Impact of immunization technology and assay application on antibody performance--a systematic comparative evaluation

Abstract

Antibodies are quintessential affinity reagents for the investigation and determination of a protein's expression patterns, localization, quantitation, modifications, purification, and functional understanding. Antibodies are typically used in techniques such as Western blot, immunohistochemistry (IHC), and enzyme-linked immunosorbent assays (ELISA), among others. The methods employed to generate antibodies can have a profound impact on their success in any of these applications. We raised antibodies against 10 serum proteins using 3 immunization methods: peptide antigens (3 per protein), DNA prime/protein fragment-boost ("DNA immunization"; 3 per protein), and full length protein. Antibodies thus generated were systematically evaluated using several different assay technologies (ELISA, IHC, and Western blot). Antibodies raised against peptides worked predominantly in applications where the target protein was denatured (57% success in Western blot, 66% success in immunohistochemistry), although 37% of the antibodies thus generated did not work in any of these applications. In contrast, antibodies produced by DNA immunization performed well against both denatured and native targets with a high level of success: 93% success in Western blots, 100% success in immunohistochemistry, and 79% success in ELISA. Importantly, success in one assay method was not predictive of success in another. Immunization with full length protein consistently yielded the best results; however, this method is not typically available for new targets, due to the difficulty of generating full length protein. We conclude that DNA immunization strategies which are not encumbered by the limitations of efficacy (peptides) or requirements for full length proteins can be quite successful, particularly when multiple constructs for each protein are used.

Figure 1. Regions of proteins selected for peptide and DNA immunization designs.

Each protein is represented in grey with the length in amino acids of the monomeric unit indicated. 2 proteins, TG and A2M are too large to fit proportionally and gaps in the grey band are shown in areas where no designs were selected. Designs selected for peptides are shown in red and designs for DNA shown in blue.

Figure 2. Yields of purified antibody following affinity purification from 80 ml of pooled rabbit sera.

Each symbol represents the yield for each pool of 2 rabbits. There are 3 constructs per protein for each of the DNA-Ab and Pep-Ab methods and a single data point for the pool from full length native immunization (FLP-Abs).

Figure 3. Specific activity of purified antibodies in direct bind ELISA to full length antigen.

All purified antibodies were titered against full length antigen in direct bind ELISA. Data was processed with a four parameter curve fit (XLfit, IDBS, Guildford, UK) and expressed as the quantity of purified antibody required to give an absorbance of 0.5 at 650 nm using a TMB substrate. Lower amounts of antibody required (shown on an inverted scale) are indicative of higher specific activity). Each symbol represents the activity for each pool of 2 rabbits. There are 3 constructs per protein for each of the DNA-Ab and Pep-Ab methods and a single data point for the pool from full length native immunization (FLP-Abs).

Figure 4. Sensitivity of antibodies in sandwich ELISA when paired with an antibody to full length protein.

Standard curves were processed with four parameter curve fitting software and sensitivity expressed as the amount of antigen that could be detected at an OD 650 nm of 0.1 absorbance units above background. Each symbol represents the sensitivity for each pool of 2 rabbits. There are 3 constructs per protein for each of the DNA-Ab and Pep-Ab methods and a single data point for the pool from full length native immunization (FLP-Abs).

Figure 5. Immunohistochemical staining of human colon carcinoma and normal tissue specimens for carcinoembryonic antigen (CEA).

Staining was performed with highest rank antibody R10736 and lowest rank D3305-20 generated by DNA and peptide immunization, respectively. DNA-Ab R01736 was generated against the sequence encoding amino acids 410–500. Pep-Ab D3305-20 was made from a peptide of amino acids 561–570. After full optimization both of these antibodies were found to work best at 2.5 µg/ml, the level shown here. Magnification ranges from 20× to 40×.

Figure 6. Immunohistochemical staining of human prostate carcinoma normal tissue specimens for prostate specific antigen (PSA).

Staining was performed with the highest rank antibody R10733 (2.5 µg/ml) generated by DNA immunization and the lowest rank peptide-antibody D3305-18 (10 µg/ml). DNA-Ab R01733 was generated against the sequence encoding amino acids 37–139. Pep-Ab D3305-18 was made from a peptide of amino acids 126–144. Antibodies are shown at their optimized concentration, 2.5 µg/ml for R10733 and 10 µg/ml for D3305-18. Magnification ranges from 20× to 40×.

Figure 7. Western blot analysis of full-length carcinoembryonic antigen (CEA) run under denaturing conditions.

Each immunoblot was probed with DNA (DNA-Abs) derived anti-CEA antibodies or peptide-derived (Pep-Abs) antibodies at 100 ng/ml and 1∶4000 anti-rabbit HRP. Lane 1 = molecular weight standards (kDa); Lane 2 and 3 = 10 ng and 1 ng of CEA per lane, respectively. DNA-Abs R01736, R01737, and R01738 were generated against the sequences encoding amino acids 410–500, 588–686, and 317–421 respectively. Pep-Abs D3305-19, D3305-20,and D3305-21 were made from peptides of amino acids 616–630, 561–570, and 321–339 respectively.

Figure 8. Success rates in sandwich ELISA using 3 different immunization designs for Pep-Abs or DNA-Abs.

8a) Success is defined as the percent of the time that at least one of the designs will give rise to an antibody that can pair with an FLP-Ab to the target giving sensitivity at or above the indicated concentrations. 8b) Success when at least one of the designs will form a sandwich pair with itself or with another antibody within the group (Pep-Abs with Pep-Abs, DNA-Abs with DNA Abs) to give sensitivity at or above the indicated concentrations.