Development of cell-based assay for detecting replication-competent adeno-associated virus by qPCR

Abstract

Recombinant, non-replicative adeno-associated virus (AAV) containing a therapeutic gene of interest (GOI) is widely used as a vector for gene therapy. One way to manufacture AAV is through triple-transfection of HEK293 cells, with plasmids containing the (1) GOI, (2) replication (rep2) and capsid (cap) sequences, and (3) adenovirus helper sequences. During the manufacturing of AAV, replication-competent AAV (rcAAV) could theoretically be generated via homologous and non-homologous recombination events. rcAAV contaminants could lead to reduced efficacy, or an adverse immunogenic response. Therefore, testing is required by regulatory agencies. However, there is a paucity of literature on this critical assay. Here, we have developed a sensitive, cell-based assay for detection of rcAAV in AAV8 preparations. After transducing HEK293 cells over three rounds with AAV8 in the presence of helper adenovirus 5, we performed qPCR to detect the presence of rcAAV using the rep2 gene as a marker. The optimized assay is performed at a 2 mL scale, minimizes the generation of false-positive results, and achieves a reportable result of 1 rcAAV per 10⁶ vector genomes. The same approach for rcAAV method development can be expanded to all other AAV serotypes, providing means for substantially improving process development and product safety.

Keywords: AAV; adeno-associated viral vector; critical quality attribute; quality control; rcAAV; replication-competent AAV.

Graphical abstract

Figure 1

Schematic describing the finalized rcAAV assay workflow For the first round of transduction (T1), HEK293 cells are transduced with AAV in the presence of Ad5 at MOI 3 with viral combinations listed in Table 1. R2C8 is replication competent and contains rep2-cap8 genes. After 72 ± 2 h, supernatant is harvested from T1. For the second round of transduction (T2), fresh HEK293 cells are seeded and transduced in the presence of fresh Ad5 at MOI 3 with T1 supernatant. After 72 ± 2 h, supernatant is harvested from T2. Similarly, third round of transduction (T3) is set up. After each round, DNA is extracted from the cell pellet, normalized to 5 ng/μL, and then qPCR is set up to measure rep2. In parallel, the “no transduction control” (T0) and other positive, negative, and spike controls are set up.

Figure 2

Schematic of the workflow for defining transduction conditions for AAV HEK293 cells were transduced with R2C8 in the presence of Ad5 at MOI 2 for T1. After 48 and 72 ± 2 h, supernatant and lysate were harvested from T1. A small aliquot of supernatant at each time point was set aside for measuring via qPCR. For the second round of transduction (T2), fresh HEK293 cells were seeded and transduced in the presence of fresh Ad5 at MOI 2 either with T1 supernatant or T1 lysate, from both 48 and 72 h time points. Supernatants were harvested after 48 ± 2 h from T2 for both sources of transduction (48 h supernatant T1 and 48 h lysate T1). Similarly, supernatants were harvested after 72 ± 2 h from T2 for both sources of transduction (72 h supernatant T1 and 72 h lysate T1).

Figure 3

Transduction with supernatant and 72 h incubation were selected as transduction conditions as these led to the most significant drop in ΔCt with transduction round Replication of R2C8 was assayed by qPCR of the rep2 gene. Transduction material source (lysate or supernatant) and incubation time (48 h or 72 h) over three R2C8 concentrations (10², 10⁴, and 10⁶ vg/μL) co-transduced with Ad5 at MOI 2 were studied according to the workflow as shown in Figure 2. The qPCR ΔCt values collected from the supernatant are reported. At R2C8 concentration 10² vg/μL, ΔCt failed to cross the common cycle threshold, yielding no measurement. Error bars represent 95% confidence interval, n = 2 replicates for qPCR ΔCt values. Red and blue represents 48 and 72 h samples, respectively. Solid bars represent T1 samples. Lysate seeding is represented by vertical lines and supernatant seeding by dots.

Figure 4

ΔCt values decreased relative to increasing Ad5 MOI from 0.5 to 4 R2C8 at concentration 10⁶ vg/μL was co-transduced with Ad5 MOI varied from 0 to 10 over one round of transduction. qPCR ΔCt values, assaying the rep2 gene, are reported. Error bars represent 95% confidence interval; MOI 0, n = 20 replicates; MOI 0.5–10, n = 2 replicates for qPCR ΔCt values.

Figure 5

LOD of the assay was established as 10³ vg/μL when co-transduced with Ad5 at MOI3 Changes in qPCR ΔCt value assaying the rep2 gene, over three rounds of transduction (1–3) with varying R2C8 concentration (10²–10⁴ vg/μL) co-transduced with Ad5 at MOI 3 (solid symbols, solids lines [●, ⸺]) and MOI 4 (hollow symbols, dashed lines [○, ---]) are shown. T0 is shown as no transduction control and R2C8 at 10⁶ vg/μL co-transduced with Ad5 was used as a positive control. Error bars represent 95% confidence interval, n = 3 replicates for qPCR ΔCt values.

Figure 6

rcAAV below LOD in AAV8 drug substance qPCR ΔCt values, assaying the rep2 gene, of (A) negative controls, (B) positive controls, (C) spike controls, and (D) AAV8 test article (TA) are presented here. Spike control samples and AAV8 samples were tested in replicates 1–3. Error bars represent 95% confidence interval, n = 3 replicates for qPCR ΔCt values.