Improved predictions of drug efficacy

20th June 2018

Samuel Altun, Patrik Forssén & Ian A Nicholls report on cell-based heterogeneous interaction analysis

The continuous development of new experimental technologies and growth in computational power together enable new possibilities in drug discovery and development. Experimental assays can now be designed to better mimic in vivo conditions while still maintaining high experimental resolution. Consequently, more information can be obtained both from the actual experiments and from the analysis of the experimental data.

In this paper we discuss the potential for improved prediction of drug efficacy based on a combination of label-free in flow cell-based assays and heterogeneous computer interaction analysis. To elucidate the potential of this approach, we will compare the interaction profile of a typical homogenous biochemical one-to-one interaction with that of a more complex interaction observed in a cell-based experiment. For this, we have chosen to study Parathyroid Hormone (PTH) interaction with immobilised Parathyroid Hormone Receptor (PTH1R) and Trastuzumab’s interaction with HER2-expressing SKBR3 cells.

For experiments, the label-free Attana Cell 200 Quartz Crystal Microbalance (QCM) biosensor (Fig. 1) was used, which can monitor the interaction of an analyte with cells or molecules immobilised on the biosensor surface as described in several publications. These results were then analysed with a newly developed computer software5,6 based on Adaptive Interaction Distribution Algorithm (AIDA).

In the use of the AIDA strategy presented here, the interaction dissociation rate is first calculated to reveal if there are any heterogeneous interactions. This is calculated because the dissociation rate is independent of the analyte concentration and hence has one fewer degrees of freedom and thereby is most suitable to start the analysis with. Thereafter, the AIDA is used to obtain the number of interactions for each analyte concentration. This information is then used to estimate the interaction rate constants by fitting to the measured sensorgrams one at a time. Finally, all estimated rate constants are plotted and clustered, where each cluster represents a mode of complex formation.

In the first experiment, the representative homogeneous PTH-PTH1R system was used. PTH1R is immobilised on the sensor chip and PTH is flowed over at increasing concentrations. In Fig .2A, the traditional biosensor sensorgrams are displayed showing the association and dissociation phase for the different concentrations. In Fig. 2B, the logarithm of the dissociation signal is plotted against the dissociation time of PTH from the surface, giving as expected an almost linear graph. In Fig. 2C, the logarithm of the association rate for all the concentration against the logarithm of the dissociation rate for the different concentrations is plotted. In this case, all concentrations superimpose in one position in the graph, indicating a homogenous interaction.

In the second experiment, the interaction of Trastuzumab with HER2 expressing SKBR3 cells was analysed. Fig. 3A depicts the sensorgrams for the different concentrations. In Fig. 3B, analysis of the dissociation rate clearly depicts a deviation from linearity, thus indicating a heterogeneous interaction. The interaction analysis in Fig. 3C reveals that at least two different interaction types are present. On the top left, having the lower dissociation rate, is the interaction with the target3, whereas the other is less distinct, most probably arising from off-target interactions with the cell membrane.

We believe that the examples shown here demonstrate the potential for cell-based interaction analysis using the AIDA approach. Off-target interactions can play an important role for drug efficacy, either through positive or negative influences. Generally, off-target interactions tend to affect the in vivo distribution of the drug, but they can also increase the residence time of the drug on the target cell and thereby improve the interaction. Secondly, this tool can also be used to obtain information on the receptor accessibility.

We believe that understanding the heterogeneity in the association phase due to receptor accessibility may be significant for the concentration-dependent performance of the drug and hence should be considered when designing clinical phase 1 studies. The detailed characterisation improves the possibility to optimise the interaction profile by performing changes to the drug candidate that minimise off-target interactions and optimise target interaction. Compared with well-established one-point measurement technologies such as ELISA and flow cytometry, this heterogeneous interaction analysis will be a valuable complementary tool in selection and optimisation.

This paper is from a recently started project called BIO-QC: Quality Control and Purification for New Biological Drugs (grant number 20170059), co-financed by the Swedish Knowledge Foundation and including participants such as Attana, AstraZeneca and Karlstad and Linnaeus Universities.

Patrick Forssen is from Karlstad University, Ian A Nicholls is with Linnaeus University & Samuel Altun is with Attana





To receive the Scientist Live weekly email NewsBrief please enter your details below

Twitter Icon © Setform Limited