Thanks to Alain Poyau, Jerôme Nugier, Lucie Levast and Olivier Degoul for their scientific input.
A journey through twist and turns of the bioanalytical strategy supporting clinical development of bi-specific T cell engagers.
Why bispecific T-cell engagers need a specific bioanalytical strategy
The development of bispecific antibodies (bsAb), and T-cell engagers (TCE) in particular, differs from that of conventional monoclonal antibodies (mAbs) owing to their toxicity profile. Toxicity can manifest through multiple mechanisms mostly involving:
- peripheral vascular dysfunctions associated with cytokine release syndrome (CRS)
- neurotoxicity linked to immune-effector associated neurotoxicity syndrome (ICANS)
Mitigating serious adverse events shapes both the dosing strategy, notably with much lower starting doses compared to classical mAb, and the bioanalytical strategy.
Case study: a universal PK and receptor occupancy assay for four bsTCE candidates
One of our sponsors, developing multiple bsTCE for different indications in oncology, including solid tumors as well as hematologic malignancies, initially came to us for developing a universal PK assay for its 4 compounds, and receptor occupancy assay (ROA) in the context of a non-clinical toxicology study.
Finding a safe but efficacious starting dose for first in human
The goal was to determine from this study a starting dose that was safe and showed biological activity for the first-in-human (FIH) administration. Three challenges had to be addressed:
Logistics: the in-life phase of the toxicology study was executed in China, which was a major challenge for receptor occupancy.
Sensitivity: the starting dose in non-human primates (NHP) was very low and required an ultrasensitive technological platform for the development of the PK assay.
Consistency: the PK assay had to appropriately and consistently measure all 4 candidates.
Receptor occupancy on circulating T cells as a direct PD biomarker
Receptor occupancy (RO) assessment in peripheral blood by flow cytometry can inform on the relative degree of target engagement of a therapeutic bsAb to its cell surface targets. Together with drug exposure data, it helps guide the starting concentration to be used in a Phase1 dose escalation (DE) setting.
In this study, RO focused on circulating effector T cells, the common moiety for all 4 bsTCE candidates. Based on available critical reagents, the RO was determined on bound/occupied and total receptors on T cells.
Shipping whole blood from China to France without impacting the read-out
One major constraint for this RO assay development was that blood was collected by the in-life CRO located in China had to be shipped to our laboratories in France for analysis without impacting the final read-out. Our approach was the following:
- We confirmed that antibodies from the RO panel did cross-react with corresponding cell surface antigens on simian blood cells
- We identified a non-competing Ab for the total receptors
- We worked on different solutions for stabilizing whole blood while preserving receptor occupancy, investigating whether red blood cell (RBC) lysis and fixation of samples needed to be done prior or after staining, and if samples could be stored frozen pending flow cytometer acquisition at our lab facilities.
Staining after fixation and freezing proved to have minimal impact on receptor occupancy by the bsTCE.
Method development and validation of the RO assay
Method development and validation focused initially on one compound, then extending to the 4 compounds, juggling between human and NHP matrices in order to minimize animal use according to the 4R principles. Validation consisted essentially in demonstrating that:
- Nnonclinical sample analysis could be done in different runs, by different analysts In case of instrument shortage, samples could be analyzed on a back-up instrument with minimal impact on data precision.
- We were also able to validate a stability period at -80°C that allowed samples to be shipped in one single batch at the end of the nonclinical toxicology study.
Drug exposure: Adapting the bioanalytical strategy to the context-of-use
To understand the exposure-response relationship in nonclinical species and project the exposure of the FIH starting dose, a universal / generic PK assay was developed for measuring free bsTCE in parallel to the ROA.
Universal Simoa PK assay for the nonclinical toxicology study
Given the very low starting dose in the nonclinical toxicology study, our choice inclined towards the fully automated Simoa HD-X ultrasensitive platform from Quanterix. The PK assay relied on capture and detection antibodies specific to the common T cell moiety on one hand and to another domain shared by the 4 drug candidates. With this universal / generic method, we were able to achieve a limited calibration range of 2.5 log, with a lower limit of quantification (LLOQ) ranging from 40-80 pg/mL, depending on the bsTCE candidate.
How the FIH starting dose is set: NOAEL vs. MABEL
In the meantime, while preparing for FIH clinical study, the Sponsor generated anti-idiotype antibodies specific to each of their drug candidates.
For FIH studies, the anticipated starting dose is commonly set at:
- 10-fold lower than the human equivalent dose of the No Observed Adverse Effect Level (NOAEL) or
- 10-50% of the Minimal Anticipated Biological Effect Level (MABEL EC10-50).
Since the bsTCEs developed by ourSponsor were not cross-reactive with NHP orthologs, starting dose was determined based on the MABEL.
Ultrasensitive Simoa PK assay for first-in-human
For this FIH, with one of the bsTCE targeting a tumor antigen expressed on leukemic blasts, a more specific PK assay was developed and validated in human plasma according to the bioanalytical validation guidelines. The starting dose was established at less than 0.5 µg per patient for this bsTCE candidate, corresponding to MABEL EC20. Thus, the PK assay for FIH was also developed on the ultrasensitive Simoa HD-1 technological platform. The objective was to evaluate whether we could achieve a LLOQ close to 1 pg/mL and a dynamic range of 5 log.
Method development parameters tested.
Method development challenged many parameters, including:
- Assay format: capturing with either target moiety and detecting with 2 different anti-idiotype antibodies
- Assay protocol: 3-steps, (sequential incubation with bead-conjugated capture Ab then biotinylated detection Ab finally Streptavidin-coupled enzyme) vs 2-steps, in which the first incubation step includes both capture and detection reagents
- Minimal required dilution (MRD) in human biological matrix that does not interfere with drug quantification
- Reagent concentrations
- Sample volume
Validation outcome
After optimization and validation of the method, the absence of matrix interference in the intended population allowed us to test samples neat and reach a LLOQ in the low pg/mL range. Although the dynamic range was more restricted than expected, dilution linearity was demonstrated to a dilution factor appropriate for measuring samples collected at the peak plasma level (Cmax).
Clinical sample analysis and ISR
During the FIH, more than 500 samples collected from 16 patients were analyzed over approx. 40 analytical runs. Overall incurred sample reanalysis (ISR), done on approx. 13% of samples selected to cover the concentration profile of the dosed study participants, met acceptance criteria, set to 2/3 of the repeats with a percent difference within +/- 30%, as per ICH M10. In addition, based on ISR results, we were able to extend the stability period from 10 months, as established during pre-study validation, to 22 months.
MSD PK assay for phase 1b/2
For a different bsTCE candidate, targeting the same T cell subpopulation but a solid tumor antigen, the starting dose for phase 1b/2 was in the range of the µg per mL. Since ultrasensitivity was not needed for measuring drug exposure, the PK assay was developed on the MSD platform.
We screened multiple anti-idiotype antibodies specific for both arms of the bsTCE, optimized all the development parameters, and determined from selectivity assessment the MRD. Optimized parameters included:
- Antibody concentrations
- Biotinylation ratio of detection Ab
- Sample volume
- Incubation times
We also evaluated robustness across:
- Antibody biotinylation scale-up
- Matrix pool change
- Inter-operator variability
- Incubation time ranges
The method was validated according to ICH M10 guidelines and achieved a sensitivity of 300 pg/mL, a dynamic range of approx. 3 logs and a dilution linearity range appropriate for measuring samples collected at the expected peak of drug exposure without hook effect. In addition, potential interference with co-administered medications was ruled out.
The method was used for the analysis of more than 2,000 clinical serum samples, all within their established stability period, set at 24 months at -80°C during pre-study validation. Among more than 50 analytical runs, less than 10% failed due to either invalid calibration curve or QC and more than 95% of ISR results met acceptance criteria.
Comparison of the three PK assays
| Universal PK assay (nonclinical tox) | FIH PK assay (leukemic blast target) | Phase 1b/2 PK assay (solid tumor target) | |
| Platform | Simoa HD-X | Simoa HD-X | MSD |
| Starting dose context | Very low (NHP) | Less than 0.5 µg per patient (MABEL EC20) | In the range of µg per mL |
| LLOQ / sensitivity | 40-80 pg/mL | Low pg/mL range (objective: close to 1 pg/mL) | 300 pg/mL |
| Dynamic range | 2.5 log | More restricted than the 5-log objective | Approx. 3 logs |
| Samples analyzed | Not reported | More than 500 (16 patients, approx. 40 runs) | More than 2,000 (more than 50 runs) |
| ISR | Not reported | Approx. 13% of samples reanalyzed; acceptance criteria met | More than 95% met acceptance criteria |
| Sample stability | Not reported | 10 months at validation, extended to 22 months | 24 months (at -80°C) |
Immune cell profiling to document the mechanism of action
In parallel, a high-parameter immunophenotyping method was developed to document the mechanism of action (MoA) of the bsTCE in the hemato-oncology indication.
We built a 23-color panel including cell surface and intranuclear markers on a Cytek Aurora spectral cytometer. This comprehensive panel aimed at:
- Characterizing immune cell subpopulations, including T cell subsets and antigen presenting cells
- Determining the level of expression of markers relevant to the MoA
- Evaluating activation and proliferation of the target T cell component
Safety monitoring alongside FIH studies: cytokine release syndrome (CRS)
Given the higher doses administered in this phase 1b/2 study, safety monitoring ran in parallel with the PK analyses. Circulating cytokines that could signal CRS were measured with a validated ELLA method covering a panel of 7 pro-inflammatory cytokines.
What ties it together: a bioanalytical strategy that adapts to the clinical question
A bsTCE program rarely needs a single assay. It needs the right method for each molecule and each stage, from nonclinical toxicology to first-in-human and phase 1b/2. That takes expertise in both cellular and ligand-binding approaches, the flexibility to adapt a method as a program evolves, and a technology park broad enough to support whichever context of use.
Planning a bsTCE program? We would be glad to discuss how to adapt your bioanalytical strategy to your molecule and your next clinical milestone.