Determining the feasibility of a biotherapeutic early-on in a project is important and challenging. There’s often no experimental data, not a lot of resources and budget, and many teams resort to back-of-the-envelope calculations or general rules of thumb. Once the project does enter the portfolio, the race is on.   It's critical to have a good understanding as early as possible of what therapeutic parameters are necessary to meet the required dosing regimen and success criteria, even theoretical, and also if therapeutic properties (in the form of mathematical model parameters) are feasible (e.g., affinity, half-life, and format).   Applied BioMath AssessTM was developed to provide an interactive, web-based tool to help more easily and quickly determine the risk of developability of a therapeutic given a target profile (e.g., indication, desired dose amount and frequency, therapeutic mechanism of action and target coverage or inhibition) and biological parameters (e.g., ligand expression, ligand turnover, receptor per cell, receptor turnover, ligand-receptor affinity, etc.).

In this case study, we explore some of the analyses possible in Applied BioMath Assess using a single compartment, anti-ligand model as an example.  Our goal is to determine an acceptable binding affinity and first order half life given prerequisites such as dosing route and frequency for an anti-TNF monoclonal antibody under consideration for patients with rheumatoid arthritis who are on methotrexate.   We have simplified this analysis for illustrative purposes. For example, we will assume that the joint, or disease compartment, is well perfused, so we will combine the joint compartment with the central compartment, and use a one-compartment version of the model. We will also assume that there is no soluble receptor and that the concentration of receptor in the blood compartment is 0.1 nM.  The mechanism of action for the therapeutic is an antagonist, with an assumed success criteria target inhibition of  >95% ligand - receptor inhibition for the entire dosing interval.

  1. Our analysis shows that for an average patient (0.005 nm) or high expressing case (0.05 nM), a binding affinity of about 2.5 pM or tighter,  given a half-life of 14 days, will achieve the success criteria of >95% ligand - receptor inhibition. 

  2. For flare situations, there is no reasonable binding affinity that achieves the success criteria given the dosing regimen, half-life, and solubility.  However, if you extend the therapeutic’s first-order half-life to 21 days and increase solubility to 300 mgs, or apply two SC doses at 150mg per SC stick, then a 1pM binder will suffice.   Additionally, if you change from monthly SC dosing to every other week, a dose of 150mgs and a 4pM binder will suffice.


Learn More About Applied BioMath Assess


Applied BioMath Assess provides both 1-D and 2-D parameter scans. One-dimensional scans are a useful tool if either one parameter range is the desired output of the analysis or if you need to vary one parameter at a time to determine which parameters your system is most sensitive to.  Two-dimensional scans are useful in finding the optimal combination of two model parameters to achieve a success criteria.  In this example, we are trying to determine acceptable therapeutic binding affinity (KD) and first-order half-life given the dosing route and frequency. We begin with a 1-D scan over KD to determine if there is an optimal binding affinity that will work for all patient cases. Leveraging scenario sets, we create a scenario to mimic an average patient and see that for an average case (Figure 1, left), a KD of 2.4pM achieves the success criteria of 95% inhibition. However, in a second scenario,  if we increase the ligand concentration to that typically seen in a flare patient (Figure 1, right), we see that no KD achieves the success criteria for the given parameter values. 

1-D Scans in Applied BioMath Assess

Figure 1: 1-D scans show feasible KD for average patients (left) but no feasible KD for flare patients (right).

Continuing to delve into the flare scenario, let's explore extending half-life from 14 days to 21 days. We see that while receptor inhibition improves, it is nowhere near achieving success criteria, even if we increase Dose to 200mgs (Figure 2, left). What if we increase Dose further? As shown in Figure 2 (right), increasing dose to 300mgs achieves success criteria.

1-D Scans in Applied BioMath Assess

Figure 2: Increase half-life and dose shows varied feasibility results with Dose = 200mgs not feasible (left) and Dose = 300mgs feasible (right).

The 1-D scans showed that binding affinity and Dose impact our ability to achieve the 95% inhibition criteria. Therefore, a 2-D scan scanning over binding affinity and Dose is helpful to determine what combination of those parameters will work best for our patient subsets. As shown in Figure 3, scanning dose from 1 - 200mgs and KD from .001 to 10 nM results in some combinations that achieve 95% inhibition, as shown in the yellow, for average patients (left), but not in the case of the flare (middle). Extending the half life to 21 days (right) while varying KD and Dose results in combinations that achieve the 95% inhibition.  Identifying these parameter ranges and minimum requirements helps guide the decisions the project team must make about whether or not producing this therapeutic is feasible for all patient scenarios. 


2-D Scans in Applied BioMath Assess

Figure 3: 2-D Scans show what combinations of KD and Dose achieve success criteria for average patients (left) and flare patients (middle, right).


Determining the feasibility of developability of a biotherapeutic early-on in a project is challenging.  In this example, our goal was to determine an acceptable binding affinity and first-order half life given a dosing route and frequency for an anti-TNF monoclonal antibody under consideration for patients with rheumatoid arthritis who are on methotrexate. We wanted to target >95% receptor inhibition for the entire dosing interval. Leveraging 1-D and 2-D scans for various patient scenarios in Applied BioMath Assess, we were able to determine that for an average patient a binding affinity of approximately 2.5 pM and a half-life of 14 days will achieve the success criteria. For flare situations, there was no reasonable binding affinity that achieved the success criteria given the dosing regimen, assuming for example that typical affinites range from 0.001 - 10 nM, first-order half-life ranges from 7-30 days, and doses are limited to 150 mgs per SC dose. However, when we extended the half life to 21 days and increased solubility to 300 mg (or two doses of 150 mg), a 1pM binder sufficed. Additionally, changing from monthly SC dosing to every other week resulted in a dose of 150mgs and a 4pM binder achieving success criteria. 

With this information, estimates can be provided to inform the Lead Generation strategy to develop a therapeutic that may work for average to high end patients with RA, and one can provide stakeholders information to consider the dosing frequency and amount for patients with flares. Not shown here is a sensitivity analysis to help identify sensitive model biological parameters that can impact feasibility or prioritize experiments. For example, what would happen to this analysis if the ligand turnover rate is changed from 30 mins to 20mins, or 15 mins?

This is one example of working through a feasibility assessment. The analyses and methods shown here are easily applicable to any biotherapeutic project. Applied BioMath Assess was developed to bring these types of analyses to all biotherapeutic project teams so each is able to better assess early-on the likelihood they will be able to achieve success with their therapeutic.

Schedule a Demo