How to Redesign Mouse Antibodies for Safe Use in Humans

Genentech's 1995 patent on a systematic method for humanizing rodent antibodies by grafting their disease-targeting loops onto a human consensus framework while carefully swapping key structural support residues to maintain binding strength.

What it covers

The patent describes a method to engineer therapeutic antibodies that the human immune system won't reject. It starts with a non-human 'import' antibody (typically from a mouse) that already binds to a disease target, and a human 'consensus' antibody framework (specifically VH subgroup III). The method grafts the mouse's target-binding loops, called Complementarity Determining Regions (CDRs), onto the human framework. To prevent the loops from collapsing, the method aligns the framework sequences and identifies differences. If a mouse framework residue is different from the human consensus, and it is predicted to either bind the antigen directly, interact with a CDR loop, or help the heavy and light antibody chains fit together, that specific mouse residue is substituted back into the human framework. A concrete example is the design of trastuzumab (Herceptin), where specific framework positions like 66L or 93H are retained from the mouse sequence to preserve target affinity.

What it doesn't cover

• —Does not cover humanizing antibodies using human framework regions other than the VH subgroup III consensus sequence.
• —Does not cover fully human antibodies generated from transgenic mice or synthetic phage display libraries that do not require grafting.
• —Does not cover humanization methods that do not substitute framework residues based on the three specific criteria of direct antigen binding, CDR interaction, or VL-VH interface participation.
• —Does not cover simple CDR grafting where no framework region residues are substituted back to the import sequence.

The clever bit

Instead of just copying the binding loops, the inventors realized that the surrounding framework acts like a physical scaffold. If you change the scaffold entirely to human, the loops warp; by identifying a precise set of support residues in the framework to keep as mouse-derived, they preserved the loop shape.

Why it matters

This patent laid the technical foundation for some of the most successful cancer and autoimmune drugs in history, including Herceptin (trastuzumab) and Avastin (bevacizumab). Before this systematic approach, simply grafting mouse binding loops onto human antibodies often caused them to lose their ability to bind targets tightly, making them useless as drugs.

Real-world examples

• 1.Herceptin (trastuzumab) breast cancer therapy
• 2.Avastin (bevacizumab) cancer therapy
• 3.Xolair (omalizumab) asthma treatment
• 4.Genentech's humanized monoclonal antibody pipeline