A New Machine-Learning-Based Vaccine Design Program Called Auto Pep Vax: Use in a Pan-Cancer Vaccine Targeting EGFR Missense Mutations
Keywords:
machine learning for peptide vaccine design, new vaccine design method, an cancer vaccine, GFR vaccine design, epitopes, MHC I and II, T-Cell receptorsAbstract
More complex in silico techniques are required to identify immunologically relevant epitopes since the existing epitope
selection methods for peptide vaccines only use one-peptope binding affinity estimates.In order to speed up and enhance the in silico
epitope selection process for peptide vaccine creation, we created AutoPepVax.A new tool called AutoPepVax uses a variety of
epitope properties to automatically identify non-toxic and non-allergic epitopes that may trigger lymphocytes that infiltrate tumours.
AutoPepVax uses machine learning models based on random forest classification and linear regression that are trained using tumour
sample datasets.AutoPepVax and its installation instructions are publicly accessible on GitHub.We utilised AutoPepVax to create a
pan-cancer peptide vaccine that targets missense mutations in the epidermal growth factor receptor (EGFR), which are often seen in
head and neck squamous cell carcinoma (HNSCC), lung adenocarcinoma (LUAD), colonrecta adenocarcinoma (CRAD), and
glioblastoma multiforme (GBM).Clinical studies for EGFR-specific peptide vaccines have already addressed these mutations in GBM
and LUAD; although they show promise, they have not yet showed clinical effectiveness.Our investigation of 96 EGFR mutations
using AutoPepVax revealed 430 possible MHC-I-restricted depitope–HLA couples from 168,669 candidates and 368 potential MHC-
I-restricted pairs from 49,113 candidates.Interestingly, 19 mutations showed MHC I and II restriction epitopes that were viable.We
used our algorithm, PCOptim, to curate a basic list of epitopes with optimum population coverage in order to assess the possible effect
of a pan-cancer vaccine made up of these epitopes.For MHC Class II and Class I epitopes, our list's global population coverage varied
from 81.8% to 98.5%, respectively.We created 3D epitope–MHC models for six MHC-I-restricted and four MHC-II-restricted
depitopes based on our list of epitopes, showcasing their interaction with T-cell receptors and potential for epitope binding.The
thorough method of in silico epitope selection used by AutoPepVax tackles vaccine effectiveness, safety, and wide application. Future
research will use mouse tumour models that include the mutations under study to verify the AutoPepVax-designed vaccinations.
References
Li, F.; Deng, L.; Jackson, K.R.; Talukder, A.H.; Katailiha, A.S.; Bradley, S.D.; Zou, Q.; Chen, C.; Huo, C.; Chiu, Y.; et al. Neoantigen vaccination induces clinical and immunologic responses in non-small cell lung cancer patients harboring EGFR mutations.J. Immunother.Cancer2021,9,e002531. [CrossRef][PubMed]
Liu,W.;Tang,H.;Li,L.;Wang,X.;Yu,Z.;Li,J.Peptide-basedtherapeuticcancervaccine:Currenttrendsinclinicalapplication. CellProlif.2021,54,e13025.[CrossRef]
Savsani, K.; Jabbour, G.; Dakshanamurthy, S. A New Epitope Selection Method: Application to Design a Multi-Valent Epitope Vaccine Targeting HRAS Oncogene in Squamous Cell Carcinoma. Vaccines 2021, 10, 63. [CrossRef] [PubMed]
Kim, M.;Savsani,K.;Dakshanamurthy, S.APeptideVaccineDesignTargetingKITMutationsinAcuteMyeloidLeukemia. Pharmaceuticals2023,16,932.[CrossRef][PubMed]
Suri, S.; Dakshanamurthy, S. IntegralVac:A Machine Learning-Based Comprehensive Multivalent Epitope Vaccine Design Method.Vaccines 2022, 10, 1678. [CrossRef] [PubMed]
Ott, P.A.; Hu, Z.; Keskin, D.B.; Shukla, S.A.; Sun, J.; Bozym, D.J.; Zhang, W.; Luoma, A.; Giobbie-Hurder, A.; Peter, L.; et al. An immunogenicpersonalneoantigenvaccineforpatientswithmelanoma. Nature2017,547,217–221. [CrossRef]
Gartner, J.J.; Parkhurst, M.R.; Gros, A.; Tran, E.; Jafferji, M.S.; Copeland, A.; Hanada, K.-I.; Zacharakis, N.; Lalani, A.; Krishna, S.;et al.A machine learning model for ranking candidate HLA class I neoantigens based on known neoepitopes from multiple human tumor types. Nat. Cancer2021, 2, 563–574. [CrossRef] [PubMed]
Reardon, D.A.; Desjardins, A.; Vredenburgh, J.J.; O’Rourke, D.M.; Tran, D.D.; Fink, K.L.; Nabors, L.B.; Li, G.; Bota, D.A.; Lukas, R.V.; et al.Rindopepimut with Bevacizumab for Patients with Relapsed EGFRvIII-Expressing Glioblastoma (ReACT): Results of aDouble- BlindRandomizedPhaseIITrial. Clin.CancerRes.2020,26,1586–1594. [CrossRef]
Weller,M.;Butowski,N.;Tran,D.D.;Recht,L.D.;Lim,M.;Hirte,H.;Ashby,L.;Mechtler,L.;Goldlust,S.A.;Iwamoto,F.; et al.Rindopepimut with temozolomide for patients with newly diagnosed, EGFRvIII-expressing glioblastoma (ACT IV): A randomised,doubleblind,internationalphase3trial.LancetOncol.2017,18,1373–1385.[CrossRef]
Celldex Therapeutics. A Phase II Study of Rindopepimut/GM-CSF in Patients With Relapsed EGFRvIII-Positive Glioblastoma. clinicaltrials.gov. 2020. Availableonline: https://clinicaltrials.gov/study/NCT01498328(accessedon4August2023).
CelldexTherapeutics. AnInternational,Randomized,Double-Blind,ControlledStudyofRindopepimut/GM-CSFWithAdjuvant TemozolomideinPatientsWithNewlyDiagnosed,SurgicallyResected,EGFRvIII-PositiveGlioblastoma.clinicaltrials.gov.2018. Available online: https://clinicaltrials.gov/study/NCT01480479(accessed on 4 August 2023).
MD LZ. Phase I Study of Individualized Neoantigen Peptides in the Treatment of EGFR Mutant Non-Small Cell Lung Cancer. clinicaltrials.gov.2020.Availableonline:https://clinicaltrials.gov/ct2/show/NCT04397926(accessedon12June2023).
Pan, J.; Xiong, D.; Zhang, Q.; Palen, K.; Shoemaker, R.H.; Johnson, B.; Sei, S.; Wang, Y.; You, M. Precision immunointerception ofEGFR-driven tumorigenesis for lung cancer prevention.Front.Immunol.2023, 14, 1036563.[CrossRef]
Ebben, J.D.;Lubet,R.A.;Gad,E.;Disis,M.L.;You,M.EpidermalGrowthFactorReceptorDerivedPeptideVaccinationtoPrevent Lung Adenocarcinoma Formation: An In Vivo Study in a Murine Model of EGFR Mutant Lung Cancer. Mol. Carcinog. 2015, 55, 1517–1525. [CrossRef] [PubMed]
Putzu,C.;Canova,S.;Paliogiannis,P.;Lobrano,R.;Sala,L.;Cortinovis,D.L.;Colonese,F.DurationofImmunotherapyin Non-Small Cell Lung Cancer Survivors: A Lifelong Commitment? Cancers 2023, 15, 689. [CrossRef] [PubMed]
Oronsky,B.;Reid,T.R.;Oronsky,A.;Sandhu,N.;Knox,S.J.AReviewofNewlyDiagnosedGlioblastoma.Front.Oncol.2021,10, 574012.[CrossRef][PubMed]
Hanif,F.;Muzaffar,K.;Perveen,K.;Malhi,S.M.;Simjee,S.U.GlioblastomaMultiforme:AReviewofitsEpidemiologyand PathogenesisthroughClinicalPresentationandTreatment. AsianPac.J.CancerPrev.2017,18,3–9. [PubMed]
Hatanpaa,K.J.;Burma,S.;Zhao,D.;Habib,A.A.EpidermalGrowthFactorReceptorinGlioma:SignalTransduction,Neuropathology,Imaging,andRadioresistance. Neoplasia2010,12,675–684. [CrossRef][PubMed]
Nair,S.;Bonner,J.A.;Bredel,M.EGFRMutationsinHeadandNeckSquamousCellCarcinoma.Int.J.Mol.Sci.2022,23,3818. [CrossRef][PubMed]
Hong,Y.;Kim,J.;Choi,Y.J.;Kang,J.G.Clinicalstudyofcolorectalcanceroperation:Survivalanalysis.KoreanJ.Clin.Oncol.2020, 16,3–8.[CrossRef][PubMed]
Kim, N.; Cho, D.; Kim, H.; Kim, S.; Cha, Y.; Greulich, H.; Bass, A.; Cho, H.; Cho, J. Colorectal adenocarcinoma-derived EGFR mutants are oncogenic and sensitive to EGFR-targeted monoclonal antibodies, cetuximab and panitumumab.Int. J. Cancer 2020,146, 2194–2200. [CrossRef]
Jia, W.; Zhang, T.; Huang, H.; Feng, H.; Wang, S.; Guo, Z.; Luo, Z.; Ji, X.; Cheng, X.; Zhao, R. Colorectal cancer vaccines:The currentscenarioandfutureprospects.Front.Immunol.2022,13,942235.[CrossRef]
Hung,K.;Hayashi,R.;Lafond-Walker,A.;Lowenstein,C.;Pardoll,D.;Levitsky,H.TheCentralRoleofCD4+TCellsinthe AntitumorImmuneResponse. J.Exp. Med. 1998,188,2357–2368. [CrossRef]
Giuntoli,R.L.;Lu,J.;Kobayashi,H.;Kennedy,R.;Celis,E.DirectCostimulationofTumor-reactiveCTLbyHelperTCells PotentiateTheirProliferation,Survival,andEffectorFunction. ClinCancerRes.2002,8,922–931. [PubMed]
Bui, H.-H.; Sidney, J.; Dinh, K.; Southwood, S.; Newman, M.J.; Sette, A. Predicting population coverage of T-cell epitope-based diagnostics and vaccines. BMCBioinform. 2006, 7, 153. [CrossRef]
Gedeon,P.;Choi,B.;Sampson,J.;Bigner,D.Rindopepimut:Anti-EGFRvIIIpeptidevaccine,oncolytic.DrugsFutur.2013,38, 147–155.[CrossRef]
MHCflurry2.0:ImprovedPan-AllelePredictionofMHCClassI-PresentedPeptidesbyIncorporatingAntigenProcessing:Cell Systems.Available online: https://www.cell.com/cell-systems/fulltext/S2405-4712(20)30239- 8?_returnURL=https://linkinghub.elsevier.com/retrieve/pii/S2405471220302398?showall=true(accessed on 17 July 2023).
Wang,M.;Kurgan,L.;Li,M.AcomprehensiveassessmentandcomparisonoftoolsforHLAclassIpeptide-bindingprediction. BriefingsBioinform.2023,24,bbad150.[CrossRef][PubMed]
Reynisson,B.;Alvarez,B.;Paul,S.;Peters,B.;Nielsen,M.NetMHCpan-4.1andNetMHCIIpan-4.0: ImprovedpredictionsofMHC antigenpresentationbyconcurrentmotifdeconvolutionandintegrationofMSMHCelutedliganddata.NucleicAcidsRes.2020, 48, W449– W454. [CrossRef] [PubMed]
Pan-Specific Prediction of Peptide-MHC-I Complex Stability; A Correlate of T Cell Immunogenicity—PMC. Available online: https://www.ncbi.nlm.nih.gov/pmc/articles/PMC4976001/(accessed on 17 July 2023).
Doytchinova,I.A.;Flower,D.R.VaxiJen:Aserverforpredictionofprotectiveantigens,tumourantigensandsubunitvaccines. BMCBioinform.2007,8,4.[CrossRef][PubMed]
Gupta, S.; Kapoor, P.; Chaudhary, K.; Gautam, A.; Kumar, R.; Raghava, G.P.S. In Silico Approach for Predicting Toxicity of Peptides and Proteins. PLoS ONE 2013, 8, 9. [CrossRef] [PubMed]
Dhanda,S.K.;Vir,P.;Raghava,G.P.Designingofinterferon-gammainducingMHCclass-IIbinders.Biol.Direct2013,8,30. [CrossRef][PubMed]
Dimitrov,I.;Bangov,I.;Flower,D.R.;Doytchinova,I.AllerTOPv.2—Aserverforinsilicopredictionofallergens.J.Mol.Model. 2014,20,2278.[CrossRef]
Kyte, J.; Doolittle, R.F. A simple method for displaying the hydropathic character of a protein. J.Mol. Biol. 1982, 157, 105–132. [CrossRef]
Gasteiger, E.; Hoogland, C.; Gattiker, A.; Wilkins, M.R.; Appel, R.D.; Bairoch, A. Protein Identification and Analysis Tools on the ExPASyServer.InTheProteomicsProtocolsHandbook;SpringerProtocolsHandbooks;Walker,J.M.,Ed.;HumanaPress:Totowa, NJ, USA, 2005; pp. 571–607. [CrossRef]
EGFR—Epidermal Growth Factor Receptor—Homo sapiens (Human)|UniProtKB|UniProt. Available online: https://www.uniprot.org/uniprotkb/P00533/entry(accessedon18January2024).
Cosmic. EGFRGene—COSMIC.Availableonline: https://cancer.sanger.ac.uk/cosmic/gene/analysis?ln=EGFR(accessedon18 January 2024).
BankRPD.RCSBPDB:Homepage.Availableonline:https://www.rcsb.org/(accessedon16January2024).
Zhou, P.; Jin, B.; Li, H.; Huang, S.-Y. HPEPDOCK: A web server for blind peptide–protein docking based on a hierarchical algorithm.Nucleic Acids Res. 2018, 46, W443–W450. [CrossRef] [PubMed]
PyMOL.pymol.org.Availableonline:https://pymol.org/2/(accessedon 5 February 2024).
Yin, R.; Ribeiro-Filho, H.V.; Lin, V.; Gowthaman, R.; Cheung, M.; Pierce, B.G. TCRmodel2:High-resolution modeling of T cellreceptorrecognitionusingdeeplearning. NucleicAcidsRes.2023,51,W569–W576. [CrossRef][PubMed]
