Using additivity and bliss analysis to categorize corrector-corrector interactions Tim Vortherms*, Anne-Sophie Wesse, Arlene Manelli*, Andrew Swensen*, Ashvani Singh*, Gert de Wilde, Bertrand Heckmann, Steven van der Plas, Thierry Christophe, Katja Conrath, Xueqing Wang*, Chris Tse* Galapagos NV, Generaal De Wittelaan L11A3, 2800 Mechelen, Belgium. * AbbVie, 1 North Waukegan Road, North Chicago, IL 60064, US. 62 Placeholder for a QR code. Use if applicable. To view an electronic version of this poster, scan QR code or visit To obtain a QR code reader, go to your device app store and search for “QR code reader.” Corrector combination models ABAB Additivity Model Bliss Model C = A + B – (A*B)C = A + B Figure 6. Additivity model predicts the combined response (C) is equal to the sum of the individual correctors A and B alone. The Bliss model predicts the combined response (C) is equivalent to the sum of probabilities for correctors A and B alone. This model is based on the evaluation of the response of the compound mixture compared to the effect of each compound tested individually using this formula: Additivity index = Observed response (mixture) – Expected response (Cmpd A + Cmpd B) Additivity index can be visualized with 3-D surface plots Additive Synergistic Non Additive Cmpd A Cmpd B Cmpd C Cmpd B Cmpd D Cmpd A This model is based on the evaluation of the response of the compound mixture compared to the effect of each compound tested individually using this formula: Bliss index = Observed response (mixture) – Expected response (Cmpd A + Cmpd B – (Cmpd A x Cmpd B)) Bliss index can be visualized with 3-D surface plots Additivity Model Bliss Model Figure 7. 3-D surface plots of three representative profiles of corrector-corrector interactions using the additivity index. The surface graphs were categorized using the additivity index thresholds where > 0.2 is synergistic, 0.15 to is additive and < is non-additive. Additive Cmpd A Cmpd B Cmpd C Cmpd B Synergistic Non Additive Cmpd D Cmpd A Corrector combinations in primary HBE cells Cell Surface Expression Assays Two cell based assays were used to identify novel correctors that increase the cell surface expression of CFTR F508del. Introduction Figure 5. Three representative profiles of dose response curves generated from the combination matrix cell surface expression (CSE) assay. CSE data were normalized as a percentage of control response using the formula: (Value - Ave neg control)/(Ave pos control - Ave neg control)*100. These data were represented as a family of dose response curves for the indicated compound in the absence or presence of a second corrector. The activity of the second corrector is shown in the bar graph as the mean +/- SEM. Figure 4. Matrix of dose-response curves Cell Surface Corrector Combination Assay Representative Dose Response Curves Conclusions Potentiator Corrector Cystic fibrosis (CF) is an autosomal recessive disease that affects nearly 75,000 people worldwide. CF is caused by mutations in the cystic fibrosis transmembrane conductance regulator (CFTR) protein that results in decreased maturation and function of CFTR, as well as, decreased surface stability and increased proteosomal degradation (Figure 1). To address these underlying causes of cystic fibrosis, two biomolecular activities are required, namely correctors to increase CFTR expression at the cell surface, and potentiators to allow the effective opening of the CFTR channel. Combined, these activities markedly increase CFTR function yielding improved hydration of the lung surface and subsequent restoration of mucociliary clearance. We are developing series of compounds capable of performing both these activities individually, and we report here the identification of novel correctors and their classification based on their mode of action. To this end, a cell surface expression assay was used to better understand the interaction between different corrector series. Using a matrix combination of different correctors and subsequently applying an additivity model, the corrector series interactions were categorized as synergistic, additive or non-additive. Figure 2. CFTR F508del – HRP in CFBE41o-Figure 3. CFTR F508del – Prolink™ in U2OS These two cellular systems were utilized to screen chemical libraries and identify compounds effective in restoration of CFTR F508del surface expression. These cell based assays utilize different moieties to tag the CFTR protein and allow for chemiluminescent detection of cell surface expression. In the first system, CFBe41o- cells were transduced with CFTR F508del that was fused to the horseradish peroxidase enzyme (CFTR F508del-HRP; Phuan, PW, et.al., (2014) Mol Pharmacol 86: 42-51) and the second system utilized the U2OS osteosarcoma cell line stably expressing Prolink™- tagged CFTR F508del that is activated based on the complementation platform designed by DiscoveRx ( catalog #: C3). Figure 1. Schematic of CFTR F508del defects Figure 9. Corrector combination functional activity in primary human bronchial epithelial (HBE) cells from F508del homozygous CF patients. The equivalent current (Ieq) traces shown above were measured in a 24-well TECC system and represents CFTR-mediated current following activation with forskolin (Fsk) and CFTR potentiator as indicated. Quantification of the TECC data is shown in the cognate bar graphs and represents that AUC (Fsk + potentiator) for the indicated treatment condition. Cmpd A, C are representative of corrector combinations that exhibit synergistic interactions in the CSE assay. Cmpd E is from the same chemical series as Cmpd C. Using this additivity model, newly identified chemical series could be characterized based on corrector-corrector interactions being additive/synergistic to each other and series correcting the cell surface expression of CFTR F508del in a non-additive manner. The model permitted rapid analysis of corrector combination effects and lead to identification of series that exhibit positive interactions in the presence of other corrector series. Translation of the positive corrector-corrector interactions was confirmed in primary CF epithelial cells providing the necessary support to further develop the series. Together, this has allowed us to focus our efforts on a targeted set of chemical series for further development. Figure 8. 3-D surface plots of three representative profiles of corrector-corrector interactions using the Bliss index. Note that the Bliss model may overestimate synergism if the compounds do not function in a mutually exclusive manner (eg., the difference between additivity and Bliss models for Cmpd A, D). DiscoveRxG. Lukacs