2014³â ÁîÀ½, ³ª´Â ´ÙÆ®¸Ó½º ´ëÇÐ ÄÄÇ»ÅÍ Çаú¿¡¼ ¹Ú»ç ÈÄ ¿¬±¸¿ø »ýÈ°À» ÇÏ¸é¼ ºñÇ×ü ´Ü¹éÁú Ä¡·áÁ¦ÀÇ ¸é¿ª¿ø¼º(immunogenicity) Á¦°Å¸¦ ÇÒ ¼ö ÀÖ´Â ´ÙÁß ÃÖÀûÈ ¾Ë°í¸®ÁòÀ» °³¹ßÇÑ ÈÄ [1, 2], ´Ù¸¥ ¿©·¯ ¿¬±¸¿¡ ÀÀ¿ëÇÏ°í ÀÖ´ø ÁßÀ̾ú´Ù [3, 4]. ´ç½Ã ´ÙÆ®¸Ó½º ÀÇ´ë¿¡¼ °³¹ßÇÏ´ø CAR NK ¼¼Æ÷°¡ ÀÖ¾ú´Âµ¥, ±× ¼¼Æ÷¿¡ ¾²ÀÏ scFv ¼¿À» Àΰ£È(antibody humanization)ÇØ´Þ¶ó´Â ÀÇ·Ú°¡ µé¾î¿Ô´Ù. ±× Ç×ü°¡ ¸ðÁ¾ÀÇ ÀÌÀ¯·Î ÀüÅëÀûÀÎ CDR À̽ĹýÀÌ Àß ¾È µÈ´Ù´Â °ÍÀ̾ú´Ù. Ç×üÀΰ£È¶ó´Â °Ô °á±¹ ¸é¿ª¿ø¼ºÀ» Á¦°ÅÇϱâ À§ÇÑ °ÍÀÌ´Ï °³¹ßÇÑ ¾Ë°í¸®ÁòÀ» µµÀÔÇϸé ÇØ°áÀÌ µÉ±î ½Í¾î ¿¬¶ôÀÌ ¿Â °ÍÀ̾ú´Ù. ±×µéÀÇ ±â´ë¿Í ´Þ¸® ´ç½Ã ±× ¹æ¹ý·ÐÀº Ç×ü¿¡ ¾µ ¼ö ÀÖ´Â °ÍÀÌ ¾Æ´Ï¾î¼ ³»Ä£ ±è¿¡ Ç×ü Àΰ£È¸¦ À§ÇÑ °è»ê ¹æ¹ý·ÐÀ» ¾Æ¿¹ µû·Î °³¹ßÇÏ°Ô µÇ¾ú´Ù [5].
»ç½Ç ³ª´Â Ç×ü¿Í °ü·ÃÇÑ ¿¬±¸´Â °¡´ÉÇÑ ±âÇÇÇÏ´Â ÁßÀ̾ú´Ù. ÀÌ¹Ì Ç×ü ºÐ¾ß¿¡ ¶Ù¾î³ ¿¬±¸ÀÚ°¡ ±²ÀåÈ÷ ¸¹¾Ò°í, Ä¡·á¿ë/°ËÁø¿ë Ç×ü °³¹ßÀº »ê¾÷ü·Î ³Ñ¾î°£ »óȲÀÌ´Ù. ÇöÀç Á¦¾à ½ÃÀå¿¡¼ °¡Àå ¸¹ÀÌ Æȸ®´Â ¾àÀº ´Ü¿¬ÄÚ Ç×ü´Ù [6]. ÇÏÁö¸¸, °è»ê »ý¹°Çаú ´Ü¹éÁú ¼³°è ±â¼úÀÌ ÀÌÁ¦ ½Å¾à°³¹ß¿¡ Çʼö µµ±¸°¡ µÇ¾úÀ½¿¡µµ ºÒ±¸ÇÏ°í, Ç×ü Àΰ£È °ü·Ã Àü¹® °è»ê ¹æ¹ý·ÐÀº °ÅÀÇ ¾ø´Ù´Â »ç½ÇÀ» ¾Ë°Ô µÇ¾ú´Ù. ³í¹® ¸®ºä¾îµµ ¾ÆÁ÷±îÁö ÀÌ·± ¹æ¹ý·ÐÀº µé¾îº» ÀûÀÌ ¾ø´Ù´Â ÆòÀ» ³²°å´Âµ¥, ±× ÀÌÀ¯¸¦ 2014³â ISMB¿Í IBC Ç×ü ÇÐȸ ¿¬·Ê ¸ðÀÓ¿¡ ÃÊû °ÀÇ µîÀ¸·Î Âü¼®ÇÏ°Ô µÇ¸é¼ ¾Ë°Ô µÇ¾ú´Ù. ³» ¿¬±¸¿¡ ´ëÇÑ Æò°¡´Â ±Ø°ú ±ØÀ̾ú´Ù. ƯÈ÷ ¸ð ´ëÇü ´Ù±¹Àû Á¦¾àȸ»ç¿¡¼ ¿À½Å ºÐÀÇ ½Å¶öÇÑ ÆòÀÌ Èï¹Ì·Î¿ü´Ù. Ç×ü Àΰ£È´Â ÀÌ¹Ì Ç®¸° ¹®Á¦¶ó Çй®ÀûÀ¸·Îµµ Èï¹Ì°¡ ¾øÀ»»Ó´õ·¯, ȸ»ç ³» ¿¬±¸ ÆÀ¿¡ ±×·± À¯»çÇÑ °Ô ÀÖÀ» °ÍÀ̶ó´Â ÆòÀ̾ú´Ù. ¹Ý¸é, ¼ÒÇü º¥Ã³ Á¦¾àȸ»ç, ȤÀº ´ëÇÐ ¿¬±¸½Ç µî¿¡¼ ¿À½Å ºÐµéÀº Å« °ü½ÉÀ» º¸¿´´Ù. º¥Ã³¸¦ ½ÃÀÛÇß´Ù´Â ¹Ì±¹ÀÇ ÇÑ ±³¼ö´Â ÀÌ·± À̾߱⸦ ³»°Ô Çß´Ù. ¡°Ç×»ó ÀÌ·± °Ô ÀÖÀ¸¸é ÁÁ°Ú´Ù´Â »ý°¢À» Çß¾ú´Ù.¡±
±× »çÀÌ ½Ã°£ÀÌ È帣¸é¼ Çй®¿¡µµ ¸¹Àº º¯È°¡ »ý°å´Ù. SF ¼Ò¼³À̳ª ¿µÈ, ȤÀº öÇÐÀÚµéÀÇ Èï¹Ì·Î¿î ³íÀï°Å¸® Á¤µµ¿´´ø ¡°ÀΰøÁö´É¡±ÀÌ Çö½Ç¿¡±îÁö µé¾î¿Í, ÀÌÁ¦´Â °Ç³Ê°¡Áö ¾ÊÀ¸¸é ¾È µÉ Å« È帧ÀÌ µÇ¾ú´Ù. ƯÈ÷ ÈùÆ°HintonÀÇ µö·¯´×[7]°ú 2012³â ÀÌÈÄ [8], ÀÌÁ¦ ÀΰøÁö´ÉÀº ÄÄÇ»ÅÍ°úÇÐ»Ó ¾Æ´Ï¶ó Àü Çй® ¿µ¿ªÀ¸·Î ÀÀ¿ëµÇ°í ÀÖ´Ù. Çѵ¿¾È Ç×ü ¿¬±¸¿Í °Å¸®¸¦ µÎ¾ú´ø ³ª´Â ³í¹® ¸®ºä¸¦ ÇÏ¸é¼ Ç×ü °ü·Ã ³í¹®À» ´Ù½Ã ã¾Æº¸°Ô µÇ¾ú´Ù. ±×°£ Ç×ü °øÇÐ ±â¼ú¿¡µµ ÀΰøÁö´É ¹Ù¶÷ÀÌ ºÒ°í ÀÖÀ» °ÍÀ̶ó ±â´ëÇßÀ¸³ª ½ÇÁ¦´Â ±×·¸Áö ¾Ê¾Ò´Ù. 2019³â¿¡ µé¾î¼¾ß µö·¯´× È°¿ë ³í¹®ÀÌ ³ª¿Â ÀÌÈÄ [9], Á¢±Ù °¡´ÉÇÑ ¹æ¹ý·ÐÀº ¿©ÀüÈ÷ °³¹ßµÇÁö ¾Ê°í ÀÖ¾ú´Ù. ¿¬±¸ ºÐ¾ß ÀÚüÀÇ ¼º¼÷µµ¿Í ÀΰøÁö´ÉÀÇ È°¿ë¼º, »ê¾÷ü·ÎÀÇ ÀÀ¿ë¼ºÀÌ ¸ðµÎ ÃæÁ·µÇ¾úÀ½¿¡µµ ±Ù 10³â °¡±î¿î ¼¼¿ù µ¿¾È Å« º¯È°¡ ¾ø´Ù´Â °Ç Àǿܶó°í ¹Û¿¡ ÇÒ ¼ö ¾ø¾ú´Ù. ¾î¼¸é ´ëÇü ´Ù±¹Àû Á¦¾àȸ»ç¿¡¼± ½Ã´ë¿¡ ¹ß¸ÂÃß¾î ³»ºÎÀûÀÎ ¹æ¹ý·ÐÀ» ±¸ÃàÇØ ³õ¾ÒÀ» ¼öµµ ÀÖ´Ù. Å«µ·À» ÅõÀÚÇÏ°í ÀÖÀ¸´Ï ±×µéÀº ¹º°¡ ¾öû³ °É ÀÌ¹Ì Çسõ¾ÒÀ» °Å¾ß, ¶ó´Â °Ç ¾î¼¸é °¡´É¼ºÀÌ ÀÖ´Â ¹ÏÀ½ÀÌ´Ù. ±×¸®°í ³ª´Â ¾ó¸¶ Àü¿¡ ±¹³»ÀÇ ²Ï Å« ÇÑ Á¦¾à¾÷ü·ÎºÎÅÍ 2014³â¿¡ µé¾ú´ø °Í°ú À¯»çÇÑ À̾߱⸦ µè°í¾ß ¸»¾Ò´Ù. ¡°¾Æ¸¶ ´Ù±¹Àû ±â¾÷µéÀº ÀΰøÁö´É µµÀÔÇؼ ÀÌ¹Ì ³»ºÎÀûÀÎ ¹æ¹ý·ÐÀÌ ÀÖÁö ¾ÊÀ»±î¿ä? ¿ì¸®µµ ±×·± °Ô ÀÖ´Ù¸é Âü ÁÁÀ» °Í °°Àºµ¥ ¸»ÀÌÁÒ.¡±
¿Á½ºÆÛµå ´ëÇб³ Charlotte Deane ±³¼öÀÇ ´Ü¹éÁú Á¤º¸ÇÐ ¿¬±¸ÆÀ(Oxford Protein Informatics Group; OPIG)Àº ³»°¡ ¹Ú»ç °úÁ¤»ýÀ¸·Î ÀÖÀ» ¶§ óÀ½À¸·Î Ç×ü ¿¬±¸¸¦ ½ÃÀÛÇØ Áö±ÝÀº Ç×ü °ü·Ã °è»ê ¹æ¹ý·Ð ¿¬±¸¸¦ ÁÖµµÇÏ´Â ¿¬±¸ ±×·ìÀÌ µÇ¾ú´Ù. ±×°£ Ç×ü ¿¬±¸¿Í °ü·ÃµÈ ¿©·¯ °è»ê µµ±¸¿Í µ¥ÀÌÅͺ£À̽ºµéÀ» Â÷±ÙÂ÷±Ù ±¸ÃàÇÏ´õ´Ï À۳⠰ܿï, Hu-mabÀ̶ó´Â Ç×ü Àΰ£È °è»ê ¹æ¹ý·ÐÀ» À¥¼¹ö·Î °ø°³ÇÏ¿´´Ù [10]. ÀÌ ¹æ¹ý·ÐÀº The observed antibody space database (OAS) [11]°¡ °ø°³µÇ¸é¼ °¡´ÉÇÏ°Ô µÇ¾ú´Ù. Hu-mab ¼¹ö´Â Å©°Ô Ç×üÀÇ Àΰ£È Á¤µµ(humanness)¸¦ Æò°¡ÇÏ´Â ±â´É°ú Youden¡¯s J statistic(YJS)À» ¸ñÀûÇÔ¼ö·Î ÇÏ¿© Àΰ£È ÇÏ´Â Ç×ü Àΰ£È ºÎºÐÀÌ ÀÖ´Ù. Àΰ£È Á¤µµ´Â Random Forest (RF)·Î Àΰ£/ºñÀΰ£ Ç×ü¸¦ ±¸ºÐÇϴµ¥, ¾î´À V-gene¿¡ ¼ÓÇÏ´ÂÁö °á°ú·Î ¾Ë·ÁÁØ´Ù. ³í¹®¿¡ µû¸£¸é, Ä¡·á¿ë Ç×ü Áß Àΰ£ Ç×ü¿¡ ´ëÇÑ Á¤¹Ðµµ(precision)´Â 1ÀÌ´Ù. Ç×ü Àΰ£È ºÎºÐ¿¡´Â ¾î´À V-geneó·³ Àΰ£È ÇÒ Áö ¼±Åà °¡´ÉÇϸç, YJS °ªÀÇ ¹®ÅÎÁö¼ö¸¦ ¼³Á¤ÇÒ ¼ö ÀÖ´Ù. Âü°í·Î ¸é¿ª¿ø¼ºÀÌ ¾Ë·ÁÁø 217°³ÀÇ Ç×ü¸¦ ´ë»óÀ¸·Î ÇÑ ÈÄÇ⿬±¸¿¡¼´Â YJS °ªÀ» 0.9 ÀÌ»óÀ¸·Î ÇßÀ» ½Ã ´Ü 1°³ÀÇ ¼¿À» Á¦¿ÜÇÏ°ï ¸ðµÎ ¸é¿ª¿ø¼ºÀÌ ³·Àº Ç×ü·Î ±¸ºÐÇØ ³»¾ú´Ù.
Hu-mabÀÇ ¼º´É°ú ¾Æ¿ï·¯ ÁÖ¸ñÇÒ ºÎºÐÀº »ç¿ëÀÇ ÆíÀǼºÀÌ´Ù. ±×°£ OPIGÀÇ Ç×ü °ü·Ã ¿©·¯ À¥¼ºñ½ºµé°ú °°Àº µðÀÚÀÎÀ¸·Î °£°áÇϸç ÅëÀϼº ÀÖ°Ô ±¸¼ºµÇ¾ú°í, ÀÎÅÍÆäÀ̽º´Â Á÷°üÀûÀÌ´Ù. º¹ÀâÇÑ ¼³Á¤Àº ÈÄÇâ ¿¬±¸¸¦ ÅëÇØ °¡Àå ÁÁÀº °á°ú¸¦ ³¾ ¼ö ÀÖµµ·Ï ±âº»°ªÀ¸·Î ¼³Á¤µÇ¾î ÀÖÀ¸¸ç »ç¿ëÀÚ´Â ±âº»ÀûÀ¸·Î Ç×ü ¼¿¸¸ ³ÖÀ¸¸é ºü¸£°Ô °á°ú¸¦ º¼ ¼ö ÀÖ´Ù. OAS µ¥ÀÌÅͺ£À̽º ÀÌÈÄ, Hu-mab ¿Ü¿¡µµ BioPhi[12]¿Í °°Àº À¥¼¹ö°¡ °ø°³µÇ¾úÀ¸¸ç µö·¯´×À» È°¿ëÇÑ Ç×ü ¿¬±¸µéÀÌ ³ª¿À´Â ÁßÀÌ´Ù[13, 14].
¿Ö ÀÌ·± ¹æ¹ý·ÐÀÌ Áö±Ý±îÁö ³ª¿ÀÁö ¾Ê¾Ò´ÂÁö¿¡ ´ëÇÑ ºÎºÐÀ» Â÷Ä¡ÇÏ°íµµ, Áö±ØÈ÷ °³ÀÎÀûÀ¸·Î Èï¹Ì·Î¿î ºÎºÐÀÌ Çϳª ´õ ÀÖ´Ù. Deane ±³¼ö´Â 2010³â ÀÌÀü¸¸ Çصµ ±â°èÇнÀÀ» È°¿ëÇÑ ¿¬±¸¿¡ ºÎÁ¤ÀûÀÎ »ç¶÷À̾ú´Ù. ±â°èÇнÀ ȤÀº ÀΰøÁö´ÉÀÇ ¼º´ÉÀ̳ª °¡´É¼º¿¡ ´ëÇÑ È¸ÀÇ°¡ ¾Æ´Ï¶ó, ±â°è°¡ µ¥ÀÌÅÍ·Î ¾ò¾î³½ ±ÔÄ¢À» Àΰ£ÀÌ ÁøÁ¤À¸·Î ÀÌÇØÇÏ°í ÀÖ´Â °ÍÀ̳Ĵ °úÇÐÀڷμÀÇ È¸ÀÇ°¨ ¶§¹®À̾ú´Ù.
Ä® ¼¼ÀÌ°ÇÀÇ ¸»À» ºô¸®ÀÚ¸é ³ú°¡ ºüÁ®³ª°¥ Á¤µµ·Î ¿¸®Áö¸¸ ¾Ê´Â´Ù¸é ¿¸° ¸¶À½Àº °úÇÐÀÚÀÇ ÈǸ¢ÇÑ ´ö¸ñÀ̶ó ÇÒ ¼ö ÀÖ´Ù. ±×³à´Â ÃÖ±Ù 2022³â 1¿ù¿¡ µö·¯´×À» È°¿ëÇÑ Ç×ü ½ºÅ©¸®´× ³í¹®À» Çϳª ´õ Ãâ°£ÇÏ¿´´Ù [15]. ±×¸®°í, ÀΰøÁö´É ½Å¾à °³¹ß ȸ»çÀÎ, Exscientia¿¡ ¼ö¼® °úÇÐÀÚ °âÁ÷À¸·Î ÇÕ·ùÇÏ¿´´Ù.
• The Observed Antibody Space: http://opig.stats.ox.ac.uk/webapps/oas/
• Hu-mab: http://opig.stats.ox.ac.uk/webapps/newsabdab/sabpred/humab
• BioPhi: https://biophi.dichlab.org/
Choi, Y., et al., EpiSweep: Computationally driven reengineering of therapeutic proteins to reduce immunogenicity while maintaining function, in Computational Protein Design. 2017, Springer. p. 375-398.
Parker, A.S., et al., Structure-guided deimmunization of therapeutic proteins. Journal of Computational Biology, 2013. 20(2): p. 152-165.
Blazanovic, K., et al., Structure-based redesign of lysostaphin yields potent antistaphylococcal enzymes that evade immune cell surveillance. Molecular Therapy-Methods & Clinical Development, 2015. 2: p. 15021.
Salvat, R.S., et al., Computationally optimized deimmunization libraries yield highly mutated enzymes with low immunogenicity and enhanced activity. Proceedings of the National Academy of Sciences, 2017. 114(26): p. E5085-E5093.
Choi, Y., et al. Antibody humanization by structure-based computational protein design. in MAbs. 2015. Taylor & Francis.
Urquhart, L., Top drugs and companies by sales in 2018. Nature Reviews Drug Discovery, 2019: p. NA-NA.
Hinton, G.E., Learning multiple layers of representation. Trends in cognitive sciences, 2007. 11(10): p. 428-434.
Krizhevsky, A., I. Sutskever, and G.E. Hinton, Imagenet classification with deep convolutional neural networks. Advances in neural information processing systems, 2012. 25.
Wollacott, A.M., et al., Quantifying the nativeness of antibody sequences using long short-term memory networks. Protein Engineering, Design and Selection, 2019. 32(7): p. 347-354.
Marks, C., et al., Humanization of antibodies using a machine learning approach on large-scale repertoire data. Bioinformatics, 2021. 37(22): p. 4041-4047.
Kovaltsuk, A., et al., Observed antibody space: a resource for data mining next-generation sequencing of antibody repertoires. The Journal of Immunology, 2018. 201(8): p. 2502-2509.
Prihoda, D., et al., BioPhi: A platform for antibody design, humanization and humanness evaluation based on natural antibody repertoires and deep learning. bioRxiv, 2021.
M©©okosiewicz, J., et al., AbDiver-A tool to explore the natural antibody landscape to aid therapeutic design. bioRxiv, 2021.
Ruffolo, J.A., J. Sulam, and J.J. Gray, Antibody structure prediction using interpretable deep learning. Patterns, 2021: p. 100406.
Schneider, C., et al., DLAB: deep learning methods for structure-based virtual screening of antibodies. Bioinformatics, 2022. 38(2): p. 377-383.
¼º±Õ°ü´ëÇб³, ¹®Çлç, ÀÌÇлç (öÇÐ, Áß±¹Ã¶ÇÐ, ¹°¸®ÇÐ)
King¡¯s College London, MSc (Mathematics)
The University of Oxford (St. Cross College), DPhil (Statistics)
Dartmouth College (Computer Science), Research Associate
Àü³²´ëÇб³ ÀÇ°ú´ëÇÐ, ºÎ±³¼ö