Predicting Aberrant Fc-fusion Protein Pharmacokinetics from In Silico Structural Properties and Physiologically Based...

Area of Science:

• Biopharmaceutical Development
• Pharmacokinetics
• Computational Biology

Background:

• Fusion of therapeutic proteins to monoclonal antibody (mAb) Fc domains enhances pharmacokinetics (PK) via FcRn binding.
• Fc-fusion biologics can exhibit unpredictable rapid clearance due to non-specific binding, complicating drug development.

Purpose of the Study:

• To develop a computational approach for predicting and mitigating aberrant PK behaviors in Fc-fusion biologics.
• To extend existing physiologically based pharmacokinetic (PBPK) models using in silico protein properties.

Main Methods:

• Leveraged in silico protein structural properties to enhance a PBPK model for mAbs.
• Employed symbolic regression to identify key in silico properties for scaling PK model parameters.
• Validated the extended PBPK model against independent in vivo plasma PK data in mice.

Main Results:

• The extended PBPK model successfully predicted plasma PK profiles for novel Fc-fusion biologics.
• Achieved a median absolute average fold error (AAFE) of 1.18, indicating a good model fit (values < 2).
• Demonstrated the model's ability to generalize to new biologics within the same class.

Conclusions:

• The computational PBPK modeling approach effectively predicts and de-risks PK variability in Fc-fusion proteins.
• This method can improve the selection of Fc-fusion biologics with enhanced therapeutic value.
• Enables more efficient drug development by anticipating and addressing potential clearance issues.