Assessment of free dye in solutions of dual-labeled antibody conjugates for in vivo molecular imaging

Abstract

Purpose: Recent preclinical and clinical studies show that dyes that excite and fluoresce in the near-infrared range may be used for tracking and detecting disease targets in vivo. A method for quantifying free dye molecules in antibody conjugate preparations is required for agent batch release and for translation into the clinic.

Procedures: Herein, we developed and validated a SDS-PAGE method to determine the percentage of free IRDye 800 CW in (DTPA)(n)-trastuzumab-(IRDye 800)(m) conjugate sample preparations in which high-performance liquid chromatography (HPLC) assessment of free dye was not possible.

Results: The SDS-PAGE assay was accurate and valid for free IRDye 800 CW amounts between 38 and 4 mol% of total dye. Gel sample preparation reagent affected the specificity of the assay, and lower and upper limits of quantitation and detection were determined.

Conclusion: This method may be applicable to other near-infrared dye-conjugated antibody-based imaging agents in which HPLC assessment of purity is not feasible. This validated method for quality assurance will facilitate the translation of dual-labeled antibody conjugates for nuclear and optical imaging.

Figure 1

(a) Image of 12% gel with bands corresponding to varying picomoles of free dye alone (left) and when added to 16.5 picomoles of conjugate (right). (b) The presence of beta-mercaptoethanol in loading buffer results in an extra band on the separation gel.

Figure 2

Fluorescence intensity per mole of conjugate versus dye-to-protein ratio for varying ratios. Each point corresponds to the mean of 6 samples, and the error bars denote standard deviations. Conjugates with dye/protein ratios higher than 1 exhibit approximately linear fluorescence intensity per mole of conjugate (R=0.8739). Graph shown is representative of three separate measurements of conjugates of varying dye-to-protein ratio.

Figure 3

(a) Alternate (spaced) gel lane use prevents spillover fluorescence from neighboring gel lanes. Each gel lane contains 4.29 x 10⁻¹ pmoles IRDye 800CW. For spaced lanes, the average fluorescence intensity was 2.55 x 10⁶ arbitrary units, with a standard deviation of 13.26 x 10³ units, while for the adjacent lanes, the average fluorescence intensity was 2.50 x 10⁶ units, with a standard deviation of 95.98 x 10³ units. (b) Free dye leakage occurs if time lapse between end of gel run and gel imaging is excessive.

Figure 4

(a) HPLC chromatogram of trastuzumab at 280 nm with a retention time of 9.26 minutes. (b) chromatogram of DTPA at 280 nm with a retention time of 12.39 minutes. (c) chromatogram of IRDye 800CW at 780 nm with a retention time of 12.51. (d) chromatogram of trastuzumab conjugated with DTPA at 280 nm with a retention time of 8.85 minutes. (e) chromatograms of trastuzumab conjugated with DTPA and IRDye 800CW at 280 nm (1) and at 780 nm (2) with a retention time of 8.80 minutes.

Figure 5

Fluorescence intensity versus pmoles of free dye. Symbols denote average of 6 samples and measurements, and error bars correspond to standard deviations. The calibration curve was linear (R=0.9999) for 0.086 to 17.15 pmoles dye, when fluorescence was read at an L2.0 intensity setting of Odyssey. Graphs show three separate experiments which yielded similar results.

Figure 6

Calculations for lower limit of detection. The lower limit of detection (LOD), used to define sensitivity, was measured by assaying 6 replicates of zero (H₂O added instead of dye) four times. The grand mean fluorescence intensity and pooled standard deviation of zero were calculated. The mean of each standard (in linearity assay, Figure 5) was used for the standard curve, and the mean signal of zero plus two standard deviations, read in dose from the standard curve, is the LLD. The lower limit of detection of free IRDye 800CW in this assay was found to be 7.1 x 10⁻³ pmoles.

Figure 7

Determination of upper limit of detection (ULOD). Figure 7 shows that the highest dye amount before encountering pixel saturation (65,535 fluorescence intensity maximum, arbitrary units) and loss of accuracy when measuring fluorescence intensity is 18.5 pmoles.

Figure 8

(a) Actual mean and expected mean free dye percentages (ratios) of combined dye and conjugate samples for a conjugate sample with dye-to-protein ratio=2.27. (b) Actual mean and expected mean free dye percentages (ratios) of combined dye and conjugate samples for a conjugate sample with dye-to-protein ratio=1.01. Precisions are represented by % CVs. n=6. Each graph shows three separate experiments that yielded similar results.