Generation of functional oligopeptides that promote osteogenesis based on unsupervised deep learning of protein IDRs

Abstract

Deep learning (DL) is currently revolutionizing peptide drug development due to both computational advances and the substantial recent expansion of digitized biological data. However, progress in oligopeptide drug development has been limited, likely due to the lack of suitable datasets and difficulty in identifying informative features to use as inputs for DL models. Here, we utilized an unsupervised deep learning model to learn a semantic pattern based on the intrinsically disordered regions of ~171 known osteogenic proteins. Subsequently, oligopeptides were generated from this semantic pattern based on Monte Carlo simulation, followed by in vivo functional characterization. A five amino acid oligopeptide (AIB5P) had strong bone-formation-promoting effects, as determined in multiple mouse models (e.g., osteoporosis, fracture, and osseointegration of implants). Mechanistically, we showed that AIB5P promotes osteogenesis by binding to the integrin α5 subunit and thereby activating FAK signaling. In summary, we successfully established an oligopeptide discovery strategy based on a DL model and demonstrated its utility from cytological screening to animal experimental verification.

References

[1] Erak M et al. Peptide chemistry toolbox—transforming natural peptides into peptide therapeutics. Bioorg. Med. Chem., 2018, 26: 2759-2765

[2] Hibino E, Hoshino M. A novel mode of interaction between intrinsically disordered proteins. Biophys. Physicobiol., 2020, 17: 86-93

[3] Jiang T et al. Enhanced transdermal drug delivery by transfersome- embedded oligopeptide hydrogel for topical chemotherapy of melanoma. ACS Nano, 2018, 12: 9693-9701

[4] Ling Z et al. Effects of oligopeptide simulating DMP-1/mineral trioxide aggregate/agarose hydrogel biomimetic mineralisation model for the treatment of dentine hypersensitivity. J. Mater. Chem. B, 2019, 7: 5825-5833

[5] Oliynyk AO, Mar A. Discovery of intermetallic compounds from traditional to machine-learning approaches. Acc. Chem. Res., 2018, 51: 59-68

[6] Cunningham JM et al. Biophysical prediction of protein-peptide interactions and signaling networks using machine learning. Nat. Methods, 2020, 17: 175-183

[7] Ji Y et al. Functional oligopeptide as a novel strategy for drug delivery. J. Drug Target, 2017, 25: 597-607

[8] Adam SP et al. No free lunch theorem: a review. Approximation Optim.: Algorithms, Complex. Appl., 2019, 145: 57-82

[9] Guharoy M et al. SnapShot: intrinsic structural disorder. Cell, 2015, 161: 1230-1230

[10] Wright PE, Dyson HJ. Intrinsically disordered proteins in cellular signalling and regulation. Nat. Rev. Mol. Cell Biol., 2015, 16: 18-29

[11] Zhou BO et al. Leptin-receptor-expressing mesenchymal stromal cells represent the main source of bone formed by adult bone marrow. Cell Stem Cell, 2014, 15: 154-168

[12] Sun Y et al. The long noncoding RNA lnc-ob1 facilitates bone formation by upregulating Osterix in osteoblasts. Nat. Metab., 2019, 1: 485-496

[13] Sun X et al. The Nrf2 activator RTA-408 attenuates osteoclastogenesis by inhibiting STING-dependent NF-κb signaling. Redox Biol., 2020, 28: 101309

[14] Zhang H et al. Gain-of-function RHOA mutations promote focal adhesion kinase activation and dependency in diffuse gastric cancer. Cancer Discov., 2020, 10: 288-305

[15] Jumper J et al. Highly accurate protein structure prediction with AlphaFold. Nature, 2021, 596: 583-589

[16] Suen CY. n-Gram statistics for natural language understanding and text processing. IEEE Trans. Pattern Anal. Mach. Intell., 1979, 1: 164-172

[17] Elman JL. Finding structure in time. Cogn. Sci., 1990, 14: 179-211

[18] Vaswani, A. et al. Attention is all you need. Proc. Adv. Neural Inf. Process. Syst. 30, 5998–6008 (2017).

[19] Ho QT et al. FAD-BERT: improved prediction of FAD binding sites using pre-training of deep bidirectional transformers. Comput. Biol. Med., 2021, 131: 104258

[20] Wang, Y. et al. Improving N-gram language models with pre-trained deep transformer. arXiv:1911.10235 (2019).

[21] Gsponer J et al. Tight regulation of unstructured proteins: from transcript synthesis to protein degradation. Science, 2008, 322: 1365-1368

[22] Meinema AC et al. Long unfolded linkers facilitate membrane protein import through the nuclear pore complex. Science, 2011, 333: 90-93

[23] Oldfield CJ, Dunker AK. Intrinsically disordered proteins and intrinsically disordered protein regions. Annu. Rev. Biochem., 2014, 83: 553-584

[24] van der Lee R et al. Intrinsically disordered segments affect protein half-life in the cell and during evolution. Cell Rep., 2014, 8: 1832-1844

[25] Gibson BA et al. Organization of chromatin by intrinsic and regulated phase separation. Cell, 2019, 179: 470-484

[26] van der Lee R et al. Classification of intrinsically disordered regions and proteins. Chem. Rev., 2014, 114: 6589-6631

[27] Marie PJ. Targeting integrins to promote bone formation and repair. Nat. Rev. Endocrinol., 2013, 9: 288-295

[28] Clark AY et al. Integrin-specific hydrogels modulate transplanted human bone marrow-derived mesenchymal stem cell survival, engraftment, and reparative activities. Nat. Commun., 2020, 11: 114-128

[29] Jiang M et al. Identification of osteogenic progenitor cell-targeted peptides that augment bone formation. Nat. Commun., 2020, 11: 4278-4292

[30] Marie PJ et al. Integrin and cadherin signaling in bone: role and potential therapeutic targets. Trends Endocrinol. Metab., 2014, 25: 567-575

[31] Zhang D et al. CircRNA-vgll3 promotes osteogenic differentiation of adipose-derived mesenchymal stem cells via modulating miRNA-dependent integrin alpha5 expression. Cell Death Differ., 2021, 28: 283-302

[32] Hamidouche Z et al. Priming integrin alpha5 promotes human mesenchymal stromal cell osteoblast differentiation and osteogenesis. Proc. Natl. Acad. Sci. U.S.A., 2009, 106: 18587-18591

[33] Srouji S et al. Lentiviral-mediated integrin alpha5 expression in human adult mesenchymal stromal cells promotes bone repair in mouse cranial and long-bone defects. Hum. Gene Ther., 2012, 23: 167-172

[34] Fromigue O et al. Peptide-based activation of alpha5 integrin for promoting osteogenesis. J. Cell. Biochem., 2012, 113: 3029-3038

[35] Shao PL et al. Alpha-5 integrin mediates simvastatin-induced osteogenesis of bone marrow mesenchymal stem cells. Int. J. Mol. Sci., 2019, 20: 506-521

[36] Fosgerau K, Hoffmann T. Peptide therapeutics: current status and future directions. Drug Discov. Today, 2015, 20: 122-128

[37] Erdos G, Dosztanyi Z. Analyzing protein disorder with IUPred2A. Curr. Protoc. Bioinforma., 2020, 70: e99

[38] Yue R et al. Leptin receptor promotes adipogenesis and reduces osteogenesis by regulating mesenchymal stromal cells in adult bone marrow. Cell Stem Cell, 2016, 18: 782-796

[39] Zhang C et al. Expression of IFT140 during bone development. J. Histochem. Cytochem., 2019, 67: 723-734

[40] Zhang L et al. Overexpression of MiR-335-5p promotes bone formation and regeneration in mice. J. Bone Min. Res., 2017, 32: 2466-2475

[41] Wu CH et al. alpha-Enolase-binding peptide enhances drug delivery efficiency and therapeutic efficacy against colorectal cancer. Sci. Transl. Med., 2015, 7: 290-305

Funding

National Natural Science Foundation of China (National Science Foundation of China)(8206113022, 92049201, 81770873, 81822012, 81771043, 81802193, 81970898)