Five computational developability guidelines for therapeutic antibody profiling

Abstract

Therapeutic mAbs must not only bind to their target but must also be free from "developability issues" such as poor stability or high levels of aggregation. While small-molecule drug discovery benefits from Lipinski's rule of five to guide the selection of molecules with appropriate biophysical properties, there is currently no in silico analog for antibody design. Here, we model the variable domain structures of a large set of post-phase-I clinical-stage antibody therapeutics (CSTs) and calculate in silico metrics to estimate their typical properties. In each case, we contextualize the CST distribution against a snapshot of the human antibody gene repertoire. We describe guideline values for five metrics thought to be implicated in poor developability: the total length of the complementarity-determining regions (CDRs), the extent and magnitude of surface hydrophobicity, positive charge and negative charge in the CDRs, and asymmetry in the net heavy- and light-chain surface charges. The guideline cutoffs for each property were derived from the values seen in CSTs, and a flagging system is proposed to identify nonconforming candidates. On two mAb drug discovery sets, we were able to selectively highlight sequences with developability issues. We make available the Therapeutic Antibody Profiler (TAP), a computational tool that builds downloadable homology models of variable domain sequences, tests them against our five developability guidelines, and reports potential sequence liabilities and canonical forms. TAP is freely available at opig.stats.ox.ac.uk/webapps/sabdab-sabpred/TAP.php.

Keywords: developability guidelines; immunoglobulin gene sequencing; surface charge; surface hydrophobicity; therapeutic monoclonal antibodies.

Fig. 1.

Comparing the CDRH3 length distributions of the 137 CSTs (red), 105,458 human VdH Ig-seq nonredundant CDRH3s (blue), and 551,193 human VdH Ig-seq nonredundant heavy chains (green). The CSTs have a lower median CDRH3 length.

Fig. 2.

(A) CDR vicinity PSH scores across the 137 CST (blue) and human VdH Ig-seq (red) models. The CSTs are underrepresented at higher PSH values. (B) Galiximab (Kyte and Doolittle CDR vicinity PSH score of 167.89) has a large surface-exposed patch of hydrophobicity in its CDRH3 loop. Heavy-and light-chain surfaces outside the CDR vicinity are colored in white.

Fig. 3.

Histograms of 137 CST (blue) and human VdH Ig-seq model (red) values for the (A) PPC and (B) PNC metrics in the CDR vicinity. In both measures, the datasets are biased away from higher scores. (C) Histogram of structural Fv charge symmetry parameter values. Both datasets show a bias away from negative values.

Fig. 4.

The (A) CDR vicinity PSH and (B) CDR vicinity PNC metrics for the combined set of 242 CSTs (light blue) and MedImmune case studies (colored by assigned flag). MEDI-578, MEDI-1912, and MEDI-1912STT all have the CDR vicinity PNC value labeled by an asterisk. Amber and red dashed lines delineate the 242 CST guideline thresholds. Case studies with prohibitive developability issues (MEDI-1912, AB001) are red-flagged for the PSH and PNC metrics, respectively. Engineered versions without developability issues (MEDI-1912STT, AB001DDEN) return to the range of values previously seen in CSTs for all metrics. MEDI-578/1912/1912STT are labeled as M-578/1912/1912STT, and AB-001/008/001DDEN are labeled as A-001/008/DDEN for legibility.