by Sangon Biotech, Shanghai, China. Each sample was tested in duplicates; assay results were reported as mean genome copy number per μg genome DNA.

2.6. GAA Enzyme Activity Assay

Tissues were homogenized using stainless-steel beads in RIPA lysate buffer with protease inhibitors (Solarbio, Beijing, China). Protein concentrations were determined by Bradford assay. GAA activity in the tissue homogenates was measured at pH 3.9 by conversion of the substrate 4-methylumbelliferyl (4-MU) D-glucoside to the fluorescent product umbelliferone [11]. Data were represented as nanomoles of substrate cleaved in 1 h per milligram of total protein (nmol/mg/h). The tissues collected from negative control (received 100 μL of PBS) and positive control (received 1.2 × 10¹⁴ vg/kg of GC301) mice were concluded as process quality control.

2.7. ELISPOT

Peptide libraries (18–20 mer overlapping by 10~12 amino acids [aa]) covering the sequence of the AAV9 capsid VP1 protein or human GAA was synthesized by ChinaPeptide Qyaobio (Shanghai, China) and Scilight-Peptide (Beijing, China), respectively. Mouse spleens were harvested 4 weeks after GC301 injection, and spleen cells were isolated using EZ-SepTM mouse lymphocyte separation medium. Cells with confirmed viability > 90% were tested for antigen-specific responses to peptide pools. A total of 3 × 10⁵ cells were stimulated by the peptide pools with each peptide at the final concentration of 2.5 μg/mL. PMA and Ionomycin were included as a positive control stimulation for all of the samples; mock stimulation (0.25% DMSO) was included as a negative control. All 5 conditions were plated in duplicate for each sample and stimulated for 30–36 h at 37 °C, 5% CO₂. Mouse IFN-γ precoated ELISpot Kit (Dakewe Biotech, Shenzhen, China) were prepared and IFN-γ spot-forming units (SFUs) were developed according to the manufacturer’s instructions. SFUs in each well were enumerated using an automated spot counter (ImmunoSpot CTL S6 Micro Analyzer, Cellular Technology Limited, Shaker Heights, OH, USA). The final results were reported as SFU/million cells. An antigen-specific response equal to or greater than 50 SFUs/million cells and 2-fold over the mock response was positive.

2.8. Statistical Analysis

One-way or two-way ANOVA was performed to determine whether the means and variances were equal across datasets, and the statistical differences between groups were determined by multiple comparisons or two-tailed Student’s t test using GraphPad Prism software (version 8). Values of p < 0.05 were considered statistically significant. Data are expressed as mean ± SD or mean and 95% CI.

4. Discussion

The extensive clinical experiences and the approvals of Luxturna, Zolgensma, HEMGENIX, and Roctavian have attested AAV vectors to be the best options for in vivo gene delivery [12,13]. Despite these successes, not all patients are eligible candidates for this novel approach, and its application was recommended for the AAV immunity-naïve population. Natural AAV infection might occur at a very early stage of life and serum anti-AAV Abs in infants could be obtained from the mother. The estimated seroprevalence for NAb for the different AAV serotypes ranges from 30–90% in the population [14,15,16,17,18]. The presence of NAbs could block AAV transduction and enhance the immunogenicity of AAV vectors; thus, many efforts including capsid engineering, concomitant with immunosuppressive agents, plasma dialysis, or imlifidase treatment to remove immunoglobulin have been made to avoid or prevent immune responses [19,20,21]. In vitro assay measuring anti-AAV NAbs is usually an initial test for patient enrollment, though it cannot evaluate the impacts of NAb on AAV pharmacology and potential harms in vivo. Therefore, we utilized multiple parameters involving the efficacy and safety of the tested drug to determine the exclusion criteria in future clinical trials, based on in vitro and in vivo tests.

Considering that the total neutralizing titer of polyclonal antibodies in serum is measured in practice, both polyclonal and monoclonal antibodies can be used for this purpose. We used a mouse anti-AAV9 monoclonal antibody (MoAb) in naïve mouse serum throughout the study to generate neutralization both in vitro and in vivo and also compared it with human purified immunoglobulin (supplementary document). The in vitro assay for NAb titer measuring was the same as for patient screening in future clinical trials, and the matrix effects of murine and human serum were comparable (Table 1). By transfusion with spiked sera, an animal model with anti-AAV9 NAb titers ranging from n = 20), the anti-AAV9 NAb titers were all

Pompe disease is a neuromuscular disease due to a deficiency of the lysosomal enzyme GAA. Lack of the GAA enzyme causes glycogen accumulation in skeletal muscle, smooth muscle, and the heart. The effective transduction of the GAA gene to the heart and muscle is very important for drug efficacy. The liver is the organ with the most abundant AAV after systematic administration because of its rich blood supply; sustainably expressed GAA in the liver can be secreted into the blood and taken up by muscles and other peripheral organs to reinforce the efficacy [22,23]. As demonstrated in Figure 1, GC301 was widely distributed in the heart, liver, and muscles. The higher level of GAA biodistribution in the heart and liver than in muscles is consistent with our data on the mouse model for Pompe disease [9] and the findings reported by Wilson et al. [24]. Furthermore, we found different impacts of preexisting NAb on AAV transduction in the different tissues for the first time. A low level of NAb (titer

Results from the current study showed that GC301 vector clearance was accelerated in peripheral blood, and transduction was completely blocked in the presence of a high level of preexisting NAb (a titer of about 1:1000 detected in vitro). The GAA activity only slightly increased, and no significant difference was found in mice with preexisting NAb at the titer of about 1:500 compared with NC. In the presence of moderate levels of preexisting NAb, at titers between 1:100 to 1:300 corresponding to the groups pretreated with MoAb at 3 and 7 μg/mL, gene copies and transduction of GC301 decreased in the heart and liver, especially in the liver, but were still significantly enhanced compared with NC group. Gregory D Hurlbut et al. used mice pretreated with saline or serially diluted inhibitory nonhuman primate (NHP) sera as a model to demonstrate that liver AAV8-αgal vector copy number fell off rapidly for titers above ~1:40 at low vector dose (2 × 10¹² drp/kg), whereas for the higher dose (2 × 10¹³ drp/kg), this fall-off occurred for titers above ~1:640 [25]. In our study, we did not observe a dose-dependent difference at a low level of preexisting NAb. GC301 efficiently transduced the liver, heart, and muscles at both high and low dosages in mice after 1.5 μg/mL MoAb-pretreatment, at a titer of ¹³ vg/kg) used in this study is comparable to the higher doses in other studies.

The AAV-specific T cell responses tend to increase when the titer of preexisting NAb exceeds 1:200; whether it poses a safety risk needs to be determined further. Based on these results, we predict that low levels of NAb (Figure S1) and obtained a comparable preexisting NAb with the level of 1:47~1:68 (Table S1) before 1.2 × 10¹⁴ vg/kg dose of GC301 injection, and found that little effect on the transduction efficiency in heart and liver as compared with those of non-pretreated mice on Day 28 (Figure S2). Given that enzyme activity does not need to reach normal levels to clear glycogen storage, a potential limitation of the current study is the use of Balb/C mice instead of a Pompe mouse model, which makes it impossible to evaluate if the lower transduction at higher antibody concentrations still enabled enough enzymes with therapeutic effects to be produced in the tissues. Considering that there is no long-term benefit from ERT in some Pompe disease patients, e.g., patients with high anti-GAA antibodies, and the potential immune responses against AAV9 could be controlled by immunosuppressive regimens which were widely used in gene therapy, the exclusion criteria of preexisting AAV9 NAb could be elevated to 1:200, even to 1:300. In fact, we have observed one IOPD patient with preexisting anti-AAV9 NAb (the titer was 1:168 at screening) achieve apparent clinical benefit after systematic delivery of GC301 at the dosage of 1.2 × 10¹⁴ vg/kg (paper in submit). In recent years, the important role of preexisting neutralizing antibodies in activating complement was observed [26,27,28]; the risk should be monitored and mitigated by using adequate immunosuppression strategies when dosing seropositive patients with vectors.

More and more clinical and preclinical data has indicated that low levels of preexisting AAV Ab could be overwhelmed by excessively high vector dosage and have minor effects on the therapeutic gene transduction [3,24,29]. A quantitative assessment of the role of preexisting NAbs in gene therapy efficacy and safety is required for the successful clinical development of AAV vectors. We combined the in vivo model with in vitro tests and applied multiple parameters to evaluate the impact of different levels of preexisting NAb, which could help us to determine the exclusion criteria in future clinical trials. However, caution needs to be placed on host factors that may impact outcomes.