Computational studies on KRAS inhibitors using web based tools

COMPUTATIONAL STUDIES ON KRAS INHIBITORS USING WEB BASED TOOLS Trabajo de Fin de Máster Aitor Díaz Oliva Tutorizado por Dr. José M. Padrón Máster en Biomedicina Julio 2024 2 La presente memoria de investigación ha sido realizada por Aitor Díaz Oliva durante el curso académico 2023-2024 en el grupo BioLab del Instituto Universitario de BioOrgánica “Antonio González” (IUBO-AG) y bajo la dirección del Dr. José M. Padrón. El trabajo forma parte de la línea de investigación Diseño, descubrimiento y evaluación de fármacos anticancerígenos. La Laguna, en la fecha de firma al pie. 3 TABLE OF CONTENTS ABSTRACT 4 RESUMEN 4 INTRODUCTION 5 HYPOTHESIS 11 OBJECTIVES 11 MATERIALS AND METHODS 12 RESULTS & DISCUSSION 17 CONCLUSIONS 33 REFERENCES 34 4 ABSTRACT In this work, a computational study of the KRAS protein with two of its inhibitors for the G12C variant, sotorasib and adagrasib, has been carried out. Initially, the interaction between the KRAS G12C inhibitors and the protein was studied using publicly available software (Swissdock). Next the interactions were studied for the G12D, G12R and G13C variants of KRAS. In a effort to compare the results obtained with those given by professional software, these same interactions were studied using licensed software MOE. All the results obtained from the different software and the results obtained from the different KRAS variants were compared. RESUMEN En este trabajo se ha realizado un estudio computacional de la proteína KRAS con dos de sus inhibidores para la variante G12C, sotorasib y adagrasib. Inicialmente, se estudió la interacción entre los inhibidores de KRAS G12C y la proteína utilizando un software disponible públicamente (Swissdock). A continuación, se estudiaron las interacciones para las variantes G12D, G12R y G13C de KRAS. En un esfuerzo por comparar los resultados obtenidos con los proporcionados por el software profesional, estas mismas interacciones se estudiaron utilizando el software con licencia MOE. Se compararon todos los resultados obtenidos con los distintos programas informáticos y los resultados obtenidos con las distintas variantes de KRAS. INTRODUCTION 5 INTRODUCTION Nowadays, the number of cancer cases worldwide has increased significantly due to the increase in life expectancy and current environmental conditions. According to a study by the INE, in Spain in 2022, cancer mortality became the second cause of death just behind cardiovascular diseases [1]. Among the various types of cancer that exist, it is important to take into account those that cause the greatest mortality in the population. For women they are lung, breast, colorectal and pancreatic cancer, while for men they are lung, prostate, colorectal, pancreatic and liver cancer [2]. Of the deadliest cancers mentioned above, we will focus on pancreatic cancer, where the main risks are advanced age, race, gender and environmental factors such as smoking, diet, etc. In relation to molecular factors, the most frequent tumor suppressor genes that are inactivated are P16 in 95%, P53 in up to 75%, DPC4 in 50% and BRCA2. The most frequent genetic alteration of oncogenes is the KRAS gene, found in 80% to 100% of cases [3-4]. The KRAS protein (figure 1) is an important oncogene that plays a key role in the development and progression of various cancers. The KRAS gene codes for a GTPase protein that is part of the intracellular RAS/MAPK signaling pathway, which regulates cellular processes such as proliferation, differentiation and apoptosis. The most common mutations are found in codons 12, 13, 61 and 146 of the KRAS gene. Mutated KRAS is the main oncogenic driver of pancreatic ductal adenocarcinoma (PDAC) and an attractive treatment target. To date, no effective treatments have been identified for patients with KRAS-mutated PDAC, except for the unusual KRAS G12C mutations [4-8]. INTRODUCTION 6 Figure 1. Structure of wild type human GTP KRAS protein. In cancers caused by mutant KRAS, the protein is blocked in the active GTP-bound state constitutively, through a defect in the switch-off mechanism. As such, the mutant protein resembles the normal KRAS protein from a structural perspective, making therapeutic targeting extremely challenging. KRAS is a member of a large family of related proteins, which share very similar GDP/GTP binding domains, making targeted therapies difficult. The KRAS G12C mutation (Figure 2) is characterized by an amino acid substitution of cysteine in the GTP-binding site of the KRAS protein, leading to constant activation of the RAS/MAPK signaling pathway and uncontrolled cell proliferation and survival [9]. Figure 2. Structure of mutant KRAS G12C. The inhibitors that have shown to be effective against the KRAS G12C mutant are sotorasib (also known as AMG 510 or Lumakras) and adagrasib (also known as MRTX849). These direct inhibitors act by selectively forming a covalent bond with Cys 12 residue within the switch II pocket of the KRAS G12C protein, thus blocking KRAS in an inactive INTRODUCTION 7 state to stop cell proliferation (Figure 3). Preclinical studies showed that sotorasib and adagrasib selectively affect the viability of KRAS G12C mutant cell lines, but do not affect cell lines with other KRAS mutations in vitro and in vivo [10]. Figure 3. a) KRAS has GTPase activity and can convert GTP into GDP by hydrolyzing the gamma phosphate on GTP. The inactive and active states of KRAS are regulated by GAP and GEF, respectively. RTKs are the second major type of cell surface receptors with a wide range of functions, including promoting KRAS activation and subsequent multiple effector pathways, especially the RAF-MEK-ERK and PI3K-AKT-mTOR pathways. b) Compared with wild-type KRAS, which maintains a balance between inactive and activated states, Cys 12 (C12) mutations destroys GTPase activity of KRAS and locks in the GTP-bound state. In contrast, the small molecule drug sotorasib and adagrasib can form a covalent bond with C12 in the KRAS G12C protein, causing KRAS to be in an inactive state. Sotorasib Sotorasib (AMG510, Figure 4) is a pyridopyrimidine that has a pyrido[2,3-d]pyrimidin2(1H)-one substituted with 4-methyl-2-(propan-2-yl)pyridin-3-yl, (2S)-2-methyl -4-(prop-2- INTRODUCTION 8 enoyl)piperazin-1-yl, fluoro and 2-fluoro-6-hydroxyphenyl groups at positions 1, 4, 6 and 7, respectively. It is approved for the treatment of patients with lung cancer and other non-small cell solid cancers that have KRAS G12C mutations. It has a role as an antineoplastic agent. It is a member of acrylamides, an N-acylpiperazine, a pyridopyrimidine, a member of monofluorobenzenes, a member of methylpyridines, a tertiary carboxamide, a tertiary amino compound and a member of phenols. In terms of physicochemical properties, it has a molecular weight of 560.6 g/mol, a solubility of 0.03 mg/mL at pH 6.8 and a pKa of 4.56 and 8.56. Figure 4. Molecular structure of sotorasib. For absorption of a once-daily dose of 960 mg of sotorasib, it achieves a C max of 7.50 µg/mL, with a median T max of 2.0 hours and an AUC 0-24 h of 65.3 h\*μg/mL. Sotorasib is 74% excreted in the feces and 6% in the urine. 53% of the dose recovered in feces and 1% of the dose recovered in urine are in the unchanged parent compound form. Typically, GTP binds to KRAS, activating the protein and promoting effectors of the MAP kinase pathway. GTP is hydrolyzed to GDP and KRAS is inactivated (Figure 5). KRAS G12C mutations alter GTP hydrolysis, leaving it in an active form. Sotorasib binds to the Cys 12 residue in KRAS G12C mutan, keeping the protein in its inactive form. The cysteine residue targeted by sotorasib is not present in wild-type KRAS, preventing off-target effects. This mutation is present in 13% of non-small cell lung cancers, 3% of colorectal and appendiceal cancers and 1-3% of solid tumors [11]. INTRODUCTION 9 Figure 5. Binding pose of sotorasib to the KRAS-G12C protein. Orange: Switch II domain; Yellow: residual C12; Blue: a cryptic pocket composed of H95, Y96 and Q99. Sotorasib (green) and GDP (grey) are shown in sticks. Adagrasib Adagrasib (Figure 6) is another drug that acts as a KRAS G12C protein inhibitor similar to sotorasib. It has a molecular weight of 604.1 g/mol. In terms of absorption, the AUC and C max of adagrasib increase dose-proportionally between 400 mg and 600 mg. The Tmax is approximately 6 hours. Furthermore, adagrasib has a high oral bioavailability and is able to penetrate the central nervous system. It is eliminated in the feces and urine. In patients receiving a single dose of radiolabelled adagrasib, 75% of the dose was recovered in feces (14% unchanged), while 4.5% was recovered in urine (2% unchanged). Adagrasib is a small molecule inhibitor of the mutant KRAS G12C protein found in up to 13% of refractory non-small cell lung cancer cases. Elevations of serum aminotransferases are common during treatment with adagrasib and a proportion of patients develop clinically apparent liver injury that can be severe [12]. RESULTS & DISCUSSION 16 Figure 11. Screenshot from the new SwissDock website. RESULTS & DISCUSSION 17 RESULTS & DISCUSSION Today there is a wide variety of cancers and tumors that can manifest in any part of the body and act in a variety of ways from being harmless to being ferociously lethal. W drew our attention to pancreatic cancer. There are several genes that increase the risk of pancreatic cancer. These genes comprise: • TP53 • CDKN2A • SMAD4/DPC4 • BRCA1 and BRCA2. • KRAS We focused our attention on the latter, as it is a protein of interest for cell growth and division. A mutation in this protein leads to increased proliferation of cancer cells. Next, we retrieved information on the protein in the PDB database for both the wild-type and mutant protein. Within the PDB we can obtain information of interest on the protein such as name, year of publication, authors who synthesized it, method by which the characterization was obtained, the organism from which it was obtained and the type of macromolecule it belongs to (Figure 12). Figure 12. Information about KRAS WT obtained from PDB. In order to be able to carry out the study, the resolution of the protein must be taken into account, given that values greater than 2 Å would have a poor resolution or would be heavy RESULTS & DISCUSSION 18 enough to carry out the study. On the other hand, it was necessary to make sure that the protein belongs to the Homo sapiens organism. For the old version of SwissDock, it was necessary to use the PDB in order to download the desired protein in PDB format. Subsequently, the SMILE formats of the two KRAS G12C inhibitors sotorasib and adagrasib were obtained from Pubchem. Once the formats have been obtained, OpenBabel software was used to transform the SMILE format into another that can be read by the free SwissDock software. This format is mol2. Subsequently, the PDB format for the protein and mol2 for the inhibitor was introduced into SwissDock. As for the docking time, since it is a free software, it can take from minutes to days. To avoid having to wait, SwissDock gives you the option of entering your email address and you will be notified when the docking is finished (Figure 13). Figure 13. Screenshot from the SwissDock website for docking. RESULTS & DISCUSSION 19 Unluckily due to the software update, it is currently impossible to access the information of the docking performed before the update, giving an error and the impossibility to access the results (Figure 14). Figure 14. Image of the error when accessing the results after docking. As for the new software update, it has been designed in such a way that it is more userfriendly. It is no longer necessary to download the PDB files from the protein, nor is it necessary to transform the SMILE formats to mol2. As for the protein-target interaction, it is visualized in such a way that ionic bonds, hydrogen bonds, π-bonds, hydrophobic contacts can be observed and shows the protein surface (Figure 15). Figure 15. Example of different types of protein-inhibitor interactions (left). Example of interaction between protein surface and inhibitor (right). SwissDock is a free software and has some limitations such as the size of the protein. In addition, with the new version they avoid docking for several days. Thus, if it is expected that the result is prolonged in time, they warn that it is not possible to perform the calculations. It can also lead to errors when entering the desired protein (Figure 16). RESULTS & DISCUSSION 20 Figure 16. Error of the software to recognise the protein. Sotorasib As aforementioned, sotorasib is a drug that inhibits the KRAS G12C mutated protein. We run a computational study between KRAS proteins and sotorasib. KRAS wt It was decided to perform docking for the KRAS wt protein together with the inhibitor in order to have information about how the software will act against a hypothetical proteininhibition interaction. Although the resolution of the PDB structure 6VC8 is larger than 2Å (Figure 17), the computational study was carried out due to the absence of better options. RESULTS & DISCUSSION 21 Figure 17. Information about KRAS wt. Figure 18. Interaction between KRAS and sotorasib. RESULTS & DISCUSSION 22 Figure 19. Chemical interaction between KRAS and sotorasib. As shown in Figures 18 and 19, sotorasib binds on the outside of the protein. In addition, there are hydrogen bonding interactions of the hydroxyl group of sotorasib with Lys 117 (red protein) and HSD 95 (white protein). In a comparison with MOE software (Figure 20), Lys 117 binds to the pyridine nitrogen. Figure 20. Chemical interaction between KRAS and sotorasib using MOE. RESULTS & DISCUSSION 23 As expected, the docking simulation of sotorasib with KRAS wt shows that the drug does not fit into the binding site. KRAS G12C The interaction between the KRAS G12C mutant (Figure 21, PDB ID: 5V9O) and sotorasib was modelled in order to compare the data obtained on the interaction with those reported in the literature. Figure 21. Information about KRAS G12C mutant. Figure 22. Snapshots of the interaction between KRAS G12C mutant and sotorasib. As can be seen in the Figures 22 & 23, the inhibitor is introduced into the protein (binding site), producing the KRAS G12C-sotorasib interaction. There is also interaction between the allylic ketone and Gly 15. RESULTS & DISCUSSION 24 Figure 23. Predicted interaction between KRAS G12C mutant and sotorasib. Figure 24 shows the interactions of the inhibitor with the protein using MOE software. It can be seen that the predicted interactions are different between the Swissdock software and MOE software. Figure 24. Predicted interaction between KRAS G12C mutant and sotorasib using MOE RESULTS & DISCUSSION 25 KRAS G12R The KRAS G12R mutant (Figure 25, PDB ID: 6CU6) has a similar structure to G12C. Among the characteristics of 6CU6 we highlight that it has an acceptable resolution and belongs to the Homo sapiens family. Figure 25. Information about KRAS G12R mutant. Figure 26. Predicted interaction between KRAS G12R mutant and sotorasib. RESULTS & DISCUSSION 32 Similarly, the web platform did not accept the mol2 file (Figure 38). Therefore, we could not run the proposed study. Figure 38. Screenshot of error message when entering the mol2 file of adagrasib. CONCLUSIONS 33 CONCLUSIONS After docking the proposed inhibitors against the KRAS protein, the following conclusions were obtained: It is observed that there are slight differences in the prediction of KRAS-sotorasib interaction depending on the software used: Swissdock (public use software) and MOE (commercial use software). Both systems predict interactions on different heteroatoms and different positions of the amino acids of the protein. With Swissdock software, it was observed that there is no interaction that is promising for KRAS G12R and G12D. However, similarity is seen between G12C and G13C. Therefore interaction studies between KRAS G13C-sotorasib and inhibitor should be carried out for this mutant. For adagrasib, it was not possible to perform docking studies with Swissdock software due to problems in entering the structure in either SMILES and mol2 format of the inhibitor. Finally, the use of public software such as Swissdock allows any student to perform studies on the interactions between proteins and desired target molecules. However, this software has several drawbacks, such as the waiting time to obtain the results, which can last several days, and errors in the software itself (data entry, data sizes, lost connections, etc.). Although it is ideal to avoid high usage of private licenses, data should be corroborated with professional software that is highly reliable. REFERENCES 34 REFERENCES [1] INE. Causas de muerte dos años después del comienzo de la pandemia. Retrieved from: https://www.ine.es/ss/Satellite?L=es_ES&c=INECifrasINE_C&cid=1259954008107& p=1254735116567&pagename=ProductosYServicios%2FINECifrasINE_C%2FPYSD etalleCifrasINE [2] Bray, F., Laversanne, M., Sung, H., Ferlay, J., Siegel, R. L., Soerjomataram, I., & Jemal, A. (2024). Global cancer statistics 2022: GLOBOCAN estimates of incidence and mortality worldwide for 36 cancers in 185 countries. CA: a cancer journal for clinicians, 74(3), 229-263. [3] Pereira, A. G. M., Noal, L. B., de Camargo, L. E. A., Iatskiu, C. E. A., Vieira, J. B., Figueiredo, E. F., & Colacite, J. (2022). Avaliação epidemiológica do Câncer de pâncreas no Brasil: mortalidade e fatores de risco: Epidemiological evaluation of pancreatic Cancer in Brazil: mortality and risk factors. Brazilian Journal of Development, 8(10), 68595-68603. [4] Ureña, J. F. (2017). Cáncer de páncreas. Revista Médica Sinergia, 2(8), 3-6. [5] McCormick, F. (2016). KRAS protein as a drug target. Journal of molecular medicine, 94, 253-258. [6] Luo, J. (2021, February). KRAS mutation in pancreatic cancer. In Seminars in oncology (Vol. 48, No. 1, pp. 10-18). WB Saunders. [7] Olmedo Pimentel, Y. (2015). Mutaciones en el Gen KRAS en muestra de tejidos tumorales (Doctoral dissertation, Universidad de Panamá. Vicerrectoría de Investigación y Postgrado.). [8] Strumberg, D., Schultheis, B., Ebert, M.P., Siveke, J.T., Kerkhoff, A., Hofheinz, R., Behringer, D., Schmidt, W.E., Göker, E., Dosso, S.D., Kneba, M., Yalcin, S., Overkamp, F., Schlegel, F., Dommach, M., Rohrberg, R., Steinmetz, T., Bach, F., Reuter, D., Ilyas, R.A., & Escobar, N.E. (2016). Traitement de patients diagnostiqués avec un adénocarcinome canalaire du pancréas au moyen de l'utilisation d'anticorps monoclonaux contre le récepteur de facteur de croissance épidermique (egfr). [9] Liu, J., Kang, R., & Tang, D. (2022). The KRAS-G12C inhibitor: activity and resistance. Cancer Gene Therapy, 29(7), 875-878. [10] Seton-Rogers, S. (2020). KRAS-G12C in the crosshairs. Nature Reviews Cancer, 20(1), 3-3. REFERENCES 35 [11] Yang, A., Li, M., & Fang, M. (2021). The research progress of direct KRAS G12C mutation inhibitors. Pathology & Oncology Research, 27, 631095. [12] Wu, F., Zhou, Y., Li, L., Shen, X., Chen, G., Wang, X., ... y Huang, Z. (2020). Enfoques computacionales en estudios preclínicos sobre descubrimiento y desarrollo de fármacos. Fronteras en química , 8 , 726. [13] Jiménez, C. S. H., & Albo, M. V. (2021). Pensamiento computacional como una habilidad genérica: una revisión sistemática. Ciencia Latina Revista Científica Multidisciplinar, 5(1), 1055-1078. [14] Cali, D. I. Q., & Peñafiel, J. E. R. (2022). Estudio computacional del comportamiento de dos cuerpos con cargas iguales. AlfaPublicaciones, 4(3.2), 6-20. [15] Burley, S. K., Berman, H. M., Kleywegt, G. J., Markley, J. L., Nakamura, H., & Velankar, S. (2017). Protein Data Bank (PDB): The Single Global Macromolecular Structure Archive. Methods in molecular biology (Clifton, N.J.), 1607, 627–641. [16] Berman, H., Henrick, K., Nakamura, H., & Markley, J. L. (2007). The worldwide Protein Data Bank (wwPDB): ensuring a single, uniform archive of PDB data. Nucleic acids research, 35(Database issue), D301–D303. [17] Kim S. (2016). Getting the most out of PubChem for virtual screening. Expert opinion on drug discovery, 11(9), 843–855. [18] Kim, S., Chen, J., Cheng, T., Gindulyte, A., He, J., He, S., Li, Q., Shoemaker, B. A., Thiessen, P. A., Yu, B., Zaslavsky, L., Zhang, J., & Bolton, E. E. (2021). PubChem in 2021: new data content and improved web interfaces. Nucleic acids research, 49(D1), D1388–D1395. [19] O'Boyle, N. M., Banck, M., James, C. A., Morley, C., Vandermeersch, T., & Hutchison, G. R. (2011). Open Babel: An open chemical toolbox. Journal of cheminformatics, 3, 1-14. [20] Bitencourt-Ferreira, G., & de Azevedo, W. F. (2019). Docking with swissdock. Docking screens for drug discovery, 189-202. [21] Patil, N. S., & Rohane, S. H. (2021). Organization of Swiss Dock: in study of computational and molecular docking study.