Monoklonale Antikörper als Rezeptor-Analoga für die kombinatorische Chemie : Entwicklung und Evaluation des Konzeptes am Beispiel von GPIIbIIIa-Rezeptor-Antagonisten

In the present work the development and evaluation of a screening assay for combinatorial libraries
on solid-phase by means of fluorescence labelled monoclonal antibodies as well as the development
and optimisation of a linker method for mass spectrometry of short peptides on one single bead is
described.
In a screening-assay the antibody, which has been raised against an antagonist of a receptor serves
as a substitution of the receptor. Since it represents sort of a template of the receptor it is utilised as
a screening agent. On the basis of Lamifiban, a GPIIbIIIa-receptor antagonist from Roche, and the
anti-Lamifiban monoclonal antibody LMF-2, a screening system for compound libraries on solid
phase was developed and optimised. This was then employed for the selection of active
tetrapeptides from Split&Mix-libraries as well as for the stepwise optimisation of Lamifibanmimetics
in Parallel Synthesis. Besides, the quality of the antibody-assay was validated with several
compounds having varying activity from the GPIIbIIIa-project. By means of three different approaches, using a monoclonal antibody that was raised against
Lamifiban, the principle, the applicability and the efficacy of an antibody as a receptor substituent
and as a screening tool was evaluated: Ten compounds with various IC50-values descending from the Lamifiban/GPIIbIIIa-project
were examined with an antibody screening assay. The compounds were coupled onto PEGA
resin, incubated with fluorescence labelled mAB and analysed with the fluorescence microscope
and a suitable image processing software. Afterwards their light intensity values were analysed
statistically. The statistical evaluation resulted in a good correlation with the IC50-values of the
compounds against the receptor as well as with the KD-values measured with BiaCore. With this
investigation the evidence was provided that the monoclonal antibody LMF-2 recognises
Lamifiban and related compounds tendentiously in the same manner as the GPIIbIIIa receptor
does. By the use of tetrapeptide libraries on solid phase which were produced with Split&Mix
methodology, other peptide inhibitors for the GPIIbIIIa-receptor were supposed to be found.The
very first screening with the mAB and the subsequent analysis by Edman-sequencing led to
some short peptide sequences which then were synthesised in milligramme scale. More
thoroughly testing of the affinities of these peptides against the antibody by BiaCore resulted in
only very weak binding (lower mM range). Also in the receptor-assay these peptides showed
only a marginal inhibition. The most active peptides (His-Ile-Phe-Phe-Gly and His-Ile-Ile-Phe-
Gly) showed a IC50-value of 2.6 mM against the receptor. Thus they are appx. 700 times less
active than the natural substrate RGDS (IC50 = 3.7 μM). A synthesis sequence on Solid Phase was developed to produce Lamifiban-analogues. The
compounds should have been analysed individually, analogous to the peptides. However, the Single-Bead-analysis turned out not to be realisable, whereupon a statistical evaluation method
was applied. In order to optimize the Lamifiban-mimetics backbone, an iterative approach was
chosen. The different building blocks were optimised sequentially and analysed with
monoclonal antibody on each step. The subsequent synthesis of the compounds in a larger scale
however proved to be very difficult and the products were not stable. The compounds whose
affinity against the antibody was measured with BiaCore quantitatively, showed only a low
affinity against the antibody. The most active compound showed a KD-value of 300 μM. Yet,
the statistical evaluation of each variation showed tendentiously a match between the active
building blocks and the building blocks of Lamifiban, the hapten of the antibody. A further part of the work included the analysis and the sequencing of short peptides on solidphase-
beads. Therefore, different linker-systems (acid labile linkers and photolabile linkers) were
utilised and tested. However, for Single-Bead analysis they were found to be unsuited. Thus a
linker- and cleavage-method based on a disulfide linker was developed and optimised, which
allowed doing reproducible mass spectrometric analysis of the compounds on one single bead. The
cleavage of the compounds from the solid support over different mechanisms is achieved by
treatment with basic or reductive media (Figure 1). It was possible to reproducibly analyse several short peptides (tetra- up to heptapeptides) on single
beads and to determine their sequences. Furthermore several variations for an optimisation of the
linker synthesis directly on the solid support were realised and tested successfully. The concept of a monoclonal antibody as a receptor substitution seems to work and indeed some
peptides and non-peptide compounds were identified, which really have an affinity against the
monoclonal antibody and the receptor.
However, the problem that will have to be solved is the big variation/standard deviation of the light
intensities within the same bead-species. Due to the narrow bandwidth of the available scale, this
leads to a considerable source of error. That is also the reason why this screening-method is
problematic regarding to Split&Mix-libraries. The light intensity on one single bead is not
representative enough to make a prediction about the compound on the bead.
Another problem is the quality of the utilised reactions. Even small differences in the turnover of
the reaction lead to a substantial corruption of the result. Therefore, the utilised reactions for this
approach should have turnovers comparable to those in peptide chemistry.