Going beyond target identification and ligand discovery: connecting AI and cell signalling to design drugs inducing desired pathway modulation for GPCRs
InterAx is setting an industry example of how to better understand the biology of a disease and de-risk drug candidates. Instead of only leveraging large quantities of data with machine learning algorithms, we simultaneously incorporate crucial biological knowledge of the underlying system into the intelligent drug discovery algorithms.
Rigorously building such a bottom-up understanding of the target systems is a crucial milestone on the path to the rational design of effective and safe drug candidates.
How It Works
1. G Protein Coupled Receptors & Ligands
Our platform is applicable to all GPCR targets, indications, and ligand modalities, e.g. full/partial agonists, antagonists, orthosteric or allosteric ligands. We perform all our assays using unlabeled ligands, thereby excluding labeling artifacts.
We express the target GPCR in cell lines containing signaling biosensors. We use a small set of reference ligands to calibrate the signaling responses of the target GPCR.
Our platform is applicable to all GPCR targets, indication or type of compound, e.g. biologics or small molecules, full/partial agonists, antagonists, orthosteric or allosteric ligands. We perform all our assays using unlabeled ligands, thereby excluding labeling artefacts.
We express your target GPCR in cell lines containing signaling biosensors. We use a small set of reference ligands to calibrate the signaling responses of your target GPCR.
2. Time resolved cellular signaling assays
We perform time-resolved cellular signaling assays to obtain kinetic profiles of compounds. This approach captures cellular signaling cascades to a much higher extent than static end-point assays.
We perform time-resolved cellular signaling assays to obtain kinetic profiles of your test compounds. This approach captures cellular signaling cascades to a much higher extent than static end-point assays.
3. Proprietary systems biology models
We generate mathematical models using ordinary differential equations (ODEs) to describe the specific signaling pathways of the target GPCR. These models incorporate crucial biological knowledge, including the signaling pathways modulated by a given target, cross-talks between signaling pathways, signaling amplification mechanisms, and influence of protein expression levels.
Validation, calibration, and numerical simulations of these systems biology models generate more and better information than initially contained in the experimental data. Our models indeed derive mechanistic signaling parameters (i) quantifying the activation of proteins not directly measured (more information) and (ii) which are independent of the experimental conditions and cellular background where the measurements were made (better information).
Additional information provided by the InterAx Systems Biology platform enables early selection of compounds, prediction of therapeutic effects from in vitro data, and AI-based generation of new drug candidates with high efficacy and potency. Examples are available upon request.
We generate mathematical models using ordinary differential equations (ODEs) to describe the specific signaling pathways of your target GPCR. We then analyze the kinetic profiles of your test compounds using this unique systems biology model.
4. Early selection of high quality leads
Our systems biology analysis provides a multitude of novel ligand parameters. We then cluster test compounds based on this analysis. We can easily spot outliers (off-target effects) and identify compounds with the most promising profiles for in vivo testing.
Our systems biology analysis provides you with a multitude of novel ligand parameters. We then cluster your test compounds based on this analysis. We can easily spot outliers (off-target effects) and help you to identify compounds with the most promising profiles for in vivo testing
5. Better prediction of in vivo therapeutic effects
We correlate the new compound parameter profiles to in vivo data – thereby honing in on promising compounds at an early stage. This accelerates the hit-to-lead identification process.
We generate mathematical models using ordinary differential equations (ODEs) to describe the specific signaling pathways of your target GPCR. We then analyze the kinetic profiles of your test compounds using this unique systems biology model.We generate mathematical models using ordinary differential equations (ODEs) to describe the specific signaling pathways of your target GPCR. We then analyze the kinetic profiles of your test compounds using this unique systems biology model.
6. AI-based generation of drug candidates with specific cellular response
InterAx AI models learn from drug-target molecular dynamics and systems biology parameters – to predict which drug chemistries will show the desired efficacy and biological response.
The InterAx machine learning algorithms are trained on proprietary systems biology datasets to understand both the biological effects of GPCR drugs on cell signaling and the chemical-structural rules of drug-target interactions. This union of systems biology and machine learning enables InterAx technology to guide the design of new chemical entities with desired cellular biology signaling parameters.
The InterAx Edge
Full ligand range
No limitations on ligand types
No ligand labeling
Ensuring better predictions
Time-resolved assays
More accurately characterize in-cell
behaviour
behaviour
Asthma Case Study
We exemplified the power of our platform by predicting asthma drug action using our proprietary beta2-adrenergic receptor ODE model.
Selected Supporting Literature
Systems Biology
The European Research Network on Signal Transduction (ERNEST): Toward a Multidimensional Holistic Understanding of G Protein-Coupled Receptor Signaling
Sommer ME, Selent J, Carlsson J, De Graaf C, Gloriam DE, Keseru GM, Kosloff M, Mordalski S, Rizk A, Rosenkilde MM, Sotelo E, Tiemann JKS, Tobin A, Vardjan N, Waldhoer M, Kolb P. (2020) ACS Pharmacol Transl Sci. 3:361-370.
G Protein-Coupled Receptor Signaling Networks from a Systems Perspective
Roth S, Kholodenko BN, Smit MJ, Bruggeman FJ (2015). Mol Pharmacol., 88: 604-16.
A conformation-equilibrium model captures ligand-ligand interactions and ligand-biased signalling by G-protein coupled receptors
Roth S, Bruggeman FJ (2014). FEBS J, 281: 4659-71.
Competing G protein-coupled receptor kinases balance G protein and β-arrestin signaling
Heitzler D, Durand G, Gallay N, Rizk A, Ahn S, Kim J, Violin JD, Dupuy L, Gauthier C, Piketty V, Crépieux P, Poupon A, Clément F, Fages F, Lefkowitz RJ,Reit er E. (2012). Mol Syst Biol. 8:590.
A general computational method for robustness analysis with applications to synthetic gene networks
Rizk A, Batt G, Fages F, Soliman S (2009). Bioinformatics, 25:i169-78
On the Analysis of Numerical Data Time Series in Temporal Logic
Fages F, Rizk A. (2007). Computational Methods in Systems Biology, CMSB’07 Edinburgh. Springer-Verlag LNBI, 4695 :48-63.
Drug design and Computational chemistry
Mechanistic insights into dopaminergic and serotonergic neurotransmission – concerted interactions with helices 5 and 6 drive the functional outcome
Tomasz Maciej Stepniewski, Arturo Mancini, Richard Ågren, Mariona Torrens-Fontanals, Meriem Semache, Michel Bouvier, Kristoffer Sahlholm, Billy Breton and Jana Selent (2021) Chemical Science, Issue 33, The Royal Society of Chemistry.
Ligand with Two Modes of Interaction with the Dopamine D2 Receptor–An Induced-Fit Mechanism of Insurmountable Antagonism
Richard Ågren, Hugo Zeberg, Tomasz Maciej Stępniewski, R. Benjamin Free, Sean W. Reilly, Robert R. Luedtke, Peter Århem, Francisco Ciruela, David R. Sibley, Robert H. Mach, Jana Selent, Johanna Nilsson, and Kristoffer Sahlholm (2020) ACS Chemical Neuroscience 3130–3143.
How Do Molecular Dynamics Data Complement Static Structural Data of GPCRs
Mariona Torrens-Fontanals, Tomasz Maciej Stepniewski, David Aranda-García, Adrián Morales-Pastor, Brian Medel-Lacruz and Jana Selent (2020) Int. J. Mol. Sci. 21(16), 5933.
A Focus on Unusual ECL2 Interactions Yields β2 -Adrenergic Receptor Antagonists with Unprecedented Scaffolds
Scharf MM, Zimmermann M, Wilhelm F, Stroe R, Waldhoer M, Kolb P. (2020) ChemMedChem. 15:882-890.
Technology - Biosensors & Assays
New Insights into Arrestin Recruitment to GPCRs
Spillmann M, Thurner L, Romantini N, Zimmermann M, Meger B, Behe M, Waldhoer M, Schertler GFX, Berger P. (2020) Int J Mol Sci. 21:4949.
Systems NMR: single-sample quantification of RNA, proteins and metabolites for biomolecular network analysis
Yaroslav Nikolaev, Nina Ripin, Martin Soste, Paola Picotti, Dagmar Iber & Frédéric H.-T. Allain (2019) Nature Methods volume 16, pages743–749
Quantification of Molecular Interactions in Living Cells in Real Time using a Membrane Protein Nanopattern
Reichmuth AM, Zimmermann M, Wilhelm F, Frutiger A, Blickenstorfer Y, Fattinger C, Waldhoer M, Vörös J. (2020) Anal Chem. 92:8983-8991.
Real-time trafficking and signaling of the glucagon-like peptide-1 receptor
Roed SN, Wismann P, Underwood CR, Kulahin N, Iversen H, Cappelen KA, Schäffer L, Lehtonen J, Hecksher-Soerensen J, Secher A, Mathiesen JM, Bräuner-Osborne H, Whistler JL, Knudsen SM, Waldhoer M. (2014) Mol Cell Endocrinol. 382:938-49.
Segmentation and quantification of subcellular structures in fluorescence microscopy images using Squassh
Rizk A, Paul G, Incardona P, Bugarski M, Mansouri M, Niemann A, Ziegler U, Berger P, Sbalzarini IF (2014). Nat Protoc., 9:586-96.
Arrestin-1 engineering facilitates complex stabilization with native rhodopsin.
Haider RS, Wilhelm F, Rizk A, Mutt E, Deupi X, Peterhans C, Mühle J, Berger P, Schertler GFX, Standfuss J, Ostermaier MK (2019). Sci Rep, 9:439.
Molecular mechanism of phosphorylation-dependent arrestin activation
Ostermaier MK, Schertler GF, Standfuss J. (2014) Curr Opin Struct Biol. 29:143-51.
