Medicinal Chemistry Research Group
Research Centre for Natural Sciences, Budapest

Current projects and collaborations

A. JAK inhibitors (Hungarian Science Foundation K 116904)

Janus kinases (JAK) are intracellular tyrosine kinases participating in cellular signal transduction. They function as a separate tyrosine kinase attached to different cytokine receptors of the JAK signaling pathways. They were discovered in the beginning of the 90’s, but their critical role in the immune system and haematopoiesis attracted lots of research interest to this kinase family.

From the chemical point of view, small molecules as kinase inhibitors usually bind to the ATP-binding site of the enzyme. Thus, many heterocyclic compounds showing structural similarities with ATP are tested as potent kinase inhibitors.

In 2009 together with Peter Sayesky from the University of Florida, a new efficient JAK2 inhibitor structure: (2-[(diethylamino)methyl]-4-[(3E)-4-{3-[(diethylamino)methyl]-4-hydroxyphenyl}hex-3-en-3-yl]phenol) (G6) was published. In the beginning of the co-operation in 2006, the crystal structure of JAK2 kinase domain was still unknown and just few JAK inhibitors’ structure was available, so the group made a decision to develop a structure-based virtual JAK2 model using the already disclosed crystal structure of JAK3 kinase from the Protein Data Bank via homology modelling. The compound library of the National Cancer Institute (NCI) was screened in silico to find a lead candidate. According to the screen scores, 60 compounds were chosen and 39 tested in human erythroleukemia (HEL) cell growth inhibition assay at low concentration. Among the compounds, G6 showed the highest activity. It was tested on mutant JAK2-V617F in binding assay, and it suppressed JAK2-dependent patological cell growth in vitro (inhibition of HEL cell growth with IC50 =60 nM), ex vivo, in addition to, in vivo. According to the calculations it is suggested, that one of the hydroxyl groups forms a hydrogen bond with the amino group of Leu932 and the carbonyl group of Glu930. Further analysis showed the possibility of a molecular interaction, as a new salt bridge between G6 and Asp994.

After the success of G6, several analogs1,2 were synthetized, but there are still plenty of questions to answer. The already mapped structure-activity relations demonstrated that these stilbene compounds can be modulated to increase their activity and enhance selectivity. G6 lacks any heterocyclic moiety, meanwhile the hitherto reported efficient JAK inhibitors (TG101348, CYT387, INCB018424, AZD1480, CEP-701, AG490) are all heterocyclic (imidazopyridazine, quinoxaline, aminopyridine or aminopyrazine core).

Main goals of the research

  1. Our first goal is to set up an efficient virtual screening protocol for JAK1 and JAK2 ligands. Our preliminary studies indicate that finding JAK ligands by virtual screening is feasible. On the other hand, identifying isozyme selective JAK ligands is challenging owing to the close similarity of the isozymes. We plan to combine docking scores and interaction fingerprint analysis for recognizing isozyme specific ligands. Virtual screening hits – after validation by biochemical and biophysical measurements – will be used as chemical starting points for identifying active and selective JAK1 and JAK2 inhibitors. G6 and its analogues form another set of starting points for isosyme selective ligand optimization.

  2. In order to have deeper insight how ligands interact with JAKs we invoke X-ray structure determination of ligand-JAK complexes. This is based on published protocols. Structural information obtained for the complexes of in-house ligands will contribute to a better understanding of the structural basis of ligand affinity and selectivity.

  3. Structure activity relationships will be set up both for G6 analogues and for compound families from virtual screening. The biological measurements will be done in cooperation with the Institute of Enzymology (binding and functional assays) and with the University of Florida HEL cell assays. Structural information, biochemical and biophysical data together with molecular modeling tools will be used to set up and rationalize structure-activity relationships.

  4. Physico-chemical properties, calculated descriptors (logP, logD, MW,..) and thermodynamic profiling (by Isotermic Titration Calorometric measurements) of ligand-protein binding will be jointly used to estimate pharmaco-kinetic properties.

  5. Compounds will be designed based on the recognized structure-activity and structure-property relationships. Activity, selectivity, and pharmacokinetic profile will guide the design. Using the available literature databases (Beilstein-Reaxys), synthetic routes will be carried out and optimized in larger amount (1-2 g) to enable structural characterization and biological assays.

  6. The above tools will be used to iteratively improve compound quality in terms of activity, selectivity and pharmacokinetic profile. Double digit nanomolar activity, over 10 fold and preferably over 100 fold isozyme selectivity together with drug-like physico-chemical properties are targeted.

  7. Owing to the JAK2 selectivity of G6 it is expected that G6 serves as a starting point for active, JAK2 selective ligands with improved drug-like properties. Optimizations starting from virtual screening hits will target JAK1 selective compounds, as well. The parallel development of JAK1 and JAK2 selective compounds is expected to be mutually corroborative.

B. Fragment based approaches against CNS targets (National Brain Reserach Program KTIA NAP_13)

Fragment based drug discovery (FBDD) became a general approach used by both academic and industry players on a wide variety of targets in the last decade. The foundation of FBDD was provided by two fundamental concepts, the theory of chemical space and molecular complexity. It has been suggested and later shown that the chemical space of low molecular weight compounds is much smaller than that of the more complex molecules. This results that the fragment space could be more effectively sampled by smaller collection of fragments relative to the conventional compound libraries screened in conventional hit finding campaigns (HTS).

Based on their complexity theory Hann et al. concluded that less complex fragments have typically higher chance forming favourable interactions with the binding site as compared to more complex compounds. Screening a limited set of low complexity compounds (fragments) would therefore improve the odds of identifying suitable chemical starting points for even the most challenging targets. Low molecular weight low complexity fragments, however, show limited affinity towards these binding sites. There are multiple conditions to detect the binding of fragments. Theoretically fragments should form high quality polar interactions within the binding site that suggests their low lipophilicity. At the technical level the identification of fragment hits requires high screening concentration that needs highly soluble hydrophilic fragments. Fragment screening therefore yields low complexity polar hits that were previously identified as preferred starting points for medicinal chemistry programs. Since penetration to the brain typically requires compounds with a similar character except for limited polarity we plan to design fragments that allow the effective exploration of CNS relevant chemistry space.

To achieve this goal we design compounds that are overlap with both Rule of three and CNS rule of thumb. These compounds could provide more operational freedom for structural modifications during multidimensional optimizations. In addition, the limited size and lipophylicity of the designed fragments would allow controlling their physicochemical profile more effectively in fragment optimizations.

In summary, fragment based approaches represent a powerful way to identify lead-like small molecule ligands for a number of potential drug targets. These high quality starting points might be readily optimized to the drug discovery sweet spot identified as the potency, ADME (ADmission, Metabolism and Excretion) and toxicology optimum in the property space.

In this project we use fragment based approaches for the validation of new central nervous system (CNS) targets. For this purpose we first design and develop CNS specific fragment libraries. These libraries are used to explore the chemical tractability of the targets, and to identify and develop pharmacological tools by fragment based chemogenomics. Finally, fragment hits identified could serve as starting points of drug discovery programs. In this context we perform early phase optimization of these to viable leads.

Main goals of the research

  1. Design and development of CNS specific fragment library useful for chemogenomics based target identification and validation and fragment based lead discovery for potential CNS targets.
  2. Target identification and validation using fragment based chemogenomics by screening and evaluation of the CNS specific fragment library in phenotypic assays on living cells and potential drug targets.
  3. Fragment based lead discovery on validated CNS targets starting from fragment hits. Optimization of these hits would results pharmacological tools useful for further biological evaluation of the target or advanced hits that would be further optimized to viable leads. The latters are considered as starting points for drug discovery programs.

C. FRAGments training NETwork (H2020-MSCA-ITN-2015 No. 675899 FRAGNET)

The European Union has granted 3.9 M euro to our consortium from academia and SMEs from Hungary, The Netherlands, Spain, Sweden, Switzerland, and UK. The funding enables the consortium to establish the Marie Curie Innovative Training Network (ITN) FRAGNET of 15 PhD students and explore new research methods and applications in the emerging field of FBLD.

In the last ten years, Fragment-Based Lead Discovery (FBLD, also known as Fragment-Based Drug Discovery FBDD) has proven to be an effective approach towards the discovery of small molecule compounds (ligands) that can bind to biological target molecules. The key feature of this approach is that ligand discovery begins with the screening of low molecular weight compounds that have a higher chance of binding to a target in comparison to whole (drug) candidate molecules. This means that small compound libraries are sufficient for finding hits against most targets. Such hits can then be grown or merged to provide lead compounds. Overall, this means a much smaller investment is required in terms of automation and development of compounds in comparison to those required by High Throughput Screening (HTS), thereby enabling small companies and academic groups to participate in the drug hunting efforts. More recently, fragment-based approaches are also used to interrogate biological systems, helping to unravel cellular processes and identify new drug targets.

The project complements to the efforts of the Hungarian research group to extend fragment base approaches towards the development of covalent binders against relevant drug targets. Our team will design and synthesise a reactive ‘warhead’ library and establish techniques for screening covalent binders against several FragNet protein target including protein kinases. Optimization of identified hits will be supported by specific computational chemistry protocols developed in-house for modelling covalent protein binders and fragment hit evolution.

The groups of Iwan de Esch (Vrije Universiteit Amsterdam), Rod Hubbard (University of York) and the RCNS team join their efforts delivering innovative lead compounds with therapeutic potential on high unmet need diseases areas.

Academic beneficiary partners: Iwan de Esch (Coordinator, VU University Amsterdam, NL), Rod Hubbard (University of York, UK), Gyorgy M Keseru (Hungarian Academy of Sciences, HU), Xavier Barril (Universitat de Barcelona, SP). Industrial beneficiary partners: Ben Davis (Vernalis, UK), Gregg Siegal (Zobio, NL), Helena Danielson (Beactica, SE), Wolfgang Jahnke (Novartis, CH). Partner organizations: GSK (UK), Roche (CH), Servier (FR), IOTA (UK) and 24 Media Labs (NL).

Main goals of the research

  1. Virtual Screening Protocol for Fragment Libraries of Covalent Binders
    1. Designing a docking and scoring scheme for fragment sized covalent binders. Binding results complemented with reaction kinetic data will be used.
    2. Designing a computational protocol to extend the covalent fragments to covalent lead like compounds.
    3. Virtual screening of commercially available reactive fragments against various proteins including Janus kinases.
    4. Extending covalent binders identified in c) to leadlike compounds using protocol b).
    Expected Results:
    We are developing computational methods i)to identify fragment sized covalent binders to extend covalent fragments to lead like compounds contributing to the identification of inhibitors of therapeutically relevant proteins including Janus Kinases. The development of computational schemes relies on the availability of ample experimental data. Were they not sufficient for the method development literature data will be used to complement them.

  2. “Warhead Library” of Covalent Fragment Binders
    1. Designing and creating a compound library of fragment sized molecules with reactive warheads: “warhead library”
    2. Establishing technics for efficient detection of covalent binders in a screening setup
    3. Screening of the “warhead” library + virtual hits against various protein targets including Janus kinases
    4. Extending covalent binders of 3) into lead-like compounds
    Expected Results:
    Creation of a general fragment library for screening of covalent ligands. Identifying covalent inhibitors for therapeutically relevant proteins including Janus kinases.
    Efficient detection and structural data are indispensable to find relevant covalent binders and to develop them into lead like compounds. Systems planned to be studied are likely to be appropriate since their X-ray structures are described in the literature. In case we were unable to identify covalent binders of the active site of these systems alternative systems have to be selected.

D. Biomarker screening and personalized medicine for pediatric malignancies with poor prognoses (NVKP_16-1-2016-0037)

Our group has received a National Competitiveness and Excellence Program (NVKP) Grant from the National Research, Development and Innovation Office to support our proposal on the research of pediatric malignancies, which are the leading causes of death for children below 14 years (NVKP_16-1-2016-0037). Although current haemato-oncological treatments can facilitate full recovery, solid tumours are lethal within 10 years for around half of the affected patients. Through a joint research effort with the 2nd Department of Pediatrics of the Semmelweis University and Meditop Pharmaceutical Ltd., we are planning to utilize state-of-the-art DNA sequencing techniques to identify biomarkers that correlate with the occurrence of these malignancies, propose protein targets for their treatment, and ultimately, identify and synthesize drug candidates that display promising effect on the given target. During this project, we will gain novel insight into the molecular and epidemiological background of the development of solid tumors, create a DNA sequencing test for prevention and targeted diagnostics, and hopefully improve the clinical prospects of pediatric patients with solid tumors.

Main goals of the research

  1. Isolation of DNA from solid tumor samples from pediatric patients over the past ten years to determine the genes whose mutations correlate with the development and progression of solid tumors.
  2. Compilation of a minimal list of genes (70-100), to be utilized for the screening of increased risk to these diseases. We will implement this test at the 2nd Department of Pediatrics and will utilize it to screen the relatives of former patients, as well as future patients.
  3. Identification of the mutations leading to clinically relevant overexpression by comparison with the TCGA database of adult samples.
  4. Synthesis and characterization of the best inhibitors for those protein targets, whose genes display the greatest differences.
  5. Testing the best molecules in an experimental model.
  6. Preparation of a Phase II trial for the best molecules on the relevant patients.
  7. Filing patent applications for a) those elements of the gene list which have not been used previously in pediatric oncology, and b) the application of the identified genetic differences as a biomarker of the given inhibitor.
  8. Sharing the results in scientific publications with the scientific community and contributing to the efforts of pediatric cancer awareness.

E. Chemical process development for small molecule active pharmaceutical ingredients

We are involved in the development of new, improved and clean laboratory level chemistry processes in cooperation with our pharma partners.