Interpretable Deep Learning for De Novo Design of Cell-Penetrating Abiotic Polymers

There are more amino acid permutations within a 40-residue sequence than atoms on Earth. This vast chemical search space hinders the use of human learning to design functional polymers. Here we couple supervised and unsupervised deep learning with high-throughput experimentation to drive the design of high-activity, novel sequences reaching 10 kDa that deliver antisense oligonucleotides to the nucleus of cells. The models, in which natural and unnatural residues are represented as topological fingerprints, decipher and visualize sequence-activity predictions. The new variants boost antisense activity by 50-fold, are effective in animals, are nontoxic, and can also deliver proteins into the cytosol. Machine learning can discover functional polymers that enhance cellular uptake of biotherapeutics, with significant implications toward developing therapies for currently untreatable diseases. One sentence summary Deep learning generates de novo large functional abiotic polymers that deliver antisense oligonucleotides to the nucleus.

[1]  Marilena Loizidou,et al.  Liposomes and nanoparticles: nanosized vehicles for drug delivery in cancer. , 2009, Trends in pharmacological sciences.

[2]  P. Boisguérin,et al.  Delivery of therapeutic oligonucleotides with cell penetrating peptides☆ , 2015, Advanced Drug Delivery Reviews.

[3]  A. Banham,et al.  Therapeutic Antibodies against Intracellular Tumor Antigens , 2017, Front. Immunol..

[4]  Sachdev S. Sidhu,et al.  Intracellular targeting with engineered proteins , 2016, F1000Research.

[5]  Klavs F Jensen,et al.  A fully automated flow-based approach for accelerated peptide synthesis. , 2017, Nature chemical biology.

[6]  T. Nielsen,et al.  Synthesis of proteins by automated flow chemistry , 2020, Science.

[7]  S. Cannistraro,et al.  p28, A first in class peptide inhibitor of cop1 binding to p53 , 2013, British Journal of Cancer.

[8]  William E. Winkler,et al.  String Comparator Metrics and Enhanced Decision Rules in the Fellegi-Sunter Model of Record Linkage. , 1990 .

[9]  Colin M. Fadzen,et al.  Chimeras of Cell-Penetrating Peptides Demonstrate Synergistic Improvement in Antisense Efficacy. , 2019, Biochemistry.

[10]  R. Collier,et al.  Active-site mutations of diphtheria toxin: effects of replacing glutamic acid-148 with aspartic acid, glutamine, or serine. , 1990, Biochemistry.

[11]  Abhishek Das,et al.  Grad-CAM: Visual Explanations from Deep Networks via Gradient-Based Localization , 2016, 2017 IEEE International Conference on Computer Vision (ICCV).

[12]  Evon M. O. Abu-Taieh,et al.  Comparative Study , 2020, Definitions.

[13]  Nithin V. George,et al.  KELM-CPPpred: Kernel Extreme Learning Machine Based Prediction Model for Cell-Penetrating Peptides. , 2018, Journal of proteome research.

[14]  J. Bisi,et al.  Antisense and antigene properties of peptide nucleic acids. , 1992, Science.

[15]  Gajendra P. S. Raghava,et al.  CPPsite 2.0: a repository of experimentally validated cell-penetrating peptides , 2015, Nucleic Acids Res..

[16]  Ű. Langel,et al.  Cargo-dependent cytotoxicity and delivery efficacy of cell-penetrating peptides: a comparative study. , 2007, The Biochemical journal.

[17]  Russ B. Altman,et al.  Predicting HLA class II antigen presentation through integrated deep learning , 2019, Nature Biotechnology.

[18]  Katherine A. Kantardjieff,et al.  The crystal structure of diphtheria toxin , 1992, Nature.

[19]  Markus Voelter,et al.  State of the Art , 1997, Pediatric Research.

[20]  Joel Ruben Antony Moniz,et al.  Nested LSTMs , 2018, ACML.

[21]  Jon D Moulton,et al.  Gene Knockdowns in Adult Animals: PPMOs and Vivo-Morpholinos , 2009, Molecules.

[22]  Alessandro Sette,et al.  Properties of MHC Class I Presented Peptides That Enhance Immunogenicity , 2013, PLoS Comput. Biol..

[23]  Emma J. Chory,et al.  A Deep Learning Approach to Antibiotic Discovery , 2020, Cell.

[24]  Ü. Langel,et al.  Cell-penetrating peptides with intracellular organelle targeting , 2017, Expert opinion on drug delivery.

[25]  S. Hammond,et al.  Pip6-PMO, A New Generation of Peptide-oligonucleotide Conjugates With Improved Cardiac Exon Skipping Activity for DMD Treatment , 2012, Molecular therapy. Nucleic acids.

[26]  P. Iversen,et al.  Vectorization of morpholino oligomers by the (R-Ahx-R)4 peptide allows efficient splicing correction in the absence of endosomolytic agents. , 2006, Journal of controlled release : official journal of the Controlled Release Society.

[27]  A. Eshraghi,et al.  Preclinical and clinical otoprotective applications of cell-penetrating peptide D-JNKI-1 (AM-111) , 2018, Hearing Research.

[28]  C. Brophy,et al.  Internalization and intracellular trafficking of a PTD-conjugated anti-fibrotic peptide, AZX100, in human dermal keloid fibroblasts. , 2010, Journal of pharmaceutical sciences.

[29]  A. S. Sobolev,et al.  Targeted Intracellular Delivery of Antibodies: The State of the Art , 2018, Front. Pharmacol..

[30]  A. Kurihara,et al.  Distribution of KAI-9803, a Novel δ-Protein Kinase C Inhibitor, after Intravenous Administration to Rats , 2011, Drug Metabolism and Disposition.

[31]  K. Lim,et al.  Eteplirsen in the treatment of Duchenne muscular dystrophy , 2017, Drug design, development and therapy.

[32]  David Rogers,et al.  Extended-Connectivity Fingerprints , 2010, J. Chem. Inf. Model..

[33]  M. C. Cardoso,et al.  Covalent attachment of cyclic TAT peptides to GFP results in protein delivery into live cells with immediate bioavailability. , 2015, Angewandte Chemie.

[34]  Susan M. Bridges,et al.  Prediction of Cell Penetrating Peptides by Support Vector Machines , 2011, PLoS Comput. Biol..

[35]  P. Iversen,et al.  Sustained dystrophin expression induced by peptide-conjugated morpholino oligomers in the muscles of mdx mice. , 2008, Molecular therapy : the journal of the American Society of Gene Therapy.

[36]  J. Beckmann,et al.  Cell-permeable peptides induce dose- and length-dependent cytotoxic effects. , 2007, Biochimica et biophysica acta.

[37]  E. Myers,et al.  Basic local alignment search tool. , 1990, Journal of molecular biology.

[38]  Tarmo Tamm,et al.  Prediction of Cell-Penetrating Peptides Using Artificial Neural Networks. , 2010, Current computer-aided drug design.

[39]  Gary D Stormo,et al.  An Introduction to Sequence Similarity (“Homology”) Searching , 2009, Current protocols in bioinformatics.

[40]  Balachandran Manavalan,et al.  Machine-Learning-Based Prediction of Cell-Penetrating Peptides and Their Uptake Efficiency with Improved Accuracy. , 2018, Journal of proteome research.

[41]  Jie Hu,et al.  Empirical comparison and analysis of web-based cell-penetrating peptide prediction tools , 2019, Briefings Bioinform..

[42]  R. Juliano,et al.  The delivery of therapeutic oligonucleotides , 2016, Nucleic acids research.

[43]  Federica Gemignani,et al.  Systemically delivered antisense oligomers upregulate gene expression in mouse tissues , 2002, Nature Biotechnology.

[44]  Q. Zou,et al.  SkipCPP-Pred: an improved and promising sequence-based predictor for predicting cell-penetrating peptides , 2017, BMC Genomics.

[45]  Andrew L. Ferguson,et al.  Machine learning-enabled discovery and design of membrane-active peptides. , 2017, Bioorganic & medicinal chemistry.

[46]  Jessica Chery.  RNA therapeutics: RNAi and antisense mechanisms and clinical applications. , 2016, Postdoc journal : a journal of postdoctoral research and postdoctoral affairs.

[47]  V. Rotello,et al.  Promises and Pitfalls of Intracellular Delivery of Proteins , 2014, Bioconjugate chemistry.

[48]  Justin M. Wolfe,et al.  Machine Learning To Predict Cell-Penetrating Peptides for Antisense Delivery , 2018, ACS central science.

[49]  Alán Aspuru-Guzik,et al.  Deep learning enables rapid identification of potent DDR1 kinase inhibitors , 2019, Nature Biotechnology.