Biotherapeutics are generally large molecules with a heterogeneous complex structure that is directly related to their ability to modulate biological reactions. This precludes defining a meaningful structure–activity relationship. Despite this, one needs assays for potency and stability. Regulators require that at least some of the assays mimic the expected method of action (MOA). Testing in living systems (animals or cells) presents the smallest leap of faith for demonstrating a similar MOA.
Cost and technical considerations are reducing whole organism (animal) testing. Precision of animal assays is generally poor, and thus one is seldom sufficient in definitive decision support and process troubleshooting. Usually, the results of several different bioassays are grouped to provide a consensus measure of potency or stability. Admittedly, bioassays produce results that are imprecise, but bioassays are also today’s best-accepted technology, particularly for efficacy.
For this reason, manufacturers and regulators are interested in developing bioassays that measure the efficacy and stability of drug products. “Bioassays and Bioanalytical Method Development,” held October 7–9, 2013, in Berkeley, CA, attracted over 60 scientists to talk shop on bioassays. Lectures discussed:
- Assay development and comparability
- Method transfer
- Critical reagent management, including life cycle planning
- Adverse event response.
The small audience and 45-minute lecture slots encouraged discussion. Bioassays and Bioanalytical Method Development 2013 was clearly a working meeting.
Assay development and comparability
All the heterogeneity associated with living systems affects bioassays. When a drug is administered to a patient, portions of the drug may be dispersed to a variety of locations via several routes. Some remains free in the blood, and some binds differentially to a variety of cells. Some is metabolized, or simply excreted. Even for an individual, the results may be affected by diet, health, and time of day. Dr. Peng Luan of Amgen (Thousand Oaks, CA) pointed out that the free drug is in dynamic equilibrium with a heterogeneous population of binding sites. Multiple forms and multiple kinetics produce larger %CV. Individual patients may respond differentially, thus increasing the CV. The FDA wants to compare results from several independent studies.
Assay format is also an important parameter since it relates to dosage. ELISAs are traditionally performed in multiple-well plates, with one reagent on the well bottom. Thus, the analyte needs to contact the cell bottom. Binding may be slow. “Incubation time” is therefore often a strong variable. Dr. Luan presented an example in which incubation time varied from 25 to 400 min and produced a shift in the midpoint of the IC curve of about 0.5 log. Even larger differences are seen for plate-based assays vs Gyros assays (Gyros, Uppsala, Sweden).
Chinnasamy Elango of Regeneron Pharmaceuticals (Tarrytown, NY) compared assays implemented in Gyros assays and multiple-well plates. Despite the improved quantitative precision, regulators seem to be more familiar with plates. Gyros provide much higher throughput (~1000 assays per day). Developing methods is also faster. Dr. Elango reported development of four Gyros-facilitated assays in one month. Results correlate generally high to low within an assay format, but the IC50 values can differ by several logs.
Managing the CRO–innovator relationship
Drug and diagnostic innovators span the spectrum in expertise. All probably consider using contract research organizations (CROs) at some time. Today, big pharma is readjusting its development models to outsource noncritical technology to CROs. Thus assay and data comparability are frequent issues. CROs are independent firms. How can a symbiotic relationship be nurtured and maintained? How can inevitable problems be solved?
Lectures by Dr. Edwin Golez of Shire Pharmaceuticals (Dublin, Ireland) addressed this from the innovator side, while Jing Shi of Bioreliance Corp. (Bethesda, MD) provided the CRO perspective. Some in the audience added color. My distillation is: Inexperienced developers should select CROs on the basis of reputation and relevant experience first. Close proximity is important but secondary. Then start early and commit to communicate and communicate. Both need to plan, and the plans must be mutually consistent.
If the developer has some technology, such as an analytical method that the CRO needs to use or reference, then method transfer should be included in the planning. Golez reports that method transfer training at Shire usually starts with two people from the CRO working in the developer’s lab for two weeks. This is usually sufficient for the CRO’s staff to become proficient in the method, including work flow. Idiosyncrasies of the method, including critical reagents, are described and planned for, as is a validation plan. When all is in place at the CRO, the staff from the developer visits the CRO for additional training and troubleshooting.
These plans should anticipate the need for continuing management, including periodic site visits to the CRO’s lab plus proficiency testing. Control charting is useful. Ordinarily this exercise uncovers unexpected, critical variables (including reagents) that, when recognized and controlled, improve method robustness.
Capacity optimization and independent verification are two reasons for transferring an assay to other labs, including CROs. However, method transfer adds problems and costs. Managing these is a real problem. First, there is the technical problem of transferring a method from one lab to another. Rarely will the receiving lab have the same instruments as the originating lab. Vendors make changes that are not announced to users. For example, electronic filtering may be upgraded to a higher frequency with little notification. The originating lab may have an older-model UV detector than the CRO, or vice versa. One usually learns a lot of chemistry during the process. The high variability inherent in many bioassays compounds the problem. With large %CVs, how should one interpret the significance of differences in results?
Jing Shi, an experienced CRO, discussed how to overcome roadblocks and hurdles to a successful bioassay. She pointed out that most clients wind up seeing the CRO as a partner since the CRO has experience that is complementary to the client. CROs, particularly the larger ones, have experience gained from multiple studies.
Neutralizing antibody assay
Drugs may elicit an immune response in the patient. Typically the patient’s immune system develops anti-drug antibodies (ADAs) to the drug that bind with and thus neutralize (sequester) the drug. Despite poor precision, cell-based assays (CBAs) are often preferred since the assay mimics in vivo conditions. Use of live cells entails raising cells and evaluating their response. Cells often change response characteristics with subsequent generations. Once a cell line is selected, it should be divided into aliquots sufficient for one set of experiments to avoid unnecessary freeze/thaw (F/T) cycles.
Freeze/thaw cycles often change cell performance. Dr. Weifeng Xu of Bristol-Myers Squibb (New York, NY) reported that mass spectrometry aided in developing a CBA by facilitating tracking nutrients and products of the assay. As in most of the lectures, the particular analyte was not identified, apparently due to intellectual property considerations. Still, the reports were useful since they described the chemical problems and how they were solved.
Life cycle planning for assays
Drugs have a life cycle and so do assays. Dr. Mauricio Maia of Genentech (South San Francisco, CA) focused on critical assay parameters for cell-based bioassays. The particular focus was critical reagents such as antibodies, labeled antibodies, and cell lines. Critical reagents are usually not stable indefinitely. With product life cycles measured in decades, one must anticipate changes in the reagents, especially with batches, as well as changes made by vendors of glass and plasticware. I should add that analytical instruments have a life cycle that is a decade or two at the longest. Another factor is that detection limits of instruments improve with time. Featureless baselines obtained in 2000 might show many peaks in 2020.
Key metrics are control charts, GLP protocols, and strict adherence to the acceptance criteria in the method. It should be anticipated that critical reagents, staff, or environment will change. One can explore if small changes to the method can restore performance. If not, one faces method revision and revalidation. Bridging studies are common and should be anticipated. This means reserving sufficient old material to support the bridge. Of course, all of this needs to be meticulously recorded.
Adverse event investigations
Dr. Deborah Wrona of Amgen described postmarket surveillance of adverse events. Since patient samples are involved, Amgen organized the laboratory under Clinical Laboratory Improvement Amendments (CLIA) standards. CLIA compliance is managed by the Centers for Medicare and Medicaid Services (CMS) and different Code of Federal Regulations (CFR) sections than the FDA. However, the goals are sufficiently similar so that conflicting requirements are seldom an issue. Still, idiosyncrasies exist. Some assays are requested less than once per year. Yet CLIA requires proficiency testing (PT) of each analyst twice per year. So, for some assays, the PT load is much higher than the sample load. Proficiency tests are supposed to be blinded, but the low frequency makes the PT orders quite conspicuous. The most critical question is how and when to recognize that the performance of the assay has changed in a material way. If the assay has changed, then it needs to be revalidated.
The process starts when a physician notes an adverse event and contacts Amgen. This generates an Adverse Event number and the patient identity is removed. The physician orders the tests and provides available patient history. Amgen’s lab runs most assays in house, but some are farmed out for various reasons. The company’s staff is also interested in determining if the reported event is unambiguously caused by Amgen’s product, a biosimilar, or something else. The company is responsible for the package insert, which includes therapy guidance and contraindications. Amgen pays for adverse event investigations.
Isolating circulating tumor cells
Technology for isolating circulating tumor cells (CTCs) in blood was described in a booth from SRI International (Menlo Park, CA). CTCs are rare, with only a few per dL. SRI developed the FAST (fiber array scanning technology) cytometer, which provides automated, high-resolution imaging at a rate of 25 million cells/minute. The automated digital microscope stage is used to identify CTCs, including labeled biomarkers to guide personalized cancer therapy. One example involved measurement of Her2 and estrogen receptor (ER) levels of CTCs on triple negative breast cancer for patient stratification. A similar study focused on optimizing treatment with platinum-based drugs for lung cancer.
In operation, the blood is smeared over a plate. The plate is optically scanned to locate labeled CTCs. Once located, the CTCs can be selected for further study.
Assays for monitoring immune response, hypersensitivity, and vaccine safety
Viracor-IBT Laboratories, Inc. (Lee’s Summit, MO) described its services for assays utilizing biomarkers. In May 2012 the FDA announced its Drug Development Tools Initiative, which included recognition of assays employing biomarkers. These are not biomarkers for clinical diagnostics. The focus appears to support their use in bioassays for drug discovery, development, and bioprocessing. Viracor has experience in developing assays for monitoring immune response, hypersensitivity, vaccine safety, and efficacy. Technologies include ELISAs, naturally occurring antibodies (NAbs), cytokines, antibody-dependent cell-mediated cytotoxicity (ADCC), and reverse transcription PCR (rtPCR).
SQI Diagnostics (Toronto, Ontario, Canada) is a new name, at least to me, in biomarker testing. Its particular focus is anti-drug antibodies, including IgG, IgA, and IgM. SQI developed proprietary printing technology to create arrays enabling multiplexed assays of protein, isotype, and subclass in single wells. The array pattern provides about 200 spots/well with interplate %CV of <20%. This facilitates probing each sample with 20 unique biomarkers with 7× replication.
SQI showed three robotic systems for processing samples. SQiD Works provides more than 1000 results/hr, SQiDlite is a benchtop unit with throughput of more than 330/hr, and the manual SQiD-X provides more than 200/hr.
The IIR U.S.A. team deserves special recognition for organizing and planning an outstanding symposium. The task was even more difficult since major pharmaceutical firms appear to restrict description of assays. Biosimilar developers would like to claim equivalency by obtaining the same results using the innovator’s assay protocol. If assays are proprietary, including acceptance criteria, then the biosimilar developer probably will not use the same method, making claims of equivalency difficult to support.
Robert L. Stevenson, Ph.D., is Editor, American Laboratory; e-mail: firstname.lastname@example.org.