Engineering of fragment crystallizable (Fc) domains of therapeutic immunoglobulin (IgG) antibodies to eliminate their immune effector functions while retaining other Fc characteristics has numerous applications, including blocking antigens on Fc gamma (Fcγ) receptor-expressing immune cells. We previously reported on a human IgG2 variant termed IgG2σ with barely detectable activity in antibody-dependent cellular cytotoxicity, phagocytosis, complement activity, and Fcγ receptor binding assays. Here, we extend that work to IgG1 and IgG4 antibodies, alternative subtypes which may offer advantages over IgG2 antibodies. In several in vitro and in vivo assays, the IgG1σ and IgG4σ variants showed equal or even lower Fc-related activities than the corresponding IgG2σ variant. In particular, IgG1σ and IgG4σ variants demonstrate complete lack of effector function as measured by antibody-dependent cellular cytotoxicity, complement-dependent cytotoxicity, antibody-dependent cellular phagocytosis, and in vivo T-cell activation. The IgG1σ and IgG4σ variants showed acceptable solubility and stability, and typical human IgG1 pharmacokinetic profiles in human FcRn-transgenic mice and cynomolgus monkeys. In silico T-cell epitope analyses predict a lack of immunogenicity in humans. Finally, crystal structures and simulations of the IgG1σ and IgG4σ Fc domains can explain the lack of Fc-mediated immune functions. These variants show promise for use in those therapeutic antibodies and Fc fusions for which the Fc domain should be immunologically “silent”.

Enhancing protein–protein interactions is a key therapeutic strategy to ensure effective protein function in terms of pharmacokinetics and pharmacodynamics and can be accomplished with methods like directed evolution or rationale design. Previously, two papers suggested the possible enhancement of protein–protein binding affinity via destabilizing mutations. This paper reviews the results of the previous literature and adds new data to show the generality of the strategy that destabilizing the unbound protein without significantly changing the free energy of the complex can enhance protein–protein interactions for therapeutic benefit. The first example presented is that of a variant of human growth hormone (hGHv) containing 15 mutations that improve the binding to the hGH binding protein (hGHbp) by 400-fold while retaining full biological activity. The second example is that of the YTE mutations (M252Y/S354T/T256E) in the Fc region of a monoclonal antibody (mAb). The YTE mutations improve the binding of the mAb to FcRn at pH 6.0 10-fold, resulting in elongated serum half-life of the mAb. In both cases, (i) chemical titration or differential scanning calorimetry (DSC) showed the mutations destabilize the unbound mutant proteins, (ii) isothermal titration calorimetry (ITC) showed extremely favorable enthalpy (ΔH) and unfavorable entropy (ΔS) upon binding to their respective target molecule compared with the wildtype, and (iii) hydrogen/deuterium exchange–mass spectrometry (HDX-MS) revealed that these mutations increase the free energy of unbound mutant protein without significantly affecting the free energy of the bound state, resulting in an enhancement to the binding affinities. The third example presented is that of the JAWA mutations (T437R/K248E) also located in the Fc region of a mAb. The JAWA mutations facilitate antibody multimerization upon binding to cell surface antigens, allowing for enhanced agonism and effector functions. Both DSC and HDX-MS showed that the JAWA mutations destabilize the unbound Fc, although the complex was not characterized due to weak binding. Enhancement of protein–protein interactions through incorporation of mutations that increase the free energy of a protein’s unbound state represents an alternative route to decreasing the protein–protein complex free energy through optimization of the binding interface.