The Biological Activity of AAV Vectors for Choroideremia Gene Therapy Can Be Measured by In Vitro Prenylation of RAB6A

Abstract

Choroideremia (CHM) is a rare, X-linked recessive retinal dystrophy caused by mutations in the CHM gene. CHM is ubiquitously expressed in human cells and encodes Rab escort protein 1 (REP1). REP1 plays a key role in intracellular trafficking through the prenylation of Rab GTPases, a reaction that can be reproduced in vitro. With recent advances in adeno-associated virus (AAV) gene therapy for CHM showing gene replacement to be a promising approach, an assay to assess the biological activity of the vectors is of the uttermost importance. Here we sought to compare the response of two Rab proteins, RAB27A and RAB6A, to the incorporation of a biotinylated lipid donor in a prenylation reaction in vitro. First, we found the expression of REP1 to be proportional to the amount of recombinant AAV (rAAV)2/2-REP1 used to transduce the cells. Second, prenylation of RAB6A appeared to be more sensitive to REP1 protein expression than prenylation of RAB27A. Moreover, the method was reproducible in other cell lines. These results support the further development of a prenylation reaction using a biotinylated lipid donor and RAB6A to assess the biological activity of AAV vectors for CHM gene therapy.

Keywords: AAV gene therapy; choroideremia; potency assay; prenylation.

Figure 1

Both RAB27A and RAB6A Are Subject to Prenylation by Endogenous REP1 from a HEK293-Cell Lysate (A) Summary table of experimental conditions (1–8) used in prenylation reactions in vitro regarding the amount of total cell lysate (2.5, 5, 10, and 20 μg), concentration of GGT-II (0.5, 1, and 2 μM), and concentration of Rab substrate (RAB27A or RAB6A) (0.16, 0.8, and 4 μM). Positive control (+ve): cell lysate spiked with recombinant fish REP1 (25 nM). (B) Protein expression (human REP1 and β-actin) and biotin incorporation detected in prenylation reaction products following SDS-PAGE and western blot analysis (one example out of three independent experiments). (C) Plots for condition sets assessing biotin incorporation in both RAB27A and RAB6A when different amounts of total cell lysate, concentration of GGT-II, or concentration of Rab substrate were used (n = 3). (D) Summary table of statistical analysis performed in the datasets in (C). Two-way ANOVA tests were run independently for each condition (total cell lysate, concentration of GGT-II, or concentration of Rab substrate) with “condition” and “substrate” as factors. The p values and the significance of each test, as well as the Bonferroni’s multiple comparison test for comparison of RAB27A with RAB6A, are given in detail.

Figure 2

RAB6A Is More Sensitive Than RAB27A to Assess the Biological Activity of Human REP1 following rAAV2/2 Transduction of HEK293 Cells (A) The HEK293 cells were transduced with increasing MOI of rAAV2/2-REP1 (100, 300, 1,000, 3,000, 10,000, 30,000, 100,000, and 300,000). Protein expression (human REP1 and β-actin) and biotin incorporation were detected in prenylation reaction products (20 μg) following SDS-PAGE and western blot analysis (representative image of three independent experiments). (B) Nonlinear regression plot of normalized REP1 (corrected for the corresponding actin levels) per rAAV2/2-REP1 (log gc/cell). Data were analyzed using a sigmoidal four-parameter fit (95% confidence interval; R² = 0.8625). Symbols are mean of 6 replicates ± SEM. C) Nonlinear regression plots of biotin incorporation per MOI of rAAV2-REP1 (log gc/cell). Data was analyzed using a sigmoidal four-parameter fit (95% confidence interval; R² = 0.8873 for RAB6A; R² = 0.8772 for RAB27A). Symbols represent the mean of three replicates ± SEM. RAB6A showed statistically significant higher incorporation of biotin than RAB27A at MOI 10,000 (**p = 0.01) and 30,000, 100,000, and 300,000 (****p ≤ 0.0002) (two-way ANOVA with Bonferroni’s multiple comparison test). (D) Linear regression plots of biotin incorporation in substrate corrected for the untransduced control against the normalized overexpressed REP1 for RAB6A (R² = 0.8959, Y = 18.82 × X + 0.4803) and RAB27A (R² = 0.533, Y = 6.569 × X + 0.9042).

Figure 3

RAB6A Validation as a Substrate for In Vitro Prenylation by Other Cell Lines Protein expression (human REP1 and β-actin) and biotin incorporation were detected in prenylation reaction products following cell transduction, SDS-PAGE, and western blot analysis (two replicates in one experiment). HT-1080 cells (A) and ARPE-19 cells (B) were transduced with rAAV2/2-REP1 (MOI 1,000; 10,000; and 30,000 gc/cell), and prenylation reactions were prepared with 20 μg and 10 μg of total protein, respectively. Positive controls (+ rREP1) were prepared using untransduced cell lysate spiked with a recombinant fish REP1 protein (25 nM for HT-1080; 11 nM for ARPE-19).