Validation data for mAbs and B7-H3 mAbs
Selecting the best candidates from a discovery campaign can be a daunting task, especially with a large data output. When is the best time to run bioassays for antibody development? We say, the earlier the better.
Cell-based functional assays are an important step in validating an antibody candidate. The antibody development process is usually drawn as a straight arrow, but in reality the arrow often forks and swerves depending on what assays are available for the target.
We have worked extensively with Promega’s cell-based bioassays on different targets and we’ve recently presented a webinar on the data. The assays are easy to use and a valuable way to down-select the most promising candidates. Our workflow presented in the webinar outlines how to go from AbTheneum™ output to functional antibody candidates in as little as 10 weeks.
Cell-based functional tests have been used for years to determine the biological activity of a drug. They are inexpensive and fast compared to animal-based assays, so multiple drug leads could be tested at once. The basic principle is outlined in this figure from LakePharma’s Antibody Functional Assays service page.
The target cell is a cell line expressing the antigen of interest. The effector cell responds to the antibody drug bound to the target cell to induce some change. A common cell-based functional test is antibody-dependent cell-mediated cytotoxicity (ADCC), wherein the effector cell lyses the target cell when antibodies are specifically bound to its surface. ADCC is used for a variety of drug targets, and typically the effector cell interacts with IgG antibodies, which are the most common type of antibodies.
Classically, the in vitro assay uses PBMCs (peripheral blood mononuclear cells) from human blood samples to serve as effector cells. In particular, the natural killer (NK) cells induce the killing effect. The bioassays for antibody development work by incubating a target cell with an antibody drug lead, then adding PBMCs and incubate for a few hours. The small percentage of NK cells in the PBMCs will lyse the target cells if the antibody is bound to the membrane. The assay is read by counting the number of dead cells in wells that contained antibody and comparing with wells that had no antibody.
This classic ADCC bioassay has high variability, due to the differences in the samples of PBMCs. Laboratories need to run the assay on multiple PBMC samples to understand the variability in their response. The readout is also often plagued by high background, further complicating the analysis.
Promega has developed a reporter-based ADCC bioassay (along with many other bioassays) that we feel makes the work more convenient and approachable for labs. The thaw-and-use format also makes the assays quick to run. Most importantly, the high variability issues that are common in the traditional ADCC assays are addressed by using stable engineered effector cells.
The main differentiator in the Promega reporter-based assay vs. the traditional ADCC using PBMCs is the effector cells. The effector cells in Promega’s bioassay are engineered Jurkat cells that give a luminescent signal when bound to the antibody on the target cell. SCT has used the reporter-based ADCC bioassay on several different antibodies and the reproducibility and ease-of-use is a huge advantage.
Promega has also a larger line of bioassays for antibody development, including blockade, inhibition, T-cell activation, and a whole line of immune checkpoint bioassays. The assay to use always depends on the target, but more assays are being added all the time.
Many antibody developers see in bioassays for antibody development as a last step before in vivo studies. After all, many labs have equipment and staff ready to run a battery of tests like ELISA, affinity measurement by BiaCore, flow cytometry, epitope binning, etc. prior to testing for function. Our view is that the earlier you can test for function, the better.
With the output from AbTheneum™, it can be difficult to know where to start. Some campaigns yield over 1,000 unique antibody sequences. We have developed a workflow to accelerate the antibody development process and maximize success. The name of the game is diversity. Sequence diversity. If AbTheneum™ output gives 1,000 unique antibody sequences, do we need to reconstruct all 1,000 to find the best lead? No! But choosing the most diverse subset will give the picture, allowing the focus to be on the most different antibodies to validate sequence families.
Here is an example using immune checkpoint target TIGIT:
An AbTheneum™ campaign yielded 456 unique antibody sequences that bind to TIGIT. Those sequences were clustered by sequence similarity on the dendrogram. 36 antibodies were selected, denoted by black or color names below the dendrogram. The 36 antibodies that were selected maximize the sequence diversity to ensure that each antibody in the set of 36 represent a different sequence family.
Traditionally, epitope binning would be used to categorize antibodies in separate bins to ensure selected candidates are diverse. With AbTheneum™, we use the antibody sequence itself as the criteria for diversity. We could use some time and resources to put this set of 36 through assays like ELISA, flow cytometry, affinity measurement, etc. to down-select them even further. Inevitably, the question will remain, are the antibodies functional? To get to that answer quickly, and with the added advantage of Promega’s easy-to-use bioassay format, we put the 36 reconstructed antibodies directly into a TIGIT blockade bioassay. Five antibodies were clear stand-outs, comparable to other antibodies in the clinic by Merck and BMS. In one month, 456 antibody sequences got down-selected to 36 diverse candidates and were reconstructed, then the TIGIT blockade assay revealed 5 antibodies with robust function compared to clinical candidates.
There is another added benefit to performing the functional assay first. By doing the bioassays for antibody development earlier, it does not mean that ELISA, flow cytometry, affinity measurement, and other tests will not be performed. But by performing the functional assay first, the best responding antibodies have validated a few things:
- The reconstructed antibody binds to the antigen: It is always good practice to confirm that your reconstructed antibody binds to your antigen, especially if the reconstructed form has been changed from the native sequence (e.g. chimeric). This is traditionally confirmed by ELISA, but if the antibody does not bind to the antigen, the functional test will show no response.
- The antibody binds to cells: The antibody must bind to the antigen that is expressed on the surface of the target cell. This is traditionally confirmed by flow cytometry, but if the antibody does not bind to the cell surface antigen, the functional test will show no response.
- The antibody has pretty good affinity: The affinity of the antibody on the antigen has to be sufficient to bind to the target cell and induce a functional response. This is traditionally confirmed by SPR or BLI using one of many different commercial machines.
With the 5 best candidates confirmed to be functional, and implicitly confirmed to bind antigen, bind to cell-surface, and have decent affinity, the further downstream assays can be performed without risk of losing time and money.
The next blog post will cover some of the various bioassays from Promega we have used and present the data we have gathered.