Skip to main content
NCBI home page
RSC Medicinal Chemistry logoLink to RSC Medicinal Chemistry
. 2025 Apr 22;16(7):3190–3196. doi: 10.1039/d4md01051e

A data-driven journey using results from target-based drug discovery for target deconvolution in phenotypic screening

Gergely Takács a,b, György T Balogh a,c,✉, Róbert Kiss b,✉
PMCID: PMC12062751  PMID: 40352671

Abstract

In drug discovery, various approaches exist to find compounds that alter the different states in living organisms. There are two fundamental discovery strategies regarding the mechanism of action: target-based and phenotypic methods. Both have strengths and weaknesses in assay development, target selection, target validation and structure optimization. While phenotypic screening can identify chemical starting points with the desired phenotype, it is typically difficult to carry out efficient, structure-based optimization without confirming the mechanism of action of such hits. It is therefore critical to uncover the targets behind the phenotype. Target deconvolution is typically carried out by a set of highly selective compounds, where each ligand is associated with a particular target. Hits of such a high-selectivity set can provide valuable information on the phenotype's underlying targets and may also enable novel target-based therapeutic strategies. Consequently, there is a continuously high demand for novel highly-selective tool compounds for target deconvolution. In this work, the ChEMBL database, comprising over 20 million bioactivity data, was mined to identify the most selective novel ligands for a diverse set of targets. A novel method for the automated selection of such high-selectivity ligands is presented. Using these high-selectivity compounds in phenotypic screening campaigns can provide a valuable preliminary direction during target deconvolution. 87 representative compounds were purchased and screened against 60 cancer cell lines. Several compounds were found to possess selective inhibition of cell growth of a few distinct cell lines. The phenotypic assay results, along with the nanomolar activities of individual proteins obtained from the ChEMBL database suggest some novel mechanisms of action for anti-cancer drug discovery.

In drug discovery, various approaches exist to find compounds that alter the different states in living organisms.graphic file with name d4md01051e-ga.jpg

Introduction

Phenotypic screening was the predominant drug discovery methodology in the second part of the 20th century.1 It facilitated the development of therapies without extensive knowledge of complex biological systems, the interaction among various small or macromolecules, signalling and metabolic pathways and their isolation. The significant advancement of molecular biology in the 1980s contributed to the emergence of a novel drug discovery strategy, known as rational or target-based drug discovery, along with its various screening techniques. Such techniques exploit the gathered knowledge of e.g. parallel chemical synthesis, sequencing, crystallisation and modelling tools. Efforts on the identification of novel targets by/with their role in different pathways or disease pathogenesis is still one of the most important aspects of these researches.2,3 Despite all these contributions and the numerous pharmaceuticals developed by these methodologies, the majority of first-in-class drugs approved by the FDA originated from phenotypic assays.4 The primary reason of that is the insufficient therapeutic efficacy or safety issues of the candidates, which only emerge in later stages of drug research due to the complexity of biological systems.5–7

A limitation of phenotypic assays is the lack of known mechanism of actions, which prevents their efficient optimization. On the other hand, numerous established and widely applied target deconvolution technologies exist, each presenting its own challenges.8 Affinity chromatography involves the immobilisation of small molecule tool compounds or other labelling techniques affecting their structures followed by the identification of binding proteins. In activity-based profiling, small molecule tool compounds are used to target specific protein classes. These small molecules generally consist of three parts: a covalent modifier that binds to the enzyme active site, a linker and a tag for efficient separation. Label free techniques do not require any modification of a known binder but rely on the thermodynamics changes of the binding. Such changes can result in the complex's improved resistance against proteolysis9 or oxidative conditions.10 Expression cloning techniques use various ways to increase the amount or expression of the target, providing solutions for scenarios when the target of interest is hardly detectable with the previous techniques.

Increasingly, in silico techniques are offering solutions for the target prediction of small molecules, whether they are hits from phenotypic screens11,12 or established drugs.13.

These in silico approaches rely on the knowledge gathered from the continuously growing number of biological data from target-based screens collected in databases like ChEMBL14,15 or Drugbank.16,17.

While DrugBank focuses on drugs or candidates already entered into clinical trials, ChEMBL collects data of any substance that had an interaction recorded in scientific papers, patents or datasets. There are many layers of such data beside the structure of the substance, e.g. the applied concentrations, the observed activity, and the type and source of the target.

In our work, we systematically analysed the ChEMBL database and developed a workflow to identify the most selective compounds for various targets, resulting in a library suitable for phenotypic screening. This approach yielded hits that assist in identifying targets associated with specific phenotypes and furthermore can be exploited as valuable hits with well-characterised target profiles.

Results and discussion

The ChEMBL database was downloaded, and activities were extracted (over 2.5 million) along with their associated targets, assays and compounds. These activities were filtered and separated into two categories: active data points and inactive data points. Active data points contained those entries that had a pchembl value greater than 6 (active concentration below 1 μM) and the activity_comment field was different from “not active”, “undetermined”, “inconclusive”, and “inactive”. The inactive data point category contained activities with pchembl value less than 5 (inactive at a minimum concentration of 10 μM) and the activity_comment field contained exactly “inactive” or “not active”. The number of such active and inactive data points – which passed these criteria – were 213 913 and 133 421 covering 126 131 and 83 716 unique compounds, respectively.

12 281 unique compounds were found to be purchasable compounds (based on the Mcule database18) that are associated with any active substance derived from the ChEMBL database using the core of their InChI19 representation (using only the main layer). Further filters were applied to exclude compounds that (1) possess substructures known to be PAINS20–22 or (2) were present in any of the widely known drug-repurposing or bioactive libraries, yielding 7718 compounds. From these compounds the most selective ones were selected for each target. The scoring system incorporating both active and inactive data points are as follows:

• Positive score for each active data point reported on its Target.

• Positive score for each inactive data point reported on other Targets.

• Negative score for each active data point reported on other Targets.

• Exclude compound with reported inactive data point on its Target.

The selection procedure is visualized in Fig. 1.

Fig. 1. The selection procedure from the activities in the ChEMBL database until the final 500 purchasable tool compounds.

Fig. 1

Open in a new tab

In total 564 compound-target pairs have been identified and collected. The total calculated selectivity scores and their contributing subscores of those 114 compounds that scored above 4 are visualized in Fig. 2.

Fig. 2. The selected compounds with a minimum total selectivity score of 4 ranked by their total score from left (highest) to right (lowest) and the contribution of their negative, positive-on-target and positive-off-target activities. The higher scores indicate higher selectivity.

Fig. 2

Open in a new tab

The structural relationship among these datasets was visualised using t-SNE algorithm23 to demonstrate the extent to which the ultimately selected tool compounds represent both the set of active and, more critically, the set of selective tool compounds. No bias was found, and the final selection is evenly distributed within the wider purchasable and selective sets in structural terms, as shown in Fig. 3. It was interesting to see that most of the purchasable compounds cover a rather smaller portion of the chemical space of active compounds. Many structurally different clusters (green clusters closer to the edges) are – if represented – only represented by a few purchasable compounds. This phenomenon is probably caused by the bias that chemical suppliers aim for compounds that are easier to be synthesized and have reasonable pricing. Meanwhile, many of the bioactivity data in the ChEMBL database are from scientific papers, submitted by academic institutes where more exotic compounds are being synthesized and screened. However, many of these single purchasable compounds have become part of both the selective and measured subsets.

Fig. 3. Visualization of the compounds along the selection process from the active compounds of ChEMBL (green) to purchasable active compounds found in the Mcule database (orange) to the set of most selective compounds (blue) to the compounds that were tested in vitro (red).

Fig. 3

Open in a new tab

Out of the 564 identified compounds, 87 compounds – from the top contributing 12 suppliers that were available in 1 mg quantities for ≤150 USD – were acquired for in vitro testing. The purchased compounds were subsequently submitted to the NCI-6024 (National Cancer Institute - 60) panel for testing. The NCI-60 panel is a standardised non-profit platform to evaluate potential anticancer agents in a preclinical context. This panel comprises 60 human cancer cell lines derived from nine different tissues, representing a wide variety of cancer types, including leukaemia, lung, colon, central nervous system, melanoma, ovarian, renal, prostate, and breast cancers. The compounds were tested at a concentration of 10 μM that is a frequently applied screening concentration in high-throughput screens (HTSs) and the default concentration for the NCI60 panel. Results were expressed as cell count difference ratios between the drug administration and standard (no-drug) cases, where −100%, 0% and +100% indicate complete cell death, complete inhibition of cell growth, and unchanged cell growth, respectively. Out of the 87 tested compounds, 38 had relevant targets for this screen, derived from an organism belonging to the class of Mammalia, in particular, Homo sapiens or Mus musculus. Out of these 38 compounds, 10 (26%) exhibited more than 80% growth inhibition (cell count ratio <20%) on at least one cell line. Most compounds modulated only a few cell lines; however, one compound demonstrated broad-spectrum activity that can be considered as cytotoxicity. Out of the 10 compounds, 7 have a calculated total score (derived from known bioactivities) higher than 4 that we considered significantly selective and thus worth analysing further. These seven hits included the one potentially cytotoxic compound too. The remaining 6 compounds have shown lethality on one or a few cell lines only. Their structures are represented in Fig. 4, while their screening results are summarized in Fig. 5. Although these results are preliminary, there is a strong indication that the inhibitory effects of these compounds on the following six targets are linked to their observed anti-cancer activities in the NCI-60 screens. In other words, the target behind these identified selective ligands may represent valid targets for cancer therapy.

Fig. 4. The structures of screening hits exhibiting significant growth inhibition.

Fig. 4

Open in a new tab

Fig. 5. Heatmap of the cell growth of the compounds that had a total score over 4 and showed over 80% growth inhibition on at least one cell line. Colour scale changes from no growth inhibition (green) to 100% lethality (red).

Fig. 5

Open in a new tab

We therefore analyzed the available literature data of the targets of these ligands in more detail to see whether their relevance in cancer therapy has been already suggested. We summarize below the studies that link the modulation of these targets with tumour viability.

Nuclear receptor retinoic acid receptor-related orphan receptor-gamma (RORa): although decreased levels of RORa has already been shown in several tumour types,24–28 for the gamma variant the data are inconclusive. In some cancer types it showed decreased levels with positive correlation on survival times, while other studies have reported increased levels.29–31 In our study, it showed 80% cell growth inhibition in one case only (HCT-116 cell line), while average growth inhibition was 9%. Compound 1 (CHEMBL1433015) has a potency of 316 nM on nuclear receptor ROR-gamma (CHEMBL1293231) and was reported inactive 9 times on other targets.

Heat shock factor protein 1 (HSF1) is a protein that is frequently hyperactivated in cancer cells leading to the overexpression of genes associated with anti-apoptopic mechanisms and the spreading of such cells. Blocking this enzyme has improved the efficacy of other treatments like chemotherapy, radiation therapy and targeted treatments.32 Compound 2 (CHEMBL3193922) has an EC50 of 436 nM, and AC50 values of 848, 182 and 515 nM on HSF1 (target ID: CHEMBL5313). This compound showed growth inhibition on numerous cell lines of various types and caused cell death on two cell lines, BT-549 - breast cancer (13%) and NCI-H522 - lung cancer (54%). Except for these two cell lines the average growth inhibition was only 18% compared to non-drug control showing some notable selectivity.

Compound 3, with broad spectrum lethality and potential cytotoxicity, was originally measured to be active on paired box protein Pax-8, which is a transcription factor. Its exact role and mechanism on cancer viability is still unclear and requires further investigations. A few limited experiments showed that when its expression levels were restored to their original (lower) levels, it could inhibit the migration of gastric cancer cells. Furthermore, its expression level can be used as a marker to differentiate between various types of cancer (overexpressed in renal, duct and pancreas and ovarian cancers) supported by numerous publications.33 Compound 3 (CHEMBL1366838) has an AC50 of 650 nM on Pax-8 (target ID: CHEMBL2362980) and has inactivity data on 4 other targets, measured active too on protein RecA. This compound showed a broad spectrum of lethality, suggesting general cytotoxicity (average of 52%) probably due to its reactive, nucleophile chloroacetyl group. This moiety has been indicated as a threat for kidney, blood and lung tissues by Percepta software Health Effect predictions.34 While in our target deconvolution library design we applied the PAINS filter to reduce the number of false positive ligands, this result underlines the importance of applying an even broader set of filters. Nevertheless, the vast majority of the profiled ligands in this study showed selective distinct inhibition of only a few cell lines.

Protein-tyrosine phosphatase LC-PTP is a leukocyte PTP localized subcellularly. The PTP family plays a crucial role in eukaryotic signalling pathways by being responsible for phosphorylation while their counterparts, protein tyrosine kinases (PTKs) are responsible for dephosphorylation.35 In general, less information is available about PTPs than PTKs and their oncogenic potential. Studies investigating their role in cancer cells had different results ranging from inhibition of tumor growth36 to increased proliferation.37 The only known relation for LC-PTP is that the gene coding of this phosphatase falls in a location where candidate tumor suppressors might be present.38 Compound 4 (CHEMBL1536896) has an IC50 of 209 nM on LC-PTP (target ID: CHEMBL2219) and has inactivity data on 4 other targets. On average it had 56% growth inhibition while on 3 cell lines it exhibited 80% growth inhibition with 10% lethality on 1 cell line (BT-549 breast cancer cells).

Compound 5 targeting the microphthalmia-associated transcription factor showed over 80% growth reduction in one cancer type (NCI-H522, lung cancer). However this target has only been associated with certain types of melanoma.39,40 It has been shown that certain mutations of this target might lead to renal cell carcinoma41 Compound 5 (CHEMBL1601979) has a 601 nM AC50 value on the microphthalmia-associated transcription factor (CHEMBL1741165) and was confirmed in 3 parallel measurements, while proven to be inactive on 4 other targets.

Lysine-specific demethylase 4C represented by compound 6 has been associated with certain cancer types, however its exact role is yet to be discovered. In our assay, it significantly reduced the viability of NCI-H522 cell lines by 83%, while the average cell growth rate was 84% demonstrating a cell line selective effect. The compound has an AC50 of 800 nM on its target, and was reported inactive on 2 other targets.

Compound 7, targeting scavenger receptor class B member 1, showed over 80% on 3 cell lines and over 90% growth inhibition on one cell line while it had 34% growth inhibition on average. Its membrane protein target is known for being a relatively consistent marker in cancer tissues.42–44 Furthermore, this protein is commonly known to be overexpressed in breast and prostate cancer cell lines. In our assay, compound 7 showed 90 and 92% growth inhibitions in two melanoma cell lines. This compound has a 182 nM AC50 value on its target, while it was measured to be inactive on 3 other targets already.

One compound with broad inhibition and lethality profiles (potentially cytotoxic) was selected by the NCI for further testing in five doses to assess its GI50 (growth inhibition 50%), half-maximal inhibitory concentration (IC50), TGI (total growth inhibition), and LC50 (lethal concentration 50%) values. Compound 3 had a TGI value below 10 μM on 54 cell lines and below 1 μM on 10 cell lines, while compound 9 showed TGI below 10 μM on 54 and 1 μM on 2 cell lines.

The average values of GI50, IC50, LC50 and TGI values grouped by the tumour cell types are summarized in Fig. 6. Compound 3 had the best average GI50 and LC50 values on leukaemia cells with concentrations of 455 nM and 736 nM. The highest TGI and LC50 values were on renal cancer cells with 1.93 μM and 4.26 μM respectively.

Fig. 6. Growth inhibition (GI50, blue), half-maximal inhibitory concentration (IC50, red), lethal concentration (LC50, yellow) and total growth inhibition (TGI, green) values of compound 3 on various cancer cell types. Concentration values are only visualized between 4 (−log[100 μM]) and 7 (−log[100 nM]).

Fig. 6

Open in a new tab

Conclusion

In summary, we have developed an algorithm that collects highly selective compounds from the ChEMBL database and pairs them with the commercially available compounds from the Mcule database. By using our selection algorithm, we could assign selective ligands to over 500 targets. Out of the developed high-selectivity target deconvolution set, 87 compounds were purchased and tested on the NCI60 panel, leading to the identification of 7 compounds possessing relevant targets that showed at least 80% growth inhibition on cancer cell lines at a concentration of 10 μM. One compound showed a broad spectrum of cell death and was selected for further characterization by NCI: 6 compounds showed inhibition of particular cancer cell lines suggesting a selective mechanism of action. Most of the known targets of these 6 ligands have been already demonstrated to be associated with cancer, hence validating our phenotypic specificity filtering approach that favours compounds with the highest target selectivity. Furthermore, our target-selective tool compounds may have identified potentially novel cancer targets for drug discovery, which could be exploited by target-based approaches after more thorough target characterization and validation.

Our results indicate the significant potential of mining bioactivity datasets, such as ChEMBL, and utilising compounds alongside their existing bioactivity data for application in novel therapeutic areas, including anti-cancer therapy.

Author contributions

G. T. contributed with the data mining and preparation, implementation of the selection procedure alongside collecting the literature data and preparing the figures and tables. R. K. contributed with the idea of selection procedure and the preparation of compounds for NCI60 screening alongside comprehensive review of the paper. Gy. T. B. contributed with the review of the paper and the interpretation of screening results.

Conflicts of interest

There are no conflicts to declare.

Acknowledgments

We would like say thank you to Benjamin Kovats for creating Fig. 1. Also, we would like to thank the National Cancer Institute for screening our set of compounds on the NCI60 panel.

Data availability

Data for this article, including compound sets and screening results are available at Zenodo at [https://doi.org/10.5281/zenodo.14567397].

Notes and references

  1. Vincent F. Nueda A. Lee J. Schenone M. Prunotto M. Mercola M. Nat. Rev. Drug Discovery. 2022;21:899–914. doi: 10.1038/s41573-022-00472-w. [DOI] [PMC free article] [PubMed] [Google Scholar]
  2. Lindsay M. A. Nat. Rev. Drug Discovery. 2003;2:831–838. doi: 10.1038/nrd1202. [DOI] [PubMed] [Google Scholar]
  3. Overington J. P. Al-Lazikani B. Hopkins A. L. Nat. Rev. Drug Discovery. 2006;5:993–996. doi: 10.1038/nrd2199. [DOI] [PubMed] [Google Scholar]
  4. Swinney D. Anthony J. Nat. Rev. Drug Discovery. 2011;10:507–519. doi: 10.1038/nrd3480. [DOI] [PubMed] [Google Scholar]
  5. Arrowsmith J. Miller P. Nat. Rev. Drug Discovery. 2013;12:569–570. doi: 10.1038/nrd4090. [DOI] [PubMed] [Google Scholar]
  6. Harrison R. K. Nat. Rev. Drug Discovery. 2016;15:817–818. doi: 10.1038/nrd.2016.184. [DOI] [PubMed] [Google Scholar]
  7. Sadri A. J. Med. Chem. 2023;66:12651–12677. doi: 10.1021/acs.jmedchem.2c01737. [DOI] [PubMed] [Google Scholar]
  8. Lee J. Bogyo M. Curr. Opin. Chem. Biol. 2013;17:118–126. doi: 10.1016/j.cbpa.2012.12.022. [DOI] [PMC free article] [PubMed] [Google Scholar]
  9. Chang Y. Schlebach J. P. VerHeul R. A. Park C. Protein Sci. 2012;21:1280–1287. doi: 10.1002/pro.2112. [DOI] [PMC free article] [PubMed] [Google Scholar]
  10. West G. M. Tang L. Fitzgerald M. C. Anal. Chem. 2008;80:4175–4185. doi: 10.1021/ac702610a. [DOI] [PubMed] [Google Scholar]
  11. Young D. W. Bender A. Hoyt J. McWhinnie E. Chirn G.-W. Tao C. Y. Tallarico J. A. Labow M. Jenkins J. L. Mitchison T. J. Feng Y. Nat. Chem. Biol. 2008;4:59–68. doi: 10.1038/nchembio.2007.53. [DOI] [PubMed] [Google Scholar]
  12. Homeyer N. van Deursen R. Ochoa-Montaño B. Heikamp K. Ray P. Zuccotto F. Blundell T. L. Gilbert I. H. J. Mol. Graphics Modell. 2020;95:107485. doi: 10.1016/j.jmgm.2019.107485. [DOI] [PMC free article] [PubMed] [Google Scholar]
  13. Lounkine E. Keiser M. J. Whitebread S. Mikhailov D. Hamon J. Jenkins J. L. Lavan P. Weber E. Doak A. K. Côté S. Shoichet B. K. Urban L. Nature. 2012;486:361–367. doi: 10.1038/nature11159. [DOI] [PMC free article] [PubMed] [Google Scholar]
  14. Gaulton A. Bellis L. J. Bento A. P. Chambers J. Davies M. Hersey A. Light Y. McGlinchey S. Michalovich D. Al-Lazikani B. Overington J. P. Nucleic Acids Res. 2012;40:D1100–D1107. doi: 10.1093/nar/gkr777. [DOI] [PMC free article] [PubMed] [Google Scholar]
  15. Zdrazil B. Felix E. Hunter F. Manners E. J. Blackshaw J. Corbett S. de Veij M. Ioannidis H. Lopez D. M. Mosquera J. F. Magarinos M. P. Bosc N. Arcila R. Kizilören T. Gaulton A. Bento A. P. Adasme M. F. Monecke P. Landrum G. A. Leach A. R. Nucleic Acids Res. 2024;52:D1180–D1192. doi: 10.1093/nar/gkad1004. [DOI] [PMC free article] [PubMed] [Google Scholar]
  16. Wishart D. S. Knox C. Guo A. C. Shrivastava S. Hassanali M. Stothard P. Chang Z. Woolsey J. Nucleic Acids Res. 2006;34:D668–D672. doi: 10.1093/nar/gkj067. [DOI] [PMC free article] [PubMed] [Google Scholar]
  17. Knox C. Wilson M. Klinger C. M. Franklin M. Oler E. Wilson A. Pon A. Cox J. Chin N. E. L. Strawbridge S. A. Garcia-Patino M. Kruger R. Sivakumaran A. Sanford S. Doshi R. Khetarpal N. Fatokun O. Doucet D. Zubkowski A. Rayat D. Y. Jackson H. Harford K. Anjum A. Zakir M. Wang F. Tian S. Lee B. Liigand J. Peters H. Wang R. Q. R. Nguyen T. So D. Sharp M. da Silva R. Gabriel C. Scantlebury J. Jasinski M. Ackerman D. Jewison T. Sajed T. Gautam V. Wishart D. S. Nucleic Acids Res. 2024;52:D1265–D1275. doi: 10.1093/nar/gkad976. [DOI] [PMC free article] [PubMed] [Google Scholar]
  18. Mcule database, https://mcule.com/database/, (accessed 28 July 2024)
  19. Heller S. McNaught A. Stein S. Tchekhovskoi D. Pletnev I. Aust. J. Chem. 2013;5:7. doi: 10.1186/1758-2946-5-7. [DOI] [PMC free article] [PubMed] [Google Scholar]
  20. Baell J. B. Holloway G. A. J. Med. Chem. 2010;53:2719–2740. doi: 10.1021/jm901137j. [DOI] [PubMed] [Google Scholar]
  21. Landrum G., Curating the PAINS filters, https://rdkit.blogspot.com/2015/08/curating-pains-filters.html, (accessed 28 July 2024)
  22. Bruns R. F. Watson I. A. J. Med. Chem. 2012;55:9763–9772. doi: 10.1021/jm301008n. [DOI] [PubMed] [Google Scholar]
  23. Van der Maaten L. Hinton G. J. Mach. Learn. Res. 2008;9:2579–2605. [Google Scholar]
  24. NCI-60 human tumor cell lines screen, https://dtp.cancer.gov/discovery_development/nci-60/methodology.htm, (accessed 28 July 2024)
  25. Vassar R. Kuhn P.-H. Haass C. Kennedy M. E. Rajendran L. Wong P. C. Lichtenthaler S. F. J. Neurochem. 2014;130:4–28. doi: 10.1111/jnc.12715. [DOI] [PMC free article] [PubMed] [Google Scholar]
  26. Farris F. Matafora V. Bachi A. J. Exp. Clin. Cancer Res. 2021;40:147. doi: 10.1186/s13046-021-01953-3. [DOI] [PMC free article] [PubMed] [Google Scholar]
  27. Brożyna A. A. Jóźwicki W. Skobowiat C. Jetten A. Slominski A. T. Onco Targets Ther. 2016;7:63261–63282. doi: 10.18632/oncotarget.11211. [DOI] [PMC free article] [PubMed] [Google Scholar]
  28. Xiong G. Wang C. Evers B. M. Zhou B. P. Xu R. Cancer Res. 2012;72:1728–1739. doi: 10.1158/0008-5472.CAN-11-2762. [DOI] [PMC free article] [PubMed] [Google Scholar]
  29. Thibaudin A. Limagne M. Hampe C. Ballot J. Truntzer M. Ghiringhelli F. Boidot L. Onco Targets Ther. 2016;5:e1055444. [Google Scholar]
  30. Irshad S. Flores-Borja F. Lawler K. Monypenny J. Evans R. Male V. Gordon P. Cheung A. Gazinska P. Noor F. Wong F. Grigoriadis A. Fruhwirth G. O. Barber P. R. Woodman N. Patel D. Rodriguez-Justo M. Owen J. Martin S. G. Pinder S. E. Gillett C. E. Poland S. P. Ameer-Beg S. McCaughan F. Carlin L. M. Hasan U. Withers D. R. Lane P. Vojnovic B. Quezada S. A. Ellis P. Tutt A. N. J. Ng T. Cancer Res. 2017;77:1083–1096. doi: 10.1158/0008-5472.CAN-16-0598. [DOI] [PubMed] [Google Scholar]
  31. Huang Q. Fan J. Qian X. Lv Z. Zhang X. Han J. Wu F. Chen C. Du J. Guo M. Hu G. Jin Y. J. Cancer Res. Clin. Oncol. 2016;142:263–272. doi: 10.1007/s00432-015-2040-0. [DOI] [PMC free article] [PubMed] [Google Scholar]
  32. Gumilar K. E. Chin Y. Ibrahim I. H. Tjokroprawiro B. A. Yang J.-Y. Zhou M. Gassman N. R. Tan M. Cancers. 2023;15:5167. doi: 10.3390/cancers15215167. [DOI] [PMC free article] [PubMed] [Google Scholar]
  33. Guo Y.-J. Pan W.-W. Liu S.-B. Shen Z.-F. Xu Y. Hu L.-L. Exp. Ther. Med. 2020;19:1997–2007. doi: 10.3892/etm.2020.8454. [DOI] [PMC free article] [PubMed] [Google Scholar]
  34. Xiang L. Kong B. Oncol. Lett. 2013;5:735–738. doi: 10.3892/ol.2013.1121. [DOI] [PMC free article] [PubMed] [Google Scholar]
  35. version 2024.1
  36. Hooft van Huijsduijnen R. Gene. 1998;225:1–8. doi: 10.1016/S0378-1119(98)00513-7. [DOI] [PubMed] [Google Scholar]
  37. Ardini E. Agresti R. Tagliabue E. Greco M. Aiello P. Yang L. T. Ménard S. Sap J. Oncogene. 2000;19:4979–4987. doi: 10.1038/sj.onc.1203869. [DOI] [PubMed] [Google Scholar]
  38. Radha V. Nambirajan S. Swarup G. FEBS Lett. 1997;409:33–36. doi: 10.1016/S0014-5793(97)00471-7. [DOI] [PubMed] [Google Scholar]
  39. Dynek J. N. Chan S. M. Liu J. Zha J. Fairbrother W. J. Vucic D. Cancer Res. 2008;68:3124–3132. doi: 10.1158/0008-5472.CAN-07-6622. [DOI] [PubMed] [Google Scholar]
  40. Ballotti R. Cheli Y. Bertolotto C. Mol. Cancer. 2020;19:170. doi: 10.1186/s12943-020-01290-7. [DOI] [PMC free article] [PubMed] [Google Scholar]
  41. Bertolotto C. Lesueur F. Giuliano S. Strub T. de Lichy M. Bille K. Dessen P. d'Hayer B. Mohamdi H. Remenieras A. Maubec E. de la Fouchardière A. Molinié V. Vabres P. Dalle S. Poulalhon N. Martin-Denavit T. Thomas L. Andry-Benzaquen P. Dupin N. Boitier F. Rossi A. Perrot J.-L. Labeille B. Robert C. Escudier B. Caron O. Brugières L. Saule S. Gardie B. Gad S. Richard S. Couturier J. Teh B. T. Ghiorzo P. Pastorino L. Puig S. Badenas C. Olsson H. Ingvar C. Rouleau E. Lidereau R. Bahadoran P. Vielh P. Corda E. Blanché H. Zelenika D. Galan P. French Familial Melanoma Study Group Aubin F. Bachollet B. Becuwe C. Berthet P. Bignon Y. J. Bonadona V. Bonafe J.-L. Bonnet-Dupeyron M.-N. Cambazard F. Chevrant-Breton J. Coupier I. Dalac S. Demange L. d'Incan M. Dugast C. Faivre L. Vincent-Fétita L. Gauthier-Villars M. Gilbert B. Grange F. Grob J.-J. Humbert P. Janin N. Joly P. Kerob D. Lasset C. Leroux D. Levang J. Limacher J.-M. Livideanu C. Longy M. Lortholary A. Stoppa-Lyonnet D. Mansard S. Mansuy L. Marrou K. Matéus C. Maugard C. Meyer N. Nogues C. Souteyrand P. Venat-Bouvet L. Zattara H. Chaudru V. Lenoir G. M. Lathrop M. Davidson I. Avril M.-F. Demenais F. Ballotti R. Bressac-de Paillerets B. Nature. 2011;480:94–98. doi: 10.1038/nature10539. [DOI] [PubMed] [Google Scholar]
  42. Boucher J. K. D. Chen C. Z. Franco S. J. Nichols N. A. K. David J. S. R. Triggle D. A. W. Front. Pharmacol. 2016;7:466. [Google Scholar]
  43. Mooberry L. K. Sabnis N. A. Panchoo M. Nagarajan B. Lacko A. G. Front. Pharmacol. 2016;7:466. doi: 10.3389/fphar.2016.00466. [DOI] [PMC free article] [PubMed] [Google Scholar]
  44. Rajora M. A. Zheng G. Front. Pharmacol. 2016;7:326. doi: 10.3389/fphar.2016.00326. [DOI] [PMC free article] [PubMed] [Google Scholar]

Associated Data

This section collects any data citations, data availability statements, or supplementary materials included in this article.

Data Availability Statement

Data for this article, including compound sets and screening results are available at Zenodo at [https://doi.org/10.5281/zenodo.14567397].

Articles from RSC Medicinal Chemistry are provided here courtesy of Royal Society of Chemistry

RESOURCES