Anion-exchange HPLC assay for separation and quantification of empty and full capsids in multiple adeno-associated virus serotypes

Figures

• 

  Figure 1. Selection of the mobile phase buffer for optimal separation of empty and full capsids (A) Stacked AEX-HPLC chromatograms representing fluorescence signal versus time trace of rAAV6-623 and rAAV6-030 along with their mixture over a gradient of 0%–100% NaCl (pH 9.0) are shown. Each chromatogram shows two overlapping peaks (Rs < 2.0), one for empty AAV (earlier eluting peak) and another for full AAV (later eluting peak) capsids, along with their respective relative quantitation. (B) Representative stacked AEX-HPLC chromatograms representing fluorescence signal versus time trace of rAAV6-623 and rAAV6-030 over 0%–100% of various salts (pH 9.0) as eluants are shown. The retention time of major peak in each sample was compared to calculate net retention time (ΔRT).

• 

  Figure 2. Representative AEX chromatograms representing fluorescence signal versus time trace of rAAV6 samples (A) The fluorescence signals of two different rAAV6 samples, empty (rAAV6-030) and full (rAAV6-623) capsids, along with a mixture containing both are shown. The discontinuous gradient containing an isocratic hold flanked by a linear gradient on each side that was used to separate empty and full AAV6 capsids is also shown. The overall gradient indicates binding, separating, cleaning, and re-equilibration phases. (B) Fluorescence signal versus time trace chromatograms of a full (rAAV6-623), an empty (rAAV6-030), and their three different AAV mixtures are shown. The peak areas of both peaks within each chromatogram of rAAV are plotted against the nominal % empty or full capsid AAV content. The linear relationship (r2 > 0.99) in each case indicates that the peak area responses for either empty or full capsid AAV6 peak are additive and consistent with increasing concentrations of peak areas. (A) and (B) show consistency of peak retention times (< ±2%) for both empty and full capsid peaks in each chromatogram, indicating consistency of the separation method.

• 

  Figure 3. Correlation of % empty and full capsids using orthogonal methods (A) Stacked chromatograms representing fluorescence signal versus time trace for different rAAV6 samples by optimized AEX-HPLC assay are shown. (B) AAV6 % empty and full capsids of these samples obtained by AUC and Cryo-TEM analysis are shown. (C–F) The correlation of % empty and full AAV6 capsids obtained by AEX-HPLC with Cryo-TEM (C and D) and AUC (E and F) are also shown. A linear correlation with an R > 0.90 was obtained across all comparisons. The rAAV6 samples tested contained AAV genomic size ranging from 2.5 to 5.2 kb.

• 

  Figure 4. Robustness of the optimized AEX-HPLC assay (A–C) Robustness of the AEX-HPLC method was evaluated on a mixture of rAAV6-030 and rAAV6-623 by implementing deliberate changes in the column temperature (A; 15°C–50°C), flow rate (B; 0.8–1.2 mL/min), and pH (C; 8.0–9.5) of the mobile phase, and their subsequent effect on peak resolution (between empty and full AAV6 peaks) and peak tailing was monitored. Each chromatogram represents fluorescence versus time trace and indicates the peak resolution (Rs) and tailing factor (Tf) for both empty and full AAV6 peaks.

• 

  Figure 5. Injection and dilution linearity of the optimized AEX-HPLC assay (A and B) AEX-HPLC assay was used to demonstrate the injection volume linearity of rAAV6-030 (A) and rAAV6-623 (B). (C and D) In addition, dilutional linearity of rAAV6-030 (C) and rAAV6-623 (D) with respect to capsid titer concentration was demonstrated. In (A)–(D), main graph contains overlaid chromatograms representing fluorescence signal versus time trace of the sample or its dilutions against injected capsid titer or volume of the sample. The inset in each graph shows a plot of AAV6 empty and full capsid peak area versus the parameter being evaluated. The regression analysis for peak area versus either injected capsid titer or volume in each graph showed r2 > 0.99, indicating linearity of the response.

• 

  Figure 6. Applications of the optimized AEX-HPLC assay (A) The overlaid AEX-HPLC chromatograms representing fluorescence signal versus time trace of two rAAV6 samples (lots 069 and 071) purified from two different production runs using the same process are shown. Lot 071 was shown to contain significantly high % empty capsid compared with lot 069, which was attributed to a defect in starting materials that may have caused impairment in the optimal packaging of the AAV cassette. (B) The overlaid chromatograms for A260 nm and A280 nm versus time trace for lot 071 are shown, which exhibit peak maxima for AAV6 empty and full capsid peak at 280 nm and 260 nm, respectively. The A260-to-280 ratios, 0.67 and 1.32, for these peaks indicated that they correspond to AAV6 empty and full capsids, respectively.24

• 

  Figure 7. AEX-HPLC chromatograms representing fluorescence signals versus time trace of rAAV samples for different serotypes including AAV1, AAV2, AAV3, AAV8 and AAV9 Top left: AAV1. Top right: AAV2. Middle left: AAV3. Middle right: AAV8. Bottom: AAV9. The separation method (gradient profile) for each serotype is also shown. For each serotype, binding of test sample to the column, isocratic hold, and both linear gradients were specifically optimized to allow for optimal separation of empty from full rAAV capsids. Within the results for each serotype, the retention times for AAV peaks were consistent (<2%) among the samples tested.