3.1. Characteristics of AE-Labeled Antibodies
First, we incubated the antibodies with AE for labeling by following preparing procedure described in Section 2.2
, and then loaded the resulting mixture onto the Sephadex G-25 for elution and collection of target antibodies. To characterize AE-labeled antibodies, RLU and protein concentration from 24 tubes of elutes were measured respectively. As shown in Figure 2
, the RLU peak of AE obviously appeared at the position of tubes 4, 5 and 6. Since the antibodies labeled with AE had the largest molecular weight, they should be first eluted. Thus, the RLU peak at the position of tubes 4, 5 and 6 was contributed by the AE that was labeled onto antibodies. The peak value indicated that a large amount of labeled antibodies were eluted as the major component collected. After that, RLU declined until tube 12, and then appeared slow rise. The decline of RLU suggests that most of the labeled antibody are eluted. A slow rise of RLU (tubes 13–24) indicated that the free AE molecules closely bound to the Sephadex G-25 column were gradually eluted. There are two peaks on the concentration curve. The first peak appeared at the position of tubes 4, 5 and 6, which overlap with the only one RLU peak, while the second appeared alone at the position of tubes 8, 9, 10, 11 and 12. Since there was no interference caused by other impure proteins, it can be concluded with certainty that the first peak was contributed by AE-labeled protein. Correspondingly, the second peak represents unlabeled antibody. Thus, products in tubes 4, 5, and 6 were mixed together for future use.
3.2. Optimization of the Labeling Conditions
In order to obtain the stable and highly-efficient AE-labeled antibodies, we optimized some critical conditions, including the molar ratio of AE to antibody, labeling time, and composition of trigger solution, which might affect the AE labeling process.
A too-low dosage of AE might cause an incomplete reaction with the antibodies and, hence, a low labeling efficiency. In contrast, an excess dosage might cause precipitation, which would affect the labeling efficiency [29
]. Thus, three dosages of AE (high, medium, and low) were selected, corresponding to the molar ratio (AE:antibody) of 5:1, 15:1, and 35:1, respectively. As shown in Figure 3
, when the molar ratio was 35:1 (Figure 3
C), the excess dosage of AE resulted in an obvious RLU-increasing trend after the position of tube 12. This is due to a large amount of free AE that was eluted and not used during the labeling process. The corresponding antibody utilization rate is only 59.4%. When the molar ratio was 5:1, although the antibody utilization rate was 78.2%, the labeling efficiency was relatively low (Figure 3
A). Finally, 15:1 was chosen as the optimal condition (Figure 3
B). Under this condition, the AE labeling efficiency was about 2.03 and the antibody utilization rate was 69.4%.
To determine the optimal labeling time, the labeling reaction time of 10 min, 20 min, 30 min, and 1 h were chosen, respectively. As shown in Figure 4
, labeling efficiency achieved a maximum when the labeling time was 20 min (Figure 4
B), while the antibody utilization rate did not change significantly under different labeling times. The results indicated that either a too short or a too long reaction time would affect the labeling efficiency.
Generally, PBS containing BSA or human IgG was used as the elution buffer for reducing the physical adsorption loss of target protein. However, the presence of BSA or IgG might interfere the determination of target protein concentration by using spectrophotometry method in this study. Therefore, we measured the possible differences between protein-free PBS buffer and PBS buffer containing 0.1% BSA when they were used as elution buffer, respectively. PBS buffer with low ionic concentration was used as control. As shown in Figure 5
, the absence of BSA in elution buffer did not affect the elution amount of the labeled- and unlabeled-antibodies.
The CL behavior of AE can be initiated in the presence of H2
and NaOH. In an alkaline environment, the C9 position of the acridine ring binds to the hydroxide ions in the solution, forming the pseudobase, which is unfavorable for the chemiluminescence of acridinium ester [30
]. There have been reports suggesting that the addition of HNO3
before the CL reaction could create an acidic environment, which inhibits the formation of pseudobase and, furthermore, a suitable surfactant could increase the CL intensity [31
]. Two microliters, at high concentration (6.25 × 10−7
) and low concentration (1.56 × 10−7
), of AE solution was excited respectively by different trigger solutions, and RLUs were collected. Finally, a suitable trigger solution was selected by comparison of the RLU value. As shown in Table 1
, when 0.1 M HNO3
was added before NaOH, RLU from the AE solution was 5–10 times more than that without HNO3
addition. Moreover, when 2% Tween-20 or Triton-100 was added into solution B, RLU value increased in various degrees. Thus, solution A, containing 0.1% H2
and 0.1 M HNO3
, and solution B, containing 0.25 M NaOH and 2% Triton-100, were selected as the trigger solution.
Finally, the optimal conditions for AE labeling and CL reaction were optimized as follows: 15:1 molar ratio of AE to antibodies, 20 min labeling time, 0.1 M PBS buffer (BSA-free) for elution, and a trigger solution added 0.1 M HNO3 and 2% Triton-100. It was noted that the activity of antibodies was also measured under different labeling efficiencies (data not shown), and no significant difference was found. The activity of antibodies after AE labeling were sufficient in the following detection.
3.3. Evaluation of the Labeling System
Based on the optimized conditions above, AE labeling experiments were performed three times. We further assessed the labeling system by using three parameters: labeling efficiency, utilization rate of antibodies, and antibody activity.
The labeling efficiency was defined as the ratio of molar concentration of AE to that of labeled antibodies in the elute mixture (4th, 5th, and 6th tubes). The antibody concentration and AE luminescence of all elutes from three experiments are shown in Figure 6
. We measured RLU values for a series of gradient AE solutions, and used the linear correlation between them to further determine the AE concentration of the sample with known RLU. After calculation, the molar concentration of AE labeled on antibodies was 14.84 ± 0.21 × 10−7
M, and the molar concentration of labeled antibodies was 7.73 ± 0.13 × 10−7
M; thus, the labeling efficiency was 1.92 ± 0.08.
The utilization rate of antibody was defined as the percentage of AE-labeled antibody amount accounted for that of total antibody we provided in advance. The average concentration of labeled antibodies from the three-tube mixed elutes was determined to be 0.116 ± 0.002 mg·mL−1. The corresponding volume was 1500 μL and the total amount of antibody we provided was 250 μg, thus, the average utilization rate of antibody was 69.77 ± 1.19%.
For a comparison of antibody activity before and after labeling, an antibody titer was determined by using ELISA. Generally, an antibody dilution factor corresponding to an absorbance value of 2.1 times that of the background is used as the antibody titer. Thus, our results showed that the antibody titers before and after labeling were both about 1:107
). This means that the immunological activity of the labeled antibody was not influenced significantly by the labeling process.
Homogeneity of labeling process was determined finally. According to the separation principles of a Sephadex chromatographic column, the components would encounter different resistance when passing through the column because of their different molecular weight. Based on this, AE-labeled antibodies with larger molecular weight were first eluted and mainly in the 4th, 5th, and 6th tube, while free AE with minimal molecular weight was, finally, eluted. We wondered whether the antibodies in the front tubes would be labeled with more AE molecules. In fact, it was found that there was no significant difference in labeling efficiency of antibodies between the three tubes, which indicated that the labeling process was stable and homogeneous.
3.4. Establishment of GMP-CLIA for Detection of HE4
After successful preparation of AE-labeled antibodies, we next aimed to develop a CLIA system for detection of HE4, a new biomarker of human ovarian cancer. In the clinic, the significant HE4 concentration in human serum for diagnosis is about 3.5 ng·mL−1
. In order to obtain a sufficiently high sensitivity that might be suitable for clinical diagnosis at the nanometer level, not only a strong signal source, but also an excellent immobilized carrier is needed. For the latter, GoldMag particles with a core-shell structure and large surface area-to-volume ratio were selected. Our previous work reported their application in quantitative detection of hsCRP in human serum, although the detection sensitivity was only achieved at the microgram level [23
To establish the detection system, the amount of magnetic particle, antigen, antibody, and immune reaction time were optimized successively according to our previous report. The standard calibration curve was successfully obtained for RLU values against HE4 concentrations of 0.25–50 ng·mL−1
. As shown in Figure 7
, there is a good linearity and the square of the correlation coefficient is 0.9996. The detection limit and functional sensitivity of the present approach is 0.084 ng·mL−1
and 0.25 ng·mL−1
, respectively (Table 3
). The results demonstrated that the proposed method could be used for the determination of HE4. Moreover, the whole detection is finished in less than one hour, and the sensitivity of the present method is similar with that of commercially-available test kit from Roche (based on ECLIA) and higher than that from CanAg (based on ELISA).