Synthesis and characterization of molecularly imprinted polymer receptors targeting the c-terminus of amyloid-beta via epitope imprinting

Full text

(1)LICENTIATUPPSATS. F O U R A P P O R T 2 016 : 4. MARK GAL AT SYNTHESIS AND CHARACTERIZATION OF MOLECULARLY IMPRINTED POLYMER RECEPTORS TARGETING THE C-TERMINUS OF AMYLOID-BETA VIA EPITOPE IMPRINTING.

(2)

(3) S Y N T H E S I S A N D C H A R AC T E R I Z AT I O N O F M O L E C U L A R LY I M P R I N T E D P O LY M E R R E C E P T O R S TA R G E T I N G T H E C - T E R M I N U S OF AMYLOID-BETA VIA EPITOPE IMPRINTING.

(4) Malmö University FoU rapport 2016:4. © Copyright Mark Galat, 2016 ISBN 978-91-7104-690-1 (print) ISBN 978-91-7104-691-8 (pdf) ISSN 1650-2337 Holmbergs, Malmö 2016.

(5) MARK GAL AT SYNTHESIS AND CHAR ACTERIZATION OF MOLECUL ARLY IMPRINTED POLYMER RECEPTORS TARGETING THE C-TERMINUS OF AMYLOID-BETA VIA EPITOPE IMPRINTING. Malmö University, 2016 Faculty of Health and Society.

(6)

(7) CONTENTS TABLE OF ABBREVIATIONS................................................................ I ACKNOWLEDGEMENTS..................................................................III 1. INTRODUCTION AND OBJECTIVES................................................ 1 1.1. Abstract............................................................................................ 1 1.2. Populärvetenskaplig sammanfattning.................................................... 2 1.3. Summary........................................................................................... 4 1.4. Research questions............................................................................. 7 2. STATE OF THE ART AND LITERATURE BACKGROUND........................ 8 2.1. Alzheimer’s Disease and its diagnosis................................................... 8 2.2. An important biomarker: Aβ.............................................................. 10 2.3. The art of plastic antibodies: molecular imprinting................................ 14 2.4. Molecular imprinting formats and techniques....................................... 19 2.5. Peptide and protein imprinting........................................................... 27 2.6. Epitope imprinting approach for recognition of macromolecules............. 29 3. MATERIALS AND METHODS........................................................ 33 3.1. Overview........................................................................................ 33 3.2. Materials......................................................................................... 33 3.3. Experimental methods....................................................................... 38 3.3.1. Elemental analysis (EA).............................................................. 38 3.3.2. Optical microscopy................................................................... 39 3.3.3. Fourier-transform infrared spectroscopy (FT-IR)............................... 39 3.3.4. Nuclear magnetic resonance spectroscopy (NMR)........................ 40 3.3.5. Ninhydrin test........................................................................... 40 3.3.6. Thermogravimetric analysis (TGA)............................................... 41 3.3.7. Template rebinding tests............................................................. 42 3.3.8. High-performance liquid chromatography (HPLC).......................... 43 3.3.9. Correlation analysis.................................................................. 46 3.3.10. List of miscellaneous instruments................................................ 46.

(8) 4. BULK IMPRINTED POLYMERS TARGETING THE C-TERMINUS OF Aß. . .. 49 4.1. Introduction..................................................................................... 49 4.2. Results and discussion....................................................................... 49 4.2.1. Selection of templates................................................................ 49 4.2.2. Bulk polymerization................................................................... 51 4.2.3. Characterization of bulk polymers............................................... 52 4.3. Conclusions and outlook................................................................... 55 5. MODIFIED BULK POLYMERS TARGETING THE C-TERMINUS OF Aß..... 56 5.1. Introduction..................................................................................... 56 5.2. Results and discussion....................................................................... 57 5.2.1. Pore-filling with fumed silica and porous DVB beads...................... 57 5.2.2. Characterization of solid support bulk polymers............................ 58 5.2.3. Bulk polymerization with modified composition............................. 63 5.2.4. Characterization of original and modified bulk polymers............... 65 5.3. Conclusions and outlook................................................................... 69 6. POROUS SILICA GRAFTED MIPS TARGETING THE C-TERMINUS OF Aß.................................................................................... 71 6.1. Introduction..................................................................................... 71 6.2. Results and discussion....................................................................... 72 6.2.1. Production and characterization of RAFT-immobilized silica............ 72 6.2.2. RAFT polymerization on modified silica surface using the grafting-from approach........................................................................ 74 6.2.3. Surface hydrophilization of grafted polymers............................... 76 6.2.4. Characterization of the core-shell polymers.................................. 77 6.3. Conclusions and outlook................................................................... 83 7. HIERARCHICAL MIPS TARGETING THE C-TERMINUS OF Aß.............. 85 7.1. Introduction..................................................................................... 85 7.2. Results and discussion....................................................................... 85 7.2.1. Template immobilization on Si-500-NH2...................................... 85 7.2.2. Template treated silica characterization....................................... 89 7.2.3. Optimization of the immobilization-to process.............................. 91 7.2.4. Hierarchical polymer synthesis.................................................... 94 7.2.5. Hierarchical polymer characterization......................................... 95 7.3. Conclusions and outlook................................................................... 98.

(9) 8. REBINDING TESTS OF ASS MIPS PART 1: MISPE AND µ-LC9�� 99 8.1. Introduction..................................................................................... 99 8.2. Results and discussion....................................................................... 99 8.2.1. Selection of templates for rebinding tests...................................... 99 8.2.2. Initial rebinding tests............................................................... 101 8.2.3. SPE method optimization and troubleshooting............................ 103 8.2.4. Elution medium considerations and finalization of the Frankenstein method................................................................. 107 8.2.5. Polymer library screening using the Frankenstein method............. 110 8.2.6. Bulk polymer screening using the Frankenstein method................ 112 8.2.7. Porous silica grafted polymer screening using the Frankenstein method ............................................................................................ 114 8.2.8. Hierarchical polymer screening using the Frankenstein method..... 115 8.2.9. Silica core-shell grafted nanoparticle screening using the Frankenstein method.................................................................... 117 8.3. Conclusions and outlook................................................................. 118 9. REBINDING TESTS OF Aß MIPS PART 2: BATCH PROCEDURE.. ........ 120 9.1. Introduction................................................................................... 120 9.2. Results and discussion..................................................................... 120 9.2.1. Buffer composition matrix......................................................... 120 9.2.2. Kinetic testing of the HEPES-ACN system.................................... 122 9.2.3. Changing buffer: kinetic and reproducibility testing of the GuHCl system.......................................................................... 125 9.2.4. Polymer library screening using the GuHCl batch method............ 126 9.2.5. Comparison of the Frankenstein- and the GuHCl batch method..... 129 9.2.6. Porous silica grafted polymer screening using the GuHCl batch method................................................................................... 132 9.2.7. Hierarchical polymer screening using the GuHCl batch method.... 133 9.2.8. Modified bulk polymer screening using the GuHCl batch method.. 135 9.3. Conclusions and outlook................................................................. 140 10. FINAL CONCLUSIONS AND OUTLOOK.................................... 143 APPENDIX 1: EXPERIMENTAL PROCEDURES......................................... 1 EP1. Synthesis of 1-(3,5-Bis-trifluoromethyl-phenyl)-3-(4-vinyl-phenyl)-urea.......... 1 EP2. Bulk polymer synthesis........................................................................ 2 EP3. Bulk polymer synthesis using fumed silica as porogen............................. 2 EP4. Bulk polymer synthesis using DVB beads as support............................... 4 EP5. Optimized bulk polymer synthesis for original and modified composition.. 4.

(10) EP6. Immobilization of the RAFT-agent on Si-500-NH2................................... 6 EP7. Grafting from Si-500-RAFT core, version 1............................................ 7 EP8. Grafting from Si-500-RAFT core, version 2............................................ 8 EP9. Surface hydrophilization of grafted polymers......................................... 9 EP10. Synthesis of carboxy-functionalized silica (Si-COOH)............................ 9 EP11. End-capping of carboxy-silica (Si-COOH/Ac)...................................... 9 EP12. Activation of end-capped carboxy-silica by NHS-ester formation (Si-NHS)................................................................................................... 9 EP13. Immobilization of C-terminal epitope peptides on NHS-activated silica. 10 EP14. Succinylation of C-terminal epitope peptides...................................... 10 EP15. Immobilization of succinylated peptide on Si-500-NH2....................... 10 EP16. End-capping of the modified silica.................................................... 11 EP17. Deprotection of immobilized peptides............................................... 11 EP18. Integrated online end-capping of modified silica................................ 11 EP19. Hierarchical polymer synthesis......................................................... 12 EP20. Standard procedure for preparing 4 M GuHCl buffer......................... 13 EP21. Standard procedure for preparing GV-8 and GA-10 stock solutions..... 13 EP22. SPE rebinding test for bulk MIPs....................................................... 14 EP23. Improved SPE rebinding test for bulk MIPs......................................... 14 EP24. HPLC-UV analysis for C-terminal Aβ-epitopes (Waters method)............. 15 EP25. SPE method optimization: short elution.............................................. 15 EP26. SPE method optimization: long elution.............................................. 16 EP27. Frankenstein hybrid batch/SPE method............................................. 16 EP28. µ-LC-UV analysis for C-terminal Aβ--epitopes...................................... 17 EP29. Standard procedure for preparing 100 mM HEPES buffer stock solution.. 17 EP30. HPLC-UV analysis for C-terminal Aβ-epitopes (Agilent method)............. 17 EP31. Initial HEPES-based batch procedure................................................ 18 EP32. Optimized HEPES-based batch procedure......................................... 18 EP33. Final batch rebinding procedure...................................................... 18 APPENDIX 2: HPLC CALIBRATION CURVES.......................................... 1 Waters Alliance 2695 calibration and reproducibility.................................... 1 Dionex Ultimate 3000 calibration and reproducibility.................................... 2 Agilent 1100 calibration............................................................................ 3 APPENDIX 3: CORRELATION ANALYSIS RESULTS.................................. 1 Fumed silica polymer carvon content and gravimetric yield............................. 1 Grafted polymer conversion and gravimetric yield......................................... 1 Grafted polymer conversion and initiator/chain transfer agent ratio................ 2 Grafted polymer conversion and total weight loss (TGA)................................ 3.

(11) Hierarchical polymer gravimetric yield and carbon content............................. 4 APPENDIX 4: SACRIFICIAL TEMPLATE SYNTHESIS OPTIMIZATION........... 1 APPENDIX 5: RESOURCES................................................................ 1 List of figures............................................................................................. 1 List of tables.............................................................................................. 9 References.............................................................................................. 12.

(12)

(13) TABLE OF ABBREVIATIONS. Abbreviation ABDV Ac2O ACN AD AIBN APP ATR-IR AȾ, AB CSF CT DCM DIEA DMF DMSO DVB EAMA.HCl FT-IR GA-10 GV-8 HEPES HPLC IF. Explanation Azo-bis-dimethylvaleronitrile Acetic anhydride Acetonitrile Alzheimer's Disease Azo-bis isobutyronitrile Amyloid precursor protein Attenuated total reflectance infrared spectroscopy Beta-amyloid or amyloid-beta Cerebrospinal fluid Computer tomography Dichloromethane Diisopropylethylamine N,N'-dimethylformamide Dimethyl sulfoxide Divinylbenzene N-(2-aminoethyl) methacrylamide hydrochloride Fourier-transformed infrared spectroscopy C-terminal epitope of amyloid-beta, containing amino acids 33-42. C-terminal epitope of amyloid-beta, containing amino acids 33-40. 2-(4-(2-hydroxyethyl)piperazin-1-yl)ethanesulfonic acid High performance liquid chromatography Imprinting factor I.

(14) Abbreviation LOD LOQ MAA MCI MeOH MIP MISPE MRI MS NHS NIP NMR PyBOP RAFT RAFT polymerization RP-LC RSD RT or r.t. SCX SPE TBA-OH TEA TFA TGA THF UPLC UV μ-LC. II. Explanation Limit of detection Limit of quantification Methacrylic acid Mild cognitive impairment Methanol Molecularly imprinted polymer Molecularly imprinted polymer-assisted solid phase extraction Magnetic resonance imaging Mass spectrometry N-hydroxysuccinimide Non-imprinted polymer Nuclear magnetic resonance Benzotriazol-1-yloxytri(pyrrolidino)-phosphonium hexafluorophosphate 4-cyanopentanoic acid dithiobenzoate, a RAFT polymerization agent Reversible addition-fragmentation chain transfer polymerization Reversed phase liquid chromatography Relative standard deviation Room temperature Strong cation exchanger Solid phase extraction Tetrabutylammonium-hydroxide Triethylamine Trifluoroacetic acid Thermogravimetric analysis Tetrahydrofuran Ultra high performance liquid chromatography Ultraviolet wavelength detection Micro-scale ultra high performance liquid chromatography.

(15) ACKNOWLEDGEMENTS. First and foremost, I would like to express my sincere gratitude to my supervisors. Prof. Thomas Arnebrant for following my progress throughout the program, his insight and guidance, his calm and rational leadership that solved many issues, and for always being there when really needed. Assoc. Prof. Marité Cárdenas Gómez for her exceptionally thorough review of my thesis and meaningful comments, for her kind personality, and that short conversation, without which this thesis may never have been conceived. Sometimes, the tiniest things mean the most. I would also like to thank my former supervisors. Prof Börje Sellergren for the opportunity to work in his laboratory, his molecular imprinting knowledge he has always been enthusiastic to share, and for reviewing my thesis. Dr. Sudhirkumar Shinde for his practical advice in the laboratory, for reviewing my thesis, and for the good discussions we had about science, business, Indian culture, and a plethora of other topics actually. I am immensely grateful to my immediate colleagues for their friendship and support both within and outside the working environment. Celina Wierzbicka for helping my first steps in Sweden, the PEPMIP network and the group, and because I could ask her about synthesis and HPLC anytime. Sing Yee Yeung for her help with the NMR and IR measurements, knowledge of organic chemistry, and for being ever cheerful and positive, no matter what. I am indebted to all members of the Marie Curie ITN “PEPMIP” for their regular review and support of my work, and in particular to Dr. Roberto Boi for his friendship and immense knowledge of biochemistry and proteomics, and his constant support of the project. Without his continuous feedback, my work would have been a blind man’s rambling. I am also immensely grateful to Prabal Subedi for being a good friend and hosting my secondment at the Ruhr University of Bochum. I thank to Mariana Duarte and Kishore Kumar Jagadeesan for III.

(16) hosting my other two secondments, and Cecilia Rossetti for visiting our laboratory and teaching me some advanced SPE techniques. I would also like to express my gratitude to all coworkers and students at the department for creating a nice, welcoming atmosphere and for the time spent well together. My sincere thanks go in particular to Colleagues: Elena González Arribas, Selma Maric, Tautgirdas Ruzgas, Peter Falkman, Anna Holmberg, Eva Nilsson, Svetlana Ivanova, Anna Runnsjö, Cathrine Albér, Zahra El-Scich, Monzer el-Dakkak, Vedran Boskovic, Henrik Steen, Vida Krikstolaityte, Olga Aleksejeva, Mona Mohamed, Dmitry Pankratov Students: Vivek Chaturvedi, Axel Rüter, Peter Erzog, Madhuri Anu, Adrien Créancier, Astrid Quirici, Pablo Cornejo Planelles, Sriram Thoppe Rajendran, Klaudia Ciurkot, Sarah Nocchi, Kathrin Hering I thank for the support of our external collaborators, in particular the hosts of my secondments: Dr. Katalin Barkovits and Prof. Katrin Marcus (RUB), Dr. Ecevit Yilmaz (MIPTechnologies AB), and Simon Ekström (Lund University). Special thanks are due to Victoria Rydengård, head of the Medeon Incubator for the numerous business and entrepreneurship-related educational programs she organized, expanding my knowledge greatly in fields other than science. I am also thankful to Brigitte Müller (Johannes Gutenberg University of Mainz) for carrying out the elemental analysis measurements featured in the thesis. The financial support of the European Commission through the Marie Curie ITN “PEPMIP” under the Framework Program FP7 “People” and Malmö Högskola is gratefully acknowledged. I would like to express my thanks to my former supervisors and teachers from the Eötvös Loránd University of Budapest, Hungary for guiding and inspiring me throughout my university studies and beyond. Prof. Béla Iván, supervisor to my MSc thesis, Prof. József Rábai, supervisor to my BSc thesis and Prof. Éva Kiss, teacher of numerous colloid chemistry and materials science courses. I am utterly grateful to my friends for their moral support throughout the research, especially to Ádám Sinai and László Sajtos, who – being PhD students themselves – could readily understand all aspects thereof. IV.

(17) The most sincere gratitude of all is due to my family for having raised me and their unconditional love, and my partner Viktória Toldi, with whom we have been together for almost five years now, and who has always been the first to support me in the less cheerful moments of the work. Thank you for everything, this thesis, and many other of my achievements would never have come true without you.. V.

(18)

(19) 1. INTRODUCTION AND OBJECTIVES. 1.1. Abstract Recent advances in the field of proteomics and biomarkers have created the opportunity to see the onset and progress of Alzheimer’s Disease from a new angle. Several peptides have been identified as possible factors in the formation of the disease, probably the most studied and important being amyloid-beta (AȾ), a family of closely related peptides of slightly different chain length (isoforms). Monitoring changes in their distribution may lead to an inexpensive and early diagnosis compared to current methods, in addition to aiding efforts towards the development of advanced treatments and an eventual cure more productive. AȾ in itself poses a serious diagnostic challenge, as it is a hydrophobic peptide with extreme susceptibility to aggregation. Strongly denaturing media have been proposed to counter this and keep the peptides in their monomeric state, but this would also render all antibodies currently used to determine amyloid levels useless. The idea behind the work presented in this thesis is to develop molecularly imprinted polymers (also called plastic antibodies) for the selective recognition and enrichment of fully denatured, monomeric AȾ, thus negating the need for antibodies in this area. The thesis describes a relatively early stage of the work, where different polymer formats with various physical properties were synthesized, characterized, and tested for rebinding affinity using short, specific parts of the peptide chain called 1.

(20) epitopes. We have in particular shown three formats and working examples thereof, together with physical characterization results. Crushed bulk monoliths constitute the most basic type of imprinted polymers, and their synthesis and physical characterization using microscopy, elemental analysis, and infrared spectroscopy was an important pivotal study before experimenting on the two more advanced formats. The tests revealed important details about the prepolymerization complex formation process, but failed to show a clear rebinding selectivity. One such advanced format was a grafted, thin polymer film on a porous silica carrier, which was prepared using a controlled living polymerization technique (RAFT). In addition to the abovementioned characterization methods, the presence of the polymer film was confirmed using thermogravimetric analysis. Here, elemental and thermogravimetric results exhibited a very good correlation, and although it was hard to evaluate some rebinding tests due to low polymerization conversions, they showed promising results in some cases. The other advanced format utilized a peptide template immobilized on the surface of a porous silica particle, followed by filling of the pores with monomer mixture, polymerization, and subsequent etching of the silica carrier. These are the so-called hierarchical polymers. Tests have shown that the polymer synthesis was successful. These advances are complemented by the successful development of two rebinding test methods (solid phase extraction and a static equilibrium test), and systematic optimization work done on several synthetic steps of the sacrificial carrier used for the hierarchical format. Based on rebinding results, the best materials of each format have been chosen for further testing. It is our hope that future investigators of the field will benefit from the results presented herein, and the first steps towards a novel diagnostic method have been taken.. 1.2. Populärvetenskaplig sammanfattning De senaste framstegen avseende proteomik och biomarkörer har medfört att uppkomsten och utvecklingen av Alzheimers sjukdom kan betraktas från ett nytt perspektiv. Ett flertal peptider har identifierats som möjliga faktorer under sjukdomens etableringsfas, av vilka amyloid-beta (AȾ), med flera s.k. isoformer,. 2.

(21) ses som den viktigaste och är den mest studerade. Genom att övervaka förändringar av isoformfördelningen skulle man kunna få fram en billig och snabb diagnostisk metod, jämfört med nuvarande lösningar, och samtidigt utforma avancerade behandlingar med siktet inställt på ett effektivt botemedel. Att analysera peptiden AȾ utgör en diagnostisk utmaning i sig, på grund av dess hydrofoba natur och benägenhet att aggregera. För att motverka aggregationen och bibehålla peptiden i monomer form används kraftfullt denaturerande lösningar. Detta omöjliggör användningen av antikroppar för detektion, eftersom antikroppar också är proteiner (polypeptider) och den för detektion nödvändiga tredimensionella strukturen kan inte bevaras i den denaturerande lösningen. Idén bakom arbetet som presenteras i den här avhandlingen är att utveckla syntetiska antikroppar, även kallade antikroppar av plast, baserat på s.k. molekylavtrycksteknologi där avtrycket av målmolekylen skapas i en (organisk) polymer. Syntetiska, polymerbaserade antikroppar är okänsliga för denaturerande miljöer och skulle därför kunna ersätta proteinbaserade antikroppar i sådana miljöer. Denna avhandling beskriver de relativt tidiga stegen i framtagningsprocessen, där olika polymerformat med varierade fysikaliska och kemiska egenskaper har syntetiserats, karaktäriserats, och testats med avseende på selektivitet och igenkänning. Under hela arbetet har korta, specifika delar av den ursprungliga peptidkedjan, s.k. epitoper, använts för att skapa de molekylära avtrycken. Vi har visat på användbarheten av tre olika polymerformat, med fungerande exempel, och med omfattande karaktärisering. Polymerpartiklar framställda från massiva s.k. monoliter är den mest grundläggande typen av molekyläravtryckspolymerer. Syntes och karaktärisering av sådana polymerpartiklar, som innebar mikroskopi, elementaranalys, och infrarödspektrometri, gav viktiga inblickar i tillvägagångssätt och lämpliga analysförfaranden inför experimenten med de två mer avancerade formaten. Ett sådant avancerat format är en tunn polymerfilm, ympad på en porös silikabaserad bärare, som har syntetiseras via en kontrollerad, s.k. levande polymerisationsteknik (RAFT). Förutom de redan nämnda teknikerna för karaktärisering, fastställdes förekomsten av polymerfilmen med termogravimetrisk analys.. 3.

(22) Det andra avancerade formatet, s.k. hierarkiska polymerer, bygger på att peptidmålmolekylen immobiliseras på ytan av en porös silikabaserad bärare, följt av porfyllning, polymerisation och bortetsning av bäraren. I avhandlingen redovisas utvecklingen av två metoder för fastställandet av de olika materialens selektivitet och affinitet med avseende på målmolekylerna, samt systematiskt optimeringsarbete avseende flera syntetiska steg vid syntes av hierarkiska polymerer. Vi hoppas att framtida forskare skall ha nytta av de redovisade resultaten och konstaterar att det första steget mot en helt ny diagnosmetod har tagits.. 1.3. Summary Alzheimer’s Disease (AD), the most prevalent and debilitating type of dementia is one of the top research topics nowadays. Affecting an ever-growing number of elderly people, mostly those above 65, currently in the range of tens of millions in the developed countries alone, and causing billions of dollars of economic loss worldwide each year and placing a huge infrastructural burden on the healthcare system beyond that. As of now, there is no effective cure for AD – even existing treatments are dubious in effectivity and the most medical science can do is to make life less miserable for patients and their families until their unavoidable death. There are several diagnostic methods for AD today, including computer tomography (CT), magnetic resonance imaging (MRI) and various psychological and memory tests. There are two major problems with all of them, though. First and most importantly, they are only able to yield a sufficiently accurate diagnosis when the first stage (mild cognitive impairment, MCI) has already presented itself. Second, all these methods require large and expensive instruments and/or numerous man-hours. The concept of biomarkers is not new at all, it is enough to think of blood sugar or similar compounds the level of which is routinely measured in order to determine whether the person is healthy or not. Recent advances in molecular biology, biochemistry, and analytical chemistry, however, opened up a wide array of possibilities with regard to biological molecules normally found in trace amounts. These novel biomarkers are of particular interest because they are supposedly specific to certain conditions – while changes in blood sugar can 4.

(23) hint to a number of diseases, a trace biomarker is supposed to be directly related to a specific disease, or its absence. The number of confirmed biomarkers is low as of now, but there are many strong candidates. One of them is a group of peptide isoforms called AȾ, which is suspected to be directly involved in the development of the disease itself, thus being much more than a mere indicator. As the theory asserting the role of AȾ as the cause of AD suggests, the levels of certain AȾ isoforms decrease below the normal range long before the first symptoms manifest – this would mean extra months or even years for physicians to attempt to develop a cure and employ different treatment methods to slow the progress of the disease and extend the active and productive time left for the patient. A prime challenge regarding the use of AȾ as a diagnostic biomarker stems from its hydrophobic nature. In simple words, amyloid peptides stick to practically everything at least moderately hydrophobic they contract – be it another peptide, lipid, small organic molecule or even the wall of their container. This means a considerable masking effect, which makes the already extremely difficult task of trace analysis even more complicated and even less reliable. In principle, strongly denaturing – chaotropic – agents, like highly concentrated urea, guanidine, or formic acid solutions could break down these aggregates and provide a fully denatured, monomeric form of AȾ peptides. However, as peptide biomarkers are usually enriched and detected by antibodies, which are also proteins, such harsh treatment would irreversibly destroy them as well. A possible solution to resolve this conflict is provided by a relatively new field of polymer chemistry called molecular imprinting. This approach aims at creating polymeric materials featuring a system of appropriately oriented moieties and interaction sites that are able to undergo specific binding interactions with a given target or other, closely related structures. As we will see in Chapter 2, there is a rich literature of these so-called molecularly imprinted polymers (MIPs) utilizing both primary (e.g. covalent bond) and secondary (e.g. hydrogen bonds) interactions. MIPs are less susceptible to environmental changes than biological antibodies are; they provide an opportunity for binding molecules under denaturing conditions, for example. Although several reports exist about successful imprinting of large molecules including proteins, the approach is still more researched, proven, and reliable. 5.

(24) with smaller units. Therefore, we decided to take the epitope imprinting approach, which intends to mimic the site-specific molecular recognition mechanism of biological antibodies. This involves imprinting only smaller, specific parts, namely the C-terminal hexapeptides of the two most interesting AȾ isoforms, AȾ1-40 and AȾ1-42. The aim of the present thesis is to provide an account on our efforts to develop a successfully imprinted polymer system for the C-terminus of AȾ, together with physical characterization and developing effective template rebinding tests. Different polymer formats and testing methods were developed and applied in order to achieve this, each discussed in their own chapters. The exact topic is covered by but a handful of publications (one journal article and standalone chapters from two PhD theses), but the wider context is well researched, both from the medical-biochemical and molecular imprinting side. Chapter 2 provides a summary of the most important topics, advances, and publications around the research subject, while Chapter 3 describes the materials and main characterization techniques that have been used in the course of this work. As the experimental work involved both synthetic and analytical tasks, the practical chapters follow this distinction. Chapter 4 describes the initial synthetic efforts aiming at the simplest form of MIPs; bulk monoliths. It introduces a number of basic concepts, experimental techniques, conclusions, and open questions that provided the foundation for further work. Chapter 5 elaborates on the bulk format, introducing a number of changes and improvements, including experiments with different porogen systems and monomer compositions. It also reveals some extremely important facts about the imprinting process. Chapter 6 presents entirely novel work on polymer-grafted macroporous silica beads. This instance of surface imprinting consists of a large number of experiments, and features a number of further revelations about the imprinting and recognition process. Chapter 7 is probably the most synthesis-heavy, describing the preparation of peptide-immobilized silica templates and the optimization process of some of these reactions and the making of hierarchical MIPs based on them. The polymers are formed using modified silica beads as support during their formation, and the silica is disposed of by chemical etching afterwards.. 6.

(25) Two analytical chapters are presented in rough chronological order. Chapter 8 provides detailed account on the initial development efforts towards a reliable MIP-assisted solid-phase extraction method, where polymer particles act as the solid phase. After the method was developed, a full screening of the polymer library available at the time commenced. Chapter 9 deals with the other main approach called the batch rebinding test. This method focuses on achieving complete mass transfer between the polymer and the loading solution by mixing and shaking until no more significant change in binding ratios is observed. In order to determine the optimal equilibration time and circumstances, both combinatorial loading phase screening and kinetic tests were conducted. Testing of newer polymers made after the work described in Chapter 8 and a comparison with the solid-phase extraction analysis complemented this method development. For the sake of easier reading, all detailed experimental procedures and recipes are presented separately, at the end of the thesis (Appendix 1), together with some interesting but less crucial results acquired during the analytical method development work (Appendix 2). Different experimental procedures will be referenced as “EPX” throughout the text.. 1.4. Research questions The main questions we expect to find answer to by the end of the thesis are the following: - Is it practically feasible to achieve an isoform-specific recognition of AȾ epitopes using molecular imprinting? - What deeper understanding of the imprinting and rebinding processes can be drawn from experimenting with different formats and testing methods? - Which systems, if any are particularly worthy of further investigation? We will return to the big picture in Chapter 10, containing an overarching discussion, overall conclusions, and outlook for the future. For now, though let us get immersed into the details!. 7.

(26) 2. STATE OF THE ART AND LITERATURE BACKGROUND. 2.1. Alzheimer’s Disease and its diagnosis Alzheimer’s Disease (AD) is the most common form of senile dementia and the most researched nowadays [1-3]. It receives the majority of all dementia research funding, in total comparable to diabetes or autoimmune diseases and only one step behind cancer [4]. Affecting 6% of the population over 65, over 30 million people worldwide and being responsible for half million deaths annually, it is certainly a threat to be reckoned with. To make the picture even worse, these numbers are probably much higher due to false negatives resulting from less efficient healthcare system and poorer instrumentation in developing countries, and are most likely increasing each year. It is also one of the most expensive medical conditions in the developed world, with an estimated annual cost of $600 billion [5]. Patients in later stages require constant supervision and their agony may last for many years. The disease is always fatal. Although the vast majority of Alzheimer’s cases are observed in patients over the age of 65, roughly 5 percent of the afflicted are much younger. This earlyonset AD usually begins in the early fifties, but in certain rare cases can appear in young adults [6]. The underlying cause of Alzheimer’s is believed to be largely of genetic nature, accounting for more than 70% of the cases [7]. Head injuries, depression, hypertension, diabetes, and smoking have also been cited as possible risk factors [2]. This also means that avoiding obesity, maintaining an intensive social life and regular physical and mental exercise are probably effective in reducing the risk of Alzheimer or at least postponing its onset. Furthermore, there is limited evidence that healthy diet – Mediterranean and Japanese food, flavonoid-rich beverages like red wine or tea, etc. can also contribute to a reduced risk of AD due to their antioxidant content. There are studies pointing at beneficial effects of various drugs, but these have not yet become subject to scientific consensus [8, 9]. 8.

(27) As pointed out before, there is no known cure for Alzheimer’s. There are, however, many attempts to treat the symptoms and keep them at bay for as long as possible to extend patients’ active lifetime, including the period while they are still able to take care of themselves or even work. Similar to prevention, symptom management also relies on medication and lifestyle improvement, together with psychotherapy. There are multiple drugs under testing, including acetylcholinesterase inhibitors and NMDA receptor antagonists, but no medication has exhibited any undisputable proof so far to delay or halt disease progression despite their FDA approval – on the contrary, such psychoactive drugs tend to have considerable side effects, which can even exacerbate the situation in one way or another [8, 9]. Psychological therapeutic methods aim at correcting the different cognitive deficiencies caused by the disease. Memory training, cognitive development therapy and reality orientation are three of the most important treatments. Probably the most effective way to hinder the progression of AD is to implement a variety of beneficial changes in lifestyle [10]. Mental and physical exercise, social activities with both people and pets, and practicing different arts, like painting or playing music have all proven to yield a minor, but clear improvement. The relative inefficiency and unreliability of all clinical methods can also be related to the fact that Alzheimer’s is usually not recognized before it gets serious. There are two reasons for that. First, as AD mostly affects old people and begins with very mild symptoms, it is very hard to spot the difference between earlystage Alzheimer’s and the normal effects of aging. By the time it becomes apparent that the patient’s mental deterioration is faster and more severe than it should be, it is already too late [1-3]. The second reason is that medical science still lacks effective early diagnostic methods for Alzheimer’s. The most widespread psychological and cognitive tests are but a step more effective than the opinion of someone who knows the patient well. Imaging methods, like MRI or CT are also only able to detect changes in the brain structure – something that does not occur at the early stages. On top of that, many patients do not even undergo such examination, as both imaging tools are costly and heavily occupied by e.g cancer patients where they can indeed give an effective, early diagnosis. Most ironically, undisputable diagnosis can only be set up after the patient’s ultimate demise: cerebral histopathology is the sole method that leaves. 9.

(28) no place for doubt. Extensive brain shrinkage is one landmark of AD pathology Fig. 1) (F. Fig. 1. Comparison of healthy and shrunken, AD-afflicted brains (left), image taken from Wikimedia Commons [11]. Comparison of healthy nerve cells and neurons in AD with amyloid deposits (right), image courtesy of the Alzheimer Association [12].. At the cellular level, there are two further pathologies that are commonly associated with Alzheimer’s.: amyloid plaques and neurofibrillary tangles. Of these two, the plaques also shown in the figure is the more interesting for us.. 2.2. An important biomarker: AȾ Amyloid plaques, commonly found in brains of AD patients, are deposits of aggregated AȾ peptides and are, together with neurofibrillary tangles made up of another protein, pTau (a hyperphosphorylated form of a neuron-stabilizing peptide) the two main microscopic landmarks of the disease [1-3]. There is a transmembrane protein, the amyloid precursor protein (APP), which is known to be involved in multiple biological processes, including synapse formation, neuronal transport and iron export, although its functionality is not yet fully understood [13]. It also undergoes extensive post-translational modifications and cleavage steps, regulated by secretase enzymes [14-16].. Fig. 2. The process of the transmembrane protein APP’s (left) cleavage to yield AȾ (middle) and subsequent plaque formation (right), image taken from Wikimedia Commons [17].. 10.

(29) AȾ is formed when Ⱦ- and ɀ-secretase subsequently cleave the APP chain, withFig. 2). It has multiple isoforms, but the most comout ə-secretase intervening (F mon form is AȾ40, with AȾ42 being the second [18]. Amyloid peptides are intrinsically unstructured, having no stable tertiary structure and therefore prone to misfolding and becoming toxic [19, 20]. In these terms, the absolute worst isoform is AȾ42: it has by far the highest affinity to misfold and form aggregates. Indeed, amyloid plaques are mostly made up by this isoform, despite it accounting for less than 1/10 of the total AȾ mass [20, 21]. There are several competing hypotheses about the onset of Alzheimer’s [22], encompassing cholinergic processes [23], oxidative stress [24, 25], T-cell intrusion [26], and even viral infections [27] as possible triggers. The currently most accepted theory, however, is the so-called amyloid cascade hypothesis, which incorporates the earlier amyloid- [14, 15, 20, 21, 28-31] and tau-models [7, 32, 33]. The cascade hypothesis proposes that AȾ and the plaques are not just an indicator of AD, but the very cause. This has been confirmed by animal studies involving long-term, targeted injections of AȾ [34]. As shown in Fig. 3, it all begins with the oligomerization of AȾ, followed by fibrillation and plaque formation [20, 21, 35, 36]. Light scattering and IR studies have also shown that this fibrillation process is a delicate equilibrium between monomeric, oligomeric and fibril forms, which has a concentration dependence somewhat resembling that of micelles insofar as it has a critical concentration, above which the aggregation rate does not increase anymore [37, 38]. The equilibrium theory is further supported by another study, showing that monomeric AȾ42 can reduce the amount of plaques by reversing the equilibrium [39]. Neurodegeneration was not affected – plaques seem not to be the only cause of AD. Besides cerebrospinal fluid (CSF), red blood cells also contain a significant deposit of AȾ peptides, their concentration increasing with age. Both the oxidative stress and the amyloid hypotheses take this into consideration, although to a different extent [25].. 11.

(30) Fig. 3. Schematic representation of the amyloid cascade hypothesis of Alzheimer’s Disease, showing the roles of both AȾ and pTau, image taken from U.S. National Institute of Health via Web Books Publishing [40].. There is substantial evidence that it is not monomeric AȾ that really causes Alzheimer’s, but soluble oligomeric forms trigger a neurotoxic effect by specifically binding to certain neuroreceptors, in a way similar to prions [14, 41-45]. It was also shown that amyloid isoforms are able to amass in the endoplasmatic reticulum of neurons and cause damage there and in cell membrane during secretion [14, 46]. Pathways leading to mitochondrial damage were also reported [14, 29, 47] and oligomers have been subject to clinical investigations for possible diagnostic use [48]. Based on these similarities, several groups have conducted studies to uncover further similarities and differences between amyloid peptides and prions [14, 41-43]. Both have been shown to decrease synaptic plasticity and cause learning disabilities and memory loss. The two proteins are in fact quite 12.

(31) prone to bind to each other, a feature even used as part of an amyloid-biosensor [5]. What is interesting though, is that while prions are infectious and can cause fatal brain disease through being transmitted to conspecifics or even to a different species (like kuru or CJB), AȾ oligomers do not have the same capability [44]. Considering all the scientific evidence gathered so far, FDA gave a semiofficial approval to the use of AȾ as AD biomarker and thus validated the entire research field [49]. As mentioned above, the majority of amyloid plaques are made of AȾ42. This means that free AȾ42 concentrations in Alzheimer patients’ blood and CSF are bound to decrease, while AȾ40 levels remain largely intact [50-52]. It is also worth noting that AȾ, like other trace biomarkers tends to show a huge patientto-patient and even time-to-time variation. The level of AȾ peptides in bodily fluids is not very indicative in itself, but AȾ40/AȾ42 ratio has diagnostic potential, as shown by Pannee et al [53]. This is a very important discovery, considering that AȾ40/AȾ42 ratio starts to change as soon as plaque formation is initiated – believed to happen up to years before actual symptoms start to manifest. Other studies have also yielded similar discoveries regarding isoform ratios [16, 18, 54, 55]. Plaques take some time to cause neuroinflammation, which in turn causes some of the Tau proteins responsible for protection and integrity of neural microtubules to hyperphosphorylate [32, 33]. These abnormally modified peptides then lose their functionality and fall apart, causing tangles of pTau to appear, also marked by the increase in free pTau levels in bodily fluids. Plaques and tangles together cause further inflammation and then neurodegeneration, leading to mild cognitive impairment (MCI) and eventually full-fledged Alzheimer’s dementia. Seeing the considerable diagnostic potential in monitoring AȾ40/AȾ42 ratios, one would be right to ask the question: why is this not a commercial product yet? The answer lies in the huge challenges associated with AȾ quantification. The aggregation and fibrillation process that undergoes in an Alzheimer patient’s brain happens in vitro, too: these peptides, especially AȾ42 exhibit such degree of stickiness that even the container’s wall can act as a precipitation surface. It is therefore extremely complicated to make a stable solution of AȾ and one can usually expect a considerable masking effect. This is further exacerbated by low biological concentrations (in the pM range in CSF, even lower in blood serum) and excessive matrix effect of other biomolecules, hindering efforts towards the development of effective purification methods have been made 13.

(32) [56]. Biomarkers are most commonly quantified either by antibody-based immunoassays [48, 55, 57, 58], mass spectrometry [16, 59-61], or combinations thereof [62-65]. The latter two suffer from the effects of aggregation in addition to AȾ peptides already being poor MS flyers. Alternative methods have also been proposed, like Mai et al.’s two-step process consisting of capturing AȾ with antibodies immobilized on magnetic micro beads, followed by fluorescent detection [66]. The typical limit of detection for commercially available assays is around 100 pg/mL [67], while the experimental methods cited above may lower the detection limit by almost an order of magnitude, to 30-50 pg/mL. Although these limits are comparable to biological concentrations, the available methods are still affected by the inability to detect oligomeric AȾ species or those bound to other molecules. One possible way of overcoming these limitations would be treating the samples with strongly denaturing (chaotropic) agents, such as 8 M urea, 4 M guanidium hydrochloride [58], 80 w/w% formic acid etc. Unfortunately, these agents work by irreversibly breaking up all secondary and tertiary protein structures – making it is impossible to denature AȾ peptides exclusively. Putting all information together, the actual challenge is to find an intermediate sample preparation step, where aggregates are broken up and matrix contaminants largely removed, together with most of the chaotropic agent. In addition to that, as removal of the denaturing chemicals means further folding and aggregation, the sample has to be ready for final analysis after a few hours at the latesr. Briefly, a robust material is needed with specific sensitivity towards AȾ, but relative insensitivity to environmental effects, including temperature, pH, medium, etc.. 2.3. The art of plastic antibodies: molecular imprinting Before reviewing the past and present of molecular imprinting, perhaps the best course of action is to briefly introduce the main concept as it is perceived today. The principle could be described based on a host/guest mechanism [68-70] insofar as a theoretically perfect molecularly imprinted structure has cavities (ie. the host) that accommodate their respective templates (i.e. the guest) perfectly,. 14.

(33) while rejecting all other molecules that do not have the same moieties, shape, size, etc. as the template.. Fig. 4. The basic principle of molecular imprinting: template-monomer complex formation, crosslinking, template removal and selective rebinding of the intended target molecule, image courtesy of Biotage AB [71].. As seen in Fig. 4, the molecule to be imprinted (template) is first mixed with appropriate monomers. In this context, ‘appropriate’ means monomers with moieties that can undergo specific interactions with the template, e.g. H-bonding in a mostly apolar matrix or Ɏ-stacking where all or most of the contaminants are not aromatic. Any interaction might prove useful: first-generation imprinted polymers tended to take advantage of covalent bonds (a type of chemisorption) [68], but nowadays non-covalent interactions are preferred [69, 72], together with an increasing application of organometallic compounds and their coordinative properties [73]. In reality of course, it is not always possible to recognize a template by just one interaction. In these cases, several strong interactions have to be picked and the monomers carrying them positioned accordingly. Since imprinters usually utilize interactions that are preferred among all possibilities in the given system, a self-assembly takes place upon mixing, forming the so-called prepolymerization complex. The next step is to fix these interactions in place, creating a permanent binding site. This is realized by a select amount of crosslinker. In this step, one has to be careful: too little crosslinker and the network will not be able to keep its shape. On the other hand, too much 15.

(34) crosslinker may cause the 3D structure become rigid, making it impossible for the template to migrate in and out of the system. This reasoning also assumes that the network in question is porous, in order to ensure binding site accessibility. The extent of imprinting effect and ultimately the effectivity of a MIP is usually characterized by the imprinting factor (IF). The IF is defined as the ratio of the retention factor of the template on a column packed with MIP to an identically shaped and sized column filled with non-imprinted polymer (NIP) that only differs from its respective MIP insofar as no template has been used during its synthesis. Alternatively, the ratio of template bound by MIP to NIP under identical conditions can also be called IF. The aforementioned concept of molecular recognition is almost a century old, but researchers worked only with silica-based materials for several decades. The first observation of specific binding occurred on an inorganic polymer of sodium-silicate prepared in the presence of ammonium-carbonate followed by a subsequent treatment with different simple aromatic solvents, like toluene by Polyakov’s group [74]. The whole process was overly long (taking up to a full month) and went essentially unnoticed by the scientific community. Starting from the late 1940’s, further efforts with silica-based materials followed, including the work of Linus Pauling’s student, Dickey on methylated azo-dyes [75]. It is very interesting to have a look at the procedure described in this paper, as it contains most elements of today’s imprinting methodology outlined above. As in Polyakov’s case, aqueous sodium silicate was used as monomer, but instead of forming the silica gel first, Dickey mixed the silicate with the dye right away and let the two polymerize together in aqueous acetic acid, before drying the mixture. The particles were then ground, sieved and the dye removed by methanol. The main steps of prepolymerization complexation, 3D structure forming, template removal, and selective recognition can all be caught already in this old publication. Moreover, it featured a control substance prepared exactly the same way, but in the absence of the template (called non-imprinted polymer, NIP nowadays). The theoretical explanation given by Dickey and Pauling is also worth keeping in mind while reading this chapter: “The adsorbent is thus pictured as automatically forming pockets that fit closely enough to the foreign molecule to hold it by van der Waals' forces, hydrogen bonds, interionic attractions, and other types of intermolecular interaction.” Obviously, there is a long way from capturing simple azo dyes with silica gel to even attempting the selective recognition of a complicated template like AȾ, but the discovery of the phenomenon as such happened remarkably early. The following two decades saw 16.

(35) many publications about materials capable of selective recognition, although only silica-based systems were published [68]. Molecular imprinting, as we know it today, dawned during the early 70’s, pioneered by two independent, yet concurrent publications. Wulff prepared polymeric materials for racemate resolution [76], which was one of the principal topics of the preceding decades in selective recognition science. At the same time, Takagishi and Klotz addressed the other challenges and prepared materials for the separation of structural analogues, reaching back to the group of azo dyes [77]. These were the first two accounts of using organic materials for molecular recognition, and similarly to every other molecular imprinting research in the following two decades, featured covalent bonds. Silica-based imprints also survived, albeit with ever-decreasing importance. This still does not mean though, that silica imprinting is dead – Cheng et al. have just recently reported successful extraction of the infamous bisphenol A using entirely silica-based materials [78]. Though early molecularly imprinted polymers exclusively used covalent bonds, the idea of utilizing other interactions have been present since the beginning. A pioneering work towards the realization of that was done by Sellergren and Andersson, who reacted an L-tyrosine derivative with methacryloyl-chloride and crosslinked with DVB [79]. Upon cleavage, the resulting cavity was successfully used to bind an L-phenylalanine derivative non-covalently. This approach is called semi-covalent imprinting. The first report of fully non-covalent imprinting was released by Vlatakis et al. [80] belonging to the group of Klaus Mosbach. Their work aimed at making drug assays cheaper and more robust by employing imprinted polymers, using e.g. theophylline and diazepam as templates. More than two decades have since passed and the number of articles published annually in the field has been steadily increasing, as seen in Fig. 5. This huge number of publications covers a great variety of topics from literally every part of the chemical sciences where selective recognition might have a place. Fortunately, a number of excellent reviews are available to make the challenge of following the literature easier. In 1999, Cormack and Mosbach revised the progress since the pivotal study seven years earlier [81]. This was followed by a thorough review by Alexander et al. [68], encompassing well over a thousand articles from the imprinting literature up until 2003. Largely the same group of researchers, this time led by Whitcombe wrote the sequel, spanning 17.

(36) from 2004 to 2011, providing an even more extensive digest of virtually all areas of molecular imprinting, containing over 3800 references [69].. Fig. 5. Cumulative number of articles in the field of molecular imprinting up to current day [82]. The darker column indicates the year 1993, when the first noncovalent imprinting procedure was published by Vlatakis et al. [80].. Looking at Fig. 5., we can conclude that these two articles alone cover ca. 5200 out of 7200 articles, i.e. about 75% of the entire imprinting literature published until 2011. This is supplemented and partially overlapped by a number of other reviews: Sellergren et al.’s summary about the perspectives of MIPs acting as advanced drug carriers [83] and Yan et al.’s mini-review of the actual questions and challenges [84]. Sellergren and Hall published a book chapter about molecular imprinting in 2012 [70], in the same year as Bowen et al. published a critical evaluation of the antibody-like properties of MIPs [85] followed by Vasapollo et al. [72]. In addition to these general collections of MIP literature, several thematic reviews are also available, for example an extensive description of coordination chemistry and organometallic substances in molecular imprinting [73] from 2006. The latest years saw especially many publications, including Cheong et al.’s work on MIPs used as separation materials [86], Huang et al. on wastewater treatment applications [87], Ding et al. on surface imprinting techniques [88], Abdohalli et al. on different uses of RAFT polymerization for imprinting [89], and Figuiredo et al. on contaminants in personal care products 18.

(37) [90]. There are also some reviews on more novel and complicated subjects, like Sharma et al. on macromolecular and bacterial imprinting [91], or Poma et al. [92] and Lv et al. on different nanosized imprint formats [93]. The present thesis will not cover the literature in that detail. Instead, the main imprinted polymer formats and imprinting techniques will be described and illustrated by a few interesting and/or important examples to provide adequate understanding of the subject before moving on to the actual description of the research done. Molecular imprinting has seen limited industrial application so far, and even those are limited to smaller molecules. The French company Polyintell, for example sells sample preparation (SPE) consumables for a broad range of target molecules such as mycotoxins (Patulin, Zearelonone, Ochratoxin A, Fumonisins), drugs (amphetamines), endocrine disrupting compounds in food matrices (cereals, milk, coffee, wine etc.) such as Bisphenol A, Estrogens (Estradiol), and more [94]. The Swedish company Biotage and its subsidiary, MIPTechnologies are also focused on SPE accessories, currently available for specific analysis of chloramphenicol, tobacco specific nitrosamine derivatives, NNAL, triazines, clenbuterol, ß-agonists and ß-blockers in addition to offering custom MIP engineering services for various industries, such as pharma, food, chemical, and environmental clean-up [95]. The largest company offering MIPs in its assortment is definitely Sigma-Aldrich with its SupelMIP® product line [96]. They offer MIP SPE solutions for a number of drugs (NSAIDs, Beta-agonists and –blockers, chloramphenicol, etc.) and environmental pollutants (PAHs, Bisphenol A, etc.). Despite their theoretical and practical versatility, MIPs only take a minor share of the adsorbent market, applying products based on scientific advances dating back 15-20 years.. 2.4. Molecular imprinting formats and techniques Molecular imprinting recognizes and utilizes a multitude of different material formats. It has been mentioned above that binding sites in molecularly imprinted polymers must have good accessibility in order to show specific recognition. This can be realized in two ways: either some sort of support or a small amount of solvent (porogen) is used to generate pores in the material for the template to freely migrate in and out, or the polymeric particles are designed in such a way that the majority of the binding sites are to be found at the surface.. 19.

(38) The first approach i.e. making the material porous is technically trivial and was realized early on – essentially the works of Polyakov, Dickey and other early pioneers are all about carrying out the polymerization in a relatively concentrated solution. Hence, once polymerization is initiated and the first polymer chains and crosslinks are formed at the nucleation sites, the solvent is displaced, and since the system is concentrated, a monolith rather than separate particles are formed. However, the solvent cannot escape, and a system of pores of differing shape and size is formed. This monolithic material is called a bulk polymer and the process that generates it bulk imprinting by MIP scientists. It is worth noting, though that this name is technically incorrect as the process outlined above is in fact a highly concentrated solvent polymerization. Still, in order to follow the terminology of the field and for the sake of convenience, the present work will use the term “bulk” to describe this format and process (a generalized procedure is outlined in Fig. 6). No actual bulk polymerization (i.e. in the absence of any solvent) is discussed in this thesis.. Fig. 6. General procedure for the preparation of bulk imprinted polymers. Functional monomers are primarily responsible for forming the binding sites, while the other type of monomer, a crosslinker provides physical stability to the polymer. The porogen (here: solvent) is responsible for forming the pores that make the interior of the polymer network approachable. Image courtesy of Karsten Haupt, Compiegne, France [97].. 20.

(39) Coming back to the pores, their formation is not completely random. The amount and nature of the porogen greatly affects the polymer morphology, which in turn has a considerable effect on rebinding affinity and capacity. This is a fairly well researched topic, being of interest since at least the early 1990’s. Sellergren and Shea investigated the effect of porogen, polymerization temperature and certain polymer modifications on the morphology and binding of Lphenylalanine imprints [98]. They found that the largest influence on binding amount and selectivity was exerted by the hydrogen bonding capacity of the porogen and the polymerization temperature. The morphology itself, characterized by pore volume, swelling, and surface area, seemed to be less important. Schmidt et al. studied the effect of the porogen in a MIP system for S-propranolol, a very common model compound for molecular imprinting studies [99]. In addition to using an aromatic (toluene) and an aliphatic (diglyme) porogen, they also investigated the effect of a linear polymer, poly(vinyl-acetate) (PVAc) as a porogen additive in order to facilitate phase separation and hasten pore formation. They found that while the solvent’s chemical nature and amount influenced both rebinding capacity and selectivity (with diglyme working much better), the polymeric additive only decreased capacity, while leaving selectivity intact. They also prepared a set of thin films for the same system, where all binding sites were on the surface (therefore no need for pores) and being essentially non-porous, these films were much less affected by PVAc. In both cases, a mildly polar, non-aromatic porogen has proven to return the best imprinting results. Haginaka et al. imprinted D-chlorpheniramine using acrylate monomers and a range of porogens comprising toluene, phenylacetonitrile, benzylacetonitrile and chloroform and a multi-step swelling process prior to polymerization [100]. The materials prepared with aromatic porogens proved to be porous, while to become useful, chloroform-derived polymers required swelling in acetonitrile. When tested as HPLC column filling against the target, all exhibited comparable selectivity. This again points to the direction that morphology in itself is not necessarily indicative of rebinding capabilities. Song et al. imprinted a naturally occurring flavonol, quercetin using an acrylamide functional monomer and carried out an extensive optimization of solvent polarity (dielectric constant) and volume based on molecular modeling simulations [101]. Solvent quality is of particular importance in non-covalent imprinting because more polar solvents exert a stronger competitive effect by participating in van der Waals, dipoledipole and hydrogen-bonding interactions, making favorable prepolymerization 21.

(40) complex formation less likely. Medium polar media (e.g. THF) proved to be the best also taking solubility issues into account. Solvent volume vs. binding capacity also showed a clear maximum, indicating that pore size above the optimal is just as harmful to the template recognition process ass binding site unavailability caused by too small and/or too few pores. Working on nicotinamide, Wu et al. started with molecular modeling, and used methacrylic acid as the functional monomer. They found that when the porogens had poor hydrogen bonding capacity, the interaction energy was mainly influenced by the dielectric constant of the solvent. On the other hand, in case of a porogen having strong H-bonding capacity the interaction energy of the forming template–monomer complex was affected by both the dielectric constant of the solvent and the hydrogen bonding interference. Using an aprotic porogen with a low dielectric constant (i.e. a mildly polar solvent) realized a large interaction energy between the template and the functional monomer, giving the MIP optimal binding properties [102]. Liang et al’s simulation and experimental study on ursodeoxycholic acid, using a combination of liquid and solid porogens (the latter being sodium sulfate or calcium carbonate) further supports these arguments, although in this case toluene was proven to be the best porogen both in simulation and in practice, due to Ɏ-stacking with the template [103]. From a technical perspective, bulk MIPs are the simplest of all to prepare and therefore still see widespread use as the first step of any monomer composition screening, being a relatively inexpensive yet robust format providing indicative results. Besides its obvious advantages, the bulk format has a number of negative aspects. First, crushed monolith fragments are unevenly distributed in shape and size. This causes local differences in morphology with possible consequences on rebinding behavior and yields are normally much lower than 100% due to losses resulting from the tedious crushing/sieving procedure. Second, the binding sites are also randomly distributed and the actual number of useful sites tends to remain well below the theoretical number. Fig. 7 shows a simplified model of typical binding site types in a MIP.. 22.

(41) Fig. 7. Main types of template binding sites in a MIP. Well-defined, accessible, but not too exposed binding sites of type B (and to some extent A) are preferred over the other, non-specific sites. Image taken from Sellergren and Hall [70].. The most useful sites are of type B (in micropores), followed by type A (in macropores). The rest of the sites are mostly unspecific or inaccessible, meaning that they do not contribute to the selective rebinding capabilities of the MIP. Due to the uncontrolled growth and relatively low surface area of bulk MIPs, only a minority of the sites belong to groups B and A. During the last three decades, several ideas, techniques, and formats were proposed to overcome these limitations. Reaching back to the previous discussion about the role of porogens, an increasing amount of porogen will make the pores larger, whilst decreasing monomer concentrations will cause the growing chains and nucleation sites to be further away from each other. If the monomer concentration is too low, the growing nuclei will not connect to aggregate into a monolith, and form standalone, spherical particles instead. The first to report precipitation imprinting surfaced was Sellergren [104], including the imprinting and chromatographic evaluation of a number of templates, including e.g. L-phenylalanine, atrazine, and the 23.

(42) AIDS drug pentamidine. Polymers were prepared both in situ, inside the chromatographic column, and in separate vessels and packed in columns later. Another important, early work was carried out by Ye et al. [105]. They imprinted various hormones and alkaloids, including theophylline, caffeine, and 17a-estradiol. Although the targets themselves are typical of the field, theirs was the first account that did not involve monolith formation, crushing, sieving, and all the disadvantages thereof. Ye’s continued efforts resulted in another article a year later, further discussing the formation of precipitation microparticles and microgels [106]. They also experimented with functional monomer: crosslinker (MAA:TRIM) ratios and found that optimal morphology and imprinting effect could only be achieved between specific limits, in this case between 1:1 and 1:2.4. Yoshimatsu et al. went one step further and in their study on precipitation imprinting of (S)-propranolol found that tuning the binary crosslinker system (DVB-TRIM in this case) can actually be used to fine tune the particle size [107]. Mean particle diameters ranged between 130 nm and 2.4 μm, in consequence of the increase of the DVB concentration at the expense of TRIM, while the binding affinity remained largely intact. Yields, however, decreased for smaller particles. The effect of crosslinker density is further explained and confirmed in a recent study by Golker et al. [108]. Their system targeting bupivacaine was planned based on molecular dynamics simulation results, ensuring a theoretically perfect prepolymerization complex formation. However, they found that even if the complex forms properly, an insufficient crosslinking density can still result in poor imprinting properties, caused by decreasing surface area and pore volume, and by improper stabilization of the binding cavities. Some practical applications of precipitation MIPs include analytical materials for enrichment of drugs from urine [109], capture of a fungicide, thiabendazole from fruits and fruit products [110], and polystyrene-silica composite nano beads for the recognition of amino acid derivatives [111]. Other industrially important polymerization techniques have also been adapted to certain imprinting needs, including emulsion [112] and suspension [113] processes. As these techniques both involve the use of water, which would have an undesirable competitive effect on the H-bonds in the prepolymerization complex, they are of no interest for the present research and were not investigated in detail.. 24.

(43) In the recent years, more specific formats have been developed. Core-shell structures have been a long-time interest for imprinting scientists [88, 110]. By forming a thin film on micro- or nanoparticle surface using either the grafting-to or the grafting-from approach, both material requirements and the number of inadequate binding sites can be decreased. In order to avoid random polymerization, various controlled, quasi-living polymerization techniques, like ATRP [114] for RNAse enzyme imprinting, reversible chain transfer polymerization [115] for S-propranolol, or iniferter polymerization [116] for thiabendazole were applied. Our group’s involvement and efforts have been focused on an advanced variety of the iniferter approach, reversible addition-fragmentation transfer polymerization (RAFT) [117]. The technique controls polymerization rates and chain lengths by a chain transfer agent undergoing a reversible reaction with the growing polymer chain (Fig. 8).. Fig. 8. The general mechanism of RAFT polymerization. The key concept of the mechanism is to attain a more uniform chain length distribution by creating a stable chain transfer agent – propagating chain complex in a reversible process. This slows down the chain growth and lowers the polydispersity. Image taken from Wikimedia Commons [118].. 25.

(44) The most important characteristic of this system is that the polymer-RAFT species is much more stable than standalone growing chains and the formation of this complex is faster than chain growth. This prevents chains from growing quickly and as only a small number of chains are able to react with further monomer molecules at the same time – an only for a very short period at once – the overall rate of propagation is decreased. RAFT polymerization also features even chain growth, manifested in low polydispersity (PDI). The controlled nature of RAFT polymerization is of particular interest to imprinting chemists, as it can effectively contribute to better binding site formation and morphology. Examples include Xu et al.’s report [119], clearly showing the difference between imprints made by the RAFT technique and traditional freeradical polymerization (FRP). Triazine was used as template, and the RAFTmicrospheres exhibited almost double rebinding capacity and affinity. A recent review contains even more examples [89]. Our group has also been active in the area of RAFT MIPs [120, 121]. On the other hand, some sources, like Asman et al. suggest that RAFT polymerization does not necessarily have a beneficial effect on rebinding properties and a careful choice of monomers and chain transfer agents is needed to make any difference compared to FRP [122]. Another widely spread surface imprinting format, hierarchical imprinting also utilizes porous silica as support, but instead of retaining it as a core, the beads are used as sacrificial templates – much like a negative in photography. Template molecules are immobilized on the surface, followed by filling of the silica pores with polymerization mixture and the polymerization itself. The silica frame is then etched out, leaving a negative image of the pore system behind. The first known use of the technique was reported by Sellergren and Büchel in 1999, when they patented the general process [123]. Shortly thereafter, Yilmaz et al. immobilized 8-carboxypropyltheophylline on aminopropyl-modified porous silica (APS), end-capped the remaining amino groups using acetic anhydride and then filled the pores of the silica support with a trifluoromethylmethacrylic acid (TFMAA) – DVB mixture, followed by etching of the silica with aqueous HF [124]. The polymers thus obtained were characterized and tested for rebinding. Two years later, Titirici et al. [125] immobilized an adenine derivative on APS and subjected it to an acrylate-based monomer mixture, followed by subsequent polymerization and etching of the silica support. Another year later, the same group realized the imprint of different amino acids (Phe and Gly), both protected and unprotected, and a Fmoc-Phe-Gly-silica dipeptide 26.

No results found