Peptide Lead Optimization

A highly effective lead discovery & optimization engine for linear and CLIPS-constrained peptides

2-clips-peptide-arrays

Schematic visualization of the lead optimization proces. Based on an initial lead (left), full replacement analysis using both natural & non-natural amino acids identifies those amino acids in the lead sequence that are essential to binding, as well as the amino acids that contribute to affinity improvement. Iterative synthesis & screening cycles with affinity-optimized sequences result in an optimized candidate peptide.

HiSense is the name for Pepscan’s ultrasensitve array platform for identification and optimization of peptide leads. It is based on spatially defined, high-throughput synthesis of peptides on a solid support (455 wells polypropylene cards). The linear peptides can subsequently be transformed into conformationally constrained CLIPS constructs peptides. Binding of the target (protein, antibody) to the individual peptides can be determined via an ELISA-type read-out, to ultimately identify those CLIPS-peptides that sufficiently represent the protein interaction site of interest.

The HiSense lead optimization process

The lead optimization process involves chemical modifications of a lead peptide in order to improve potency, selectivity or pharmacokinetic parameters. It allows identifying those amino acids in the lead sequence that are essential to binding, as well as the amino acids that result in affinity improvement. In particular, the use of non-natural amino acids in our HiSense-arrays further extends the horizon for exploring new NCE’s beyond the reach of phage-display type libraries that solely rely on the use of natural amino acids.

The HiSense platform has successfully been used to identify (small) peptide binders designed/derived from whole protein sequences, and to create constrained (e.g. CLIPS) derivatives of lead sequences with significantly improved binding activity. We demonstrated in several projects that by applying the CLIPS/HiSense platform technology to selected peptide leads the affinity is increased by orders of magnitude (easily 1000 fold) and hence the bioactivity is strongly improved. As a result, this may lead to the generation of New Chemical Entities (NCEs) and thus opens novel options for IP-protection of existing peptide drugs.


Case report 1 of 2: Development of a 3rd generation HIV-fusion inhibitor

26mer-peptide-sequence-e1433729436754

X-ray of the 26-mer peptide sequence in complex with the 5 helix bundle that mimicks the gp41 target protein

In a recent collaboration with Janssen Pharmaceuticals the combined HiSense/CLIPS-technology platform was successfully used for the development of a 3rd generation HIV-fusion inhibitor. Known inhibitors either have a too low antiviral activity (i.e. Fuzeon), or are too long (38-mer) to be synthesized in a cost-effective manner. Pepscan succeeded to identify a truncated version of T2635, a known HIV fusion-inhibitor with antiviral activity in the low nanomolar range, and similar activity against >10 Fuzeon-resistant HIV-mutants. The ~50 nM antiviral activity of the lead T2635-trunc peptide (26-mer CLIPS-peptide) was successfully optimized in several iterative screening rounds to ~50 pM, simply by combining several individual amino acid mutation that were identified using the Hisense-platfrom technology. Moreover, slight variations in the type of CLIPS-scaffold and the anchor positions in the sequences also resulted in impressive activity improvement of ≥20. For the optimized 26-mer lead peptide, a X-ray crystal structure was obtained (collaboration with W. Weissenhorn, Univ. of Grenoble/France), which identified the new peptide inhibitor to be the first that does neither bind to the Lipid Binding Domain (LBD) nor to the Protein Binding Domain (PBD), but just in between the two. A patent application has meanwhile been filed.


Case report 2 of 2: Identification and optimization of a 2-CLIPS binder to an anti-GPCR antibody

We recently also used the HiSense/CLIPS-technology platform for the structural optimization of lead 2-CLIPS binders to a neutralizing anti-GPCR antibody, that were identified in phage-displayed libraries of 2-CLIPS peptides (collaboration with Bicycle Therapeutics, UK). From a first set of >20 lead binders with the consensus motif ‘XAE’ (where ‘X’ is a hydrophobic amino acid), only two of these showed clear activity in a FACS-competition-assay with the antibody. The activity of the best two lead binders (KD = 60 and 45 nM) was then optimized using HiSense, first by identifying natural replacements, and subsequently also by studying a wide variety of non-natural variants. In total, we identified 4 amino acid mutation for each lead that altogether improved the activity 500-fold, from ~50 nM to ~100pM. This shows that affinities comparable to those of antibodies can easily be achieved with these ‘small molecule’ type binders. These data illustrate the potential of the HiSense-platform for efficient peptide lead optimization, and can be easily applied to a variety of other (therapeutic) protein targets.

Lead Sequence KD (nM) Fold change
Initial lead D10 ACYKFAECANEFTCA 45.2 1
Natural a.a. Improvement 1 D10-01 ECYKFAECANEFTCA 38.8 1.2 fold
Natural a.a. Improvement 2 D10-02 ACYKYAECANEFTCA 17.4 2.6 fold
Improvements 1+ 2 combined D10-03 ECYKYAECANEFTCA 16.8 8.4 fold
Non-natural improvement 1 D10-05 ECYKYX1ECANEFTCA 1.4* 32.9 fold
Non-natural improvement 2 D10-06 ECYKYAECANEX2TCA 1.1* 42.5 fold
Non natural improvement 1+2 D10-07 ECYKYX1ECANEX2TCA 0.087* 518 fold
Table summarizing optimization results for lead D10. Only natural a.a.s and 10 non-natural a.a.s were used. Optimization detected non-natural improvement options in key binding motif and in the second loop. Improvements show good additivity. Please note: X1 and X2 are undisclosed non-natural amino acids