CHAPTER 1: Nano-sized Molecularly Imprinted Polymers as Artificial Antibodies

Molecularly imprinted polymers (MIPs) emerged about 40 years ago, when Wulff and Sahan proposed the strategy of polymerisation performed in the presence of a target-template.¹  Since then, MIPs have been exploited for multiple applications thanks to their remarkable properties, such as high affinity and selectivity, and resistance to extremes of temperature, pressure and pH variations. In molecular imprinting, functional and cross-linking monomers are polymerised in an appropriate porogenic solvent in the presence of the compounds to be imprinted (called “templates”). After removal of the template, the polymer matrix retains recognition cavities that are complementary to this template in terms of size, shape and functionality. More precisely, the three-dimensional arrangement of cavity recognizing sites is driven by the structure of the template molecule. Therefore, after template removal the synthesized polymer is capable of binding the target, which can be the analyte itself. Compared with synthesis of monoclonal antibodies (mAbs), the synthesis of MIPs is simpler and cheaper, and can be performed without any preclinical development involving animals. In addition, MIPs show high stability and excellent mechanical properties. Moreover, they can be prepared for a wide variety of targets.^2,3  Template removal, however, is often difficult and incomplete, with the possible disadvantage of subsequent analyte leaching from the matrix, thus resulting in inaccurate performance in analytical applications.⁴  Moreover, it is quite laborious to integrate them with signal transducers in sensors (unless electropolymerisation is involved, see Section 1.3.1.2, below) or to convert the template binding into an electric signal.⁵  In general, MIPs can be manufactured in different ways (Table 1.1) and in several formats, for example as films or membranes, microparticles or nanoparticles. Compared to other formats, the nanoparticles present several advantages. In particular, this format allows the system to exhibit a much higher surface-to-volume ratio and larger total surface area per weight unit of polymer. The imprinted cavities are more easily accessible by the analytes, thus improving binding kinetics and template removal and, hence, enhancing their recognition capabilities.⁶  Several authors have started developing nanoMIPs for diagnostic and therapeutic applications, for instance as drug delivery systems^7,8  and sensing elements in assays or sensors.^9,10  In virtue of their features, nanoMIPs represent an attractive option for a wide range of applications. One interesting characteristic of nanoMIPs is their property of remaining in solution, rendering them suitable for in vitro studies.¹¹  However, it is crucial to obtain particle batches with a very narrow size distribution and with high yield, especially for biomedical studies. Two other great advantages of MIPs, compared to natural ligands, are their relatively straightforward preparation and, particularly, their inexpensive fabrication. Indeed, the availability of cheap reagents for MIP syntheses has led researchers to explore novel polymerisation approaches for devising smaller and monodispersed MIPs, from the most intuitive techniques (i.e. precipitation polymerisation) to more sophisticated ones (i.e. the use of solid-phase polymerisation and automated reactors).

Molecular imprinting involves three main steps; i.e. (i) the formation of the monomer–template complexes, (ii) polymerisation, and (iii) removal of the template and collection of the MIPs. In general, MIPs can be fabricated by means of two main approaches: covalent and non-covalent imprinting. In the former, developed by Wulff,¹²  reversible chemical bonds are created between the monomer and the template during the polymerisation, and the same bonds are then re-formed in the analyte binding step. The advantage is that only the monomer’s functional groups interact with the template and cavities with more homogeneous recognizing sites are generated. However, not many compounds are suitable for this approach to be used and, therefore, they need preliminary derivatization with the monomer. Furthermore, the template removal is quite difficult and the analyte binding step is slower compared to that of other approaches.⁵  In the non-covalent approach pioneered by Mosbach and co-workers,¹³  hydrogen bonding as well as electrostatic and hydrophobic interactions are involved in the formation of the monomer–template pre-polymerisation complexes, as well as in the following analyte recognition. Since weak interactions are involved, an excess of monomer is usually employed to stabilise the monomer–template complex. This method is easier and more versatile than the covalent approach, although issues related to heterogeneity of the binding sites within cavities generated might arise.¹⁴  Considering the advantages of the aforementioned two approaches, some authors have combined them, thus using a template covalently linked to the monomer and the following analyte binding step designed in a non-covalent way, thus introducing the concept of semi-covalent imprinting.¹⁴  In this section, we will explore the main polymerisation modalities optimised so far for the synthesis of MIP nanoparticles (nanoMIPs), by focusing on (i) precipitation polymerisation, (ii) mini-/micropolymerisation and core–shell polymerisation (Scheme 1.1), (iii) atom transfer radical polymerisation (ATRP) and reversible addition–fragmentation chain transfer polymerisation (RAFT), as well as (iv) solid-phase polymerisation without and with the use of automated reactors, with particular emphasis on pros and cons of each approach (Scheme 1.2).

Among all the techniques used to synthesise nanoMIPs, certainly precipitation polymerisation is the fastest and most straightforward strategy for obtaining monodispersed nanoparticles with high yield. Typically, this approach is based on a free-radical polymerisation occurring in highly diluted mixtures of monomers. The “growing” pre-polymer keeps enriching it with monomers, until it precipitates when its size renders the polymer insoluble. In this manner, spherical and uniform nanoparticles are rapidly produced with high yield and low consumption of reagents. However, the main drawback of this method lies in its strength point. In fact, the diluted solution of monomers may decrease the interaction between the active monomers and the template molecule. This decrease leads to a less efficient formation of the monomer-template pre-polymerisation complex and, therefore, it decreases the selectivity of the resulting nanoMIP. Moreover, the increase of the concentration of monomers leads to larger and less uniform nanoMIPs.¹⁵  In the early 2010s, polymerisation in concentrated monomer solutions was performed.^16,17  The idea was to stop the polymerisation before the gelation point by abruptly diluting the solution of pre-polymers; subsequently, a poor solvent for the polymer was added for nucleation of the nanoMIPs.^6,7,18  Usually, precipitation polymerisation is performed using organic solvents. Nevertheless, it is also possible to precipitate nanoMIPs from aqueous solutions if a surfactant is added (at a low concentration) to the monomer mixture. Obviously, after synthesis it is necessary to remove not only the template but also the surfactant. As discussed above, several parameters play a crucial role in this polymerisation, thus affecting quality of the resulting nanoMIPs. For instance, the ratio of different reagents, the type and concentration of the reagents (particularly the template) and the solvent can affect the resulting nanoMIP and reaction yield.¹⁹  In conclusion, precipitation polymerisation is a cost-effective, time saving, and quick procedure for fabrication of nano- and microsized MIPs. This polymerisation allows the synthesis of uniform spherical nanoMIPs with high yield and the possibility to desirably tune their parameters.

As mentioned in Section 1.2.1 above, the polymerisation mixture can be optimised and enriched with different chemical species, e.g. surfactants. Notably, emulsion polymerisation involves organic solvent solutions of template-conjugated functional monomers and organic cross-linking monomers. The solution is then emulsified in an aqueous solution of surfactants and stabilisers (e.g. sodium dodecyl sulphate, SDS), and then the solution is usually stirred and sonicated. Generally, emulsion polymerisation can be classified as mini- and micro-emulsion polymerisation. The mini-emulsion approach exploits co-surfactant or surfactant monomers and, eventually, a stabiliser is added to produce nanoparticles with homogeneous size.²⁰  The peculiarity of mini-emulsion polymerisation lies in the synthesis of semi-covalently imprinted nanoparticles, i.e. the formation of the pre-polymerisation complex of monomers with the template engages covalent bonds while the target binding depends upon non-covalent interactions. NanoMIPs fabricated using this procedure show high affinity and selectivity for the target. On the other hand, however, micro-emulsion polymerisation generally occurs under more complicated conditions, namely, in the presence of water, oil, and one or more surfactants.²¹  Moreover, spherical nanoMIPs are obtained using different procedures. A clever approach involves the addition of monomers or peptides conjugated to fatty acid chains to increase the confinement of the template to the surface of the growing nanoMIPs.²²  The main drawback of this technique is related to the presence of SDS and other chemicals. In fact, the presence of surfactant(s) and stabilisers in solution might affect monomer–target recognition and, therefore, it requires several washing steps, which may compromise the yield of the reaction as well as the efficacy and homogeneity of the nanoMIPs. Moreover, the application of these MIPs is limited to in vitro studies because it is impossible to completely remove those chemicals, a crucial step for any in vivo application. Therefore, nanoMIPs produced via emulsion polymerisation are not recommended for biological applications.¹⁶  Nevertheless, mini- and microemulsions allow easy and rapid synthesizing of uniform nanoMIPs exhibiting remarkable recognition properties. Moreover, emulsion polymerisation, together with grafting polymerisation, can be exploited for assembling core–shell nanoMIPs. The most common “core materials” are polymeric structures, silica, or magnetic nanoparticles (i.e. Fe₃O₄ nanoparticles).²³  As discussed above, multiple advantages of emulsion polymerisation render this synthetic approach particularly suitable for devising core–shell nanoparticles. There are a few interesting strategies, which have been widely employed during the past years. In particular, the fabrication of core–shell nanoparticles is composed of two steps, i.e. (i) formation of the core and (ii) creation of the MIP shell by emulsion polymerisation. In all cases, from the use of latex cores to magnetic ones, the diameter of core–shell nanoMIPs was lower than 100 nm, thus opening up the gates to several different applications.^24,25  Generally, semi-covalent imprinting is the favourite approach, but also non-covalent imprinting has been exploited and was successful in terms of affinity and selectivity of the composite nanoparticles.²⁶  Despite the drawbacks listed above, the emulsion polymerisation still remains the most suitable approach for devising core–shell nanoMIPs, thanks to its straightforwardness and high yield.

Over the past decades, heat- and UV light-initiated polymerisations have been exploited for the synthesis of MIPs. Nevertheless, the desire to continuously optimise the polymerisation protocols has led to the development of new approaches, such as microwave-assisted synthesis initiation and controlled living radical polymerisation procedures (i.e. ATRP and RAFT).^27,28  In particular, these novel polymerisation strategies have been largely employed for either the fabrication of MIP films or core–shell nanoMIPs. Indeed, the big advantage of living radical polymerisation, compared to conventional free-radical polymerisation, lies in the control of the thickness of the MIP film, which is a crucial requirement for devising composite nanoparticles.^10,11  Moreover, the ATRP method exploits metal ion catalysts (usually copper ions) to coordinate the interactions operative in the formation of the monomer–template conjugate and the overall polymerisation reaction.²⁹  The ATRP mechanism is based on the equilibrium between radical and inactive species. Ions of metals, such as copper, iron, and molybdenum, mediate the entire catalytic process. When these metal ions are in their low oxidation state, they react intermittently with the inactive species. Therefore, “dormant” species act as both activators promoting the formation of the growing polymer and as deactivators, stopping the activity of the radicals, hence re-stabilising the equilibrium between the species in solution.³⁰  The equilibrium of the reaction and the metal ions employed are the strength point of this method. Moreover, they allow for control of the polymerisation itself. However, one drawback of ATRP is related to the presence of these metal ions. The metal ion catalyst might affect monomer–template recognition and it must be removed after the end of the polymerisation. Therefore, several and sophisticated purification steps are necessary (e.g. ion-exchange resins must be employed). With further optimisation of the polymerisation parameters, the amount of copper can be decreased, which is required for biomedical applications.

The RAFT method is a good candidate for controlled free-radical polymerisation, considering its versatility and simplicity. As with ATRP, the possibility to control the thickness of the polymer renders RAFT a suitable option for the fabrication of core–shell nanoMIPs, via the use of iniferter molecules or, more specially, RAFT agents (e.g. thiocarbonylthio compounds). With this method, uniformly distributed polymers with complex architectures can be synthesised (from linear block co-polymers to dendrimer-like polymers). While in the case of ATRP a crucial role is played by the metal ion catalyst, in RAFT polymerisation the RAFT agent is the principal effector. Once the reaction is initiated and the growing polymer is extending, this active species reacts with the RAFT agent to form the “RAFT adduct radical”. The RAFT adduct radical is the most reactive species and, therefore, it is capable of triggering the formation of other growing polymers. In the case of core–shell nanoMIPs, the RAFT agent is immobilised onto the surface of the core particles (e.g. polystyrene, silica, Fe₃O₄).³¹  Both grafting and precipitation polymerisation based on RAFT results in formation of uniform MIP shells with high affinity, selectivity, and with the possibility to tune the thickness of the grafted polymer. Although ATRP and RAFT are relatively new methods, they may be interesting starting points for further optimised polymerisation strategies thanks to the capability to control the thickness and architecture of the growing and final polymer.

An option for conventional polymerisations in solution is solid-phase polymerisation. While other methods exploit free templates in solution, which requires several washing steps to remove the unreacted species (e.g. dialysis), solid-phase polymerisation overcomes this disadvantage by covalently immobilising the template molecule onto micrometre-size beads (Scheme 1.2). Typically, these beads are made of glass and, therefore, they need to be silanised with chemicals bearing suitable functional groups for further linking with the template. The template can be immobilised onto the glass beads using different strategies, depending on the characteristics of the template molecule itself (i.e. functional groups). For instance, generally, –SH group bearing templates are immobilised onto succinimidyl iodoacetate pre-functionalised beads, while –NH₂ and –COOH group bearing molecules can be immobilised via the EDC/NHS reaction. Once the solid phase is opportunely derivatised with the template of interest, the polymerisation proceeds as per other polymerisations in solution. Presently used imprinting techniques suffer from heterogeneous “polyclonal” distribution of binding sites, poor performance in water, low capacity, and slow mass transfer. A particular difficulty arises in the case of water-soluble targets.⁴ 

The use of a solid-phase polymerisation approach overcomes these difficulties, because the immobilised template enables an oriented immobilisation, thus decreasing the “polyclonality” of the imprinted sites. Furthermore, due to the effect of the solid phase, the binding sites are formed on the surface of the nanoparticles. This is in contrast to typical protocols with the template present in solution where most of the imprinted sites are formed within the polymer, thus leading to lower accessibility for the target molecule. The great advantage of solid-phase polymerisation lies in its independence from dialysis steps. Indeed, low affinity nanoMIPs and other components of the polymerisation mixture are first washed away at reaction temperature, whereas only high-affinity nanoMIPs are retained. The latter are subsequently eluted by increasing the temperature of the elution solvent. NanoMIPs produced by this solid-phase polymerisation approach have shown high affinity and selectivity for the target molecule.^32,33  The increasing number of applications, covering in vitro assays and in vivo studies as well as the interest in producing nanoMIPs in a large scale has led MIP experts to design automated systems. Thanks to their antibody mimicking features, nanoMIPs have attracted the attention of big companies and, therefore, the development of operator-independent procedures is now necessary. The great advantage of a polymerisation within an automated reactor lies in the total removal of human error and lifting operator fatigue, together with an increased yield of the reaction (and the possibility to scale up the protocol) and improvement of the batch-to-batch repeatability. The first automated reactor for nanoMIPs was proposed by Piletsky and co-workers.³⁴  They employed an automated reactor characterised by a column with a filter on the bottom with porosity of tens of micrometres. Template-functionalised micrometre-size glass beads were packed in the column placed inside the reactor. Both the polymerisation mixture and initiator were injected automatically into the column. After the polymerisation, several washing steps were performed to remove the non-reacted species and low affinity nanoMIPs, and then the most affine nanoparticles were eluted, all in an automated manner by using the reactor software.³⁵  Moreover, the use of iniferter in this automated reactor allowed for functionalisation of the nanoMIPs in a controlled fashion.³⁶  In this way, it is possible not only to obtain nanoMIPs with homogeneous distribution of binding site affinities (typically one recognition cavity per nanoparticle), but also functionalised nanoMIPs (e.g. with fluorophores or polyethylene glycol, PEG, layer to decrease the agglomeration tendency of the nanoparticles). In conclusion, nanoMIPs produced in an automated fashion by exploiting reactor polymerisation show high recognition capability, which, together with the high yield of the reaction, is making the use of this method increasingly common, especially for large scale applications.

The use of Abs and enzymes in diagnostic assays is widely accepted and, to date, represents the gold standard in terms of sensitivity and affinity. Abs are routinely used in many diagnostic assays but, unfortunately, they suffer from short shelf-life, high costs of manufacturing and relatively low stability, especially in organic solvents and at extreme temperature and pH values. Furthermore, it is not easy to produce Abs against immunosuppressants or toxins because of their action on the immune response.³  Abs also suffer from complex manufacturing procedures, lack of oral bioavailability, low cell membrane permeability and increased patient morbidity.⁴⁷  Indeed, even humanised types can elicit immunogenic reactions. Moreover, generating Abs against small molecules is not straightforward because chemical coupling to haptens is first required.⁴⁸  Finally, it is often difficult to immobilise Abs on the supports used in diagnostic assays.⁴⁹  In light of their attractive characteristics, nanoMIPs represent a viable alternative to Abs especially with regards to their stability and manufacturing costs. The fact that no animals are required for manufacturing these plastic antibodies is a step-change, which would enable companies to develop assay/sensors based on nanoMIPs at a fraction of the usual cost of Ab-based platforms. Furthermore, as discussed above, the possibility to generate nanoMIPs by solid-phase imprinting allows, for the first time, the production of surface-imprinted nanoparticles with antibody-like features (e.g. water dispersibility). Moreover, the automatic synthesis of nanoMIPs enables any party to develop their own plastic antibody, with minimal manual intervention and, therefore, training needed. At the same time, the low template contamination achievable with the solid-phase synthesis results in no need of dialysis, thus simplifying the overall MIP production. In contrast to their natural counterparts, nanoMIPs do not require any special storage conditions, being stable at room temperature for several months. This high stability makes them attractive for sensors or assays deployed in remote geographical areas, where cold chain supply might not be available. Interestingly, very recent studies demonstrated that nanoMIPs are biocompatible in several cell lines, not evoking any immunogenic response in macrophages.⁵⁰  In particular, cytokine release after exposure to nanoMIPs of different concentrations was monitored over 72 hours. The assay performed aims to flag the potential of a compound for generating an inflammatory response in vivo. Although the assay carried out is not predictive of in vivo inflammation, it can identify compounds that might lead to a potentially severe pro-inflammatory response (i.e. release of cytokines). This response might lead to the cytokine release syndrome (CRS) that causes serious systemic symptoms including fever, hypotension, and organ failure. Thanks to these properties, undoubtedly nanoMIPs hold great potential in diagnostics and in theranostics in particular. In the next paragraphs, we will review some of the most recent and innovative applications of plastic antibodies in sensors, assays, and for imaging purposes (Table 1.2).

By definition, a sensor is a device capable of detecting events or changes in the surrounding environment, and able to process the information into a corresponding output (typically electrical or optical signals). Both chemical sensors and biosensors are now considered well-established tools in several fields of research, from analytical chemistry to clinical diagnostics as well as forensics and environmental sciences. A reliable chemo- and biosensor must exhibit a design and geometry to allow the most optimised interaction of its recognition part with the transducer, which is required to possess high sensitivity for the target. Biosensors bear the disadvantages related to the use of biomolecules, such as nucleic acids and proteins (e.g. Abs), as recognition elements. Exploiting macromolecules is advantageous for many reasons, among these their availability. On the other hand, their limited stability in non-physiological environments leads to difficulties in their storage and high costs for the production of more stable structures. Based on these considerations, researchers have moved their attention to more stable recognition tools, i.e. MIPs.⁵¹  At the beginning of their development and optimisation, MIPs were mainly utilised as stationary phases for the purification of substances (i.e. chromatography). However, given their outstanding selectivity and stability, MIPs have been acquiring increasing relevance also in the field of chemosensors and bioassays. Particularly, MIP-based chemosensors,⁵²  which reveal the following integrated features, are needed: (i) a high-sensitivity transducer to monitor and process the binding event and (ii) a high-affinity and high-selectivity MIP capable of maintaining its recognition properties when implemented with the transducer. The versatility of MIP-based systems is worth highlighting. The binding of the molecule of interest can produce chemical and/or physical changes in the analyte itself or in the MIP, such as variations in fluorescence intensity or ionisation degree. Depending on the signal triggered upon binding and the variations in the properties of the analyte or the MIP, several transduction strategies can be employed, e.g. amperometry, potentiometry, conductometry (Scheme 1.3) as well as surface plasmon resonance (SPR) spectroscopy and surface enhanced Raman spectroscopy (SERS) (particularly for the detection of proteins), and surface-acoustic wave (SAW) sensing. In this section, we will analyse and discuss possible applications of MIPs in chemosensors for the detection of biomolecules, focusing on electrochemical, optical, and other MIP-based chemosensors (i.e. calorimetric, SAW, and quartz crystal microbalance, QCM).^51,53–55 

The first electrochemical sensor (ECS) was developed in the 1950s to monitor the levels of oxygen. It was composed of an anode, a micro ammeter, a cathode and a source of electrolyte. ECSs are able to transform the electrochemical activity of the target molecule, or the changes in the electrochemical properties of the system, into electrical signals.⁵³  Each electrochemical sensor is rationally designed depending on its purpose. All ECSs are characterised by their size, shape, and components (i.e. recognition element for the detection of the molecule of interest and transducer). However, as previously stated, the most crucial step in devising a chemosensor is the integration of the recognition element within the transducer. Indeed, the ligand (i.e. the MIP) must be able to maintain its recognition properties once associated to the transducer. This can be achieved, for instance, by in situ polymerisation or by surface grafting by exploiting thermal initiation or UV irradiation.⁵⁵  Electropolymerisation is one of the most preferred strategies for fabrication of MIP chemosensors. Notably, electropolymerisation bears the big advantage of controlling the deposition of polymers onto the transducer in a precise and controlled manner, by measuring the time of applied potential or current. Nevertheless, the chemosensor is designed depending on the application and the properties of the systems itself as well as the physical and chemical characteristics of the target analyte. Planning the experiment and knowing in depth the principal features of the molecule of interest and the polymer lead to the most optimised design. If the analyte shows electrochemical activity or fluorescence, amperometry or ellipsometry can be exploited for the detection of the target. On the other hand, MIPs can be prepared by polymerisation starting from a mixture of monomers having different properties, such as electroconductive monomers. Additionally, solution for polymerisation could be enriched with metal ions or nanoparticles^19,56,57  able to improve sensitivity and enhance electron transfer or to coordinate the binding and, at the same time, inducing changes in the surrounding environment, e.g. pH changes. Moreover, MIPs have been coupled to carbon nanotubes or magnetic nanoparticles to further improve the sensitivity of the detection.^58,59  The most popular MIP-based chemosensors include (i) affinity chemosensors, (ii) receptor chemosensors, and (iii) catalytic chemosensors⁵³  (Scheme 1.3). Affinity chemosensors are similar to immune-sensors; the electroactive target accumulates at the surface of the transducer thanks to the presence of high-affinity MIPs. Receptor chemosensors are those based on changes related to the MIP; the binding with the analyte triggers variations in one or more features of the polymer, resulting in the generation of a signal subsequently processed by the transducer. Finally, catalytic or enzyme-like chemosensors are able to detect changes in the close surrounding environment due to the catalytic activity of the MIP itself.⁵³  Concerning electrochemical transducers, one can distinguish three main methods for the detection of target binding: (i) conductometry, (ii) amperometry, and (iii) potentiometry.⁶⁰  Conductometry is a technique that allows detecting conductivity alterations by exploiting the application of an electric field and, therefore, the migration of ions of opposite charges. However, it is impossible to discriminate between two or more ions in this case. Amperometry is based on the evaluation of the changes once a linear relationship between the concentration of the ionic species and their current at a fixed potential is established. For instance, Xue and co-workers developed an amperometric MIP-based electrochemical sensor.⁶¹  To enhance the conductivity of the system, the MIP was doped with functional monomer-coated gold nanoparticles. The resulting chemosensor was successfully exploited for the detection of dopamine in human samples. The main disadvantage of this technique is the requirement of a porous membrane, which subsequently necessitates the removal of the target, eventually leading to false results.

On the other hand, potentiometry can solve the issue of “extraction”. Indeed, potentiometry is based on generation of a potential difference with no need for the different charged species to diffuse through any membrane. In this case, neither size exclusion due to the porosity of the membrane occurs nor are extraction steps required. Anirudhan and Alexander⁶²  designed and fabricated a potentiometric sensor exploiting the recognition properties of MIPs and the physico-chemical features of multiwalled carbon nanotubes. This sensor showed several interesting features, starting from the possibility to tune the sensing of the target molecule at different pH, its reusability and very high detection capability of the target molecule (limit of detection, LOD, 10⁻¹⁰ M) and selectivity compared to other electrochemical sensors, either in water or in organic samples. It is also possible to measure a current at the time a potential sweep is applied. The generation of this current is due to oxidation or reduction of different species.⁶⁰  In the last 5 years, capacitance/impedance chemosensors have gained increasing importance by virtue of the possibility to investigate the binding of the analyte to the MIP by detecting variations in its thickness or electric permittivity. Notably, MIP-based chemosensors exploiting capacitance or impedance have been employed for the detection of biomarkers and organism-specific proteins.⁵⁵  Together with the development of new composite materials, these typologies of chemosensors are also leading to the preparation of even more innovative and challenging platforms for protein sensing. For instance, the impedance graphene-MIP-based chemosensor designed by Luo and colleagues⁶³  showed an excellent combination of molecular imprinting with electrochemical sensing. Notably, Luo et al. devised electrochemical sensors based on the detection of bovine haemoglobin protein via impedance measurement.

In conclusion, nowadays, electrochemical sensors are versatile, easily implemented and well-established tools for the detection of charged targets. The optimisation of MIP-functionalised electrochemical systems, using an imprinted polymer as the recognition element of the sensor, has allowed for the devising of high-sensitive sensors and also decreasing production costs and increasing robustness of the sensor itself.

As discussed above, electrical read-out is extremely useful when changes in MIP or analyte properties occur upon binding and can be transduced in electrical signals. However, there are several criteria to be satisfied.⁶⁴  Conversely, the development of optical sensors is constrained to only a few restrictions, mainly related to the polymer in the case of MIP-based sensors. Optical sensors convert electromagnetic radiation (typically in the UV–NIR range) into electric signals. With the rise of new polymerisation strategies, MIPs have been employed for devising optical sensors. Thanks to the progress in the covalent coupling of MIPs to optical transducers, innovative and robust MIP-based chemosensors have been exploited for the detection of target molecules with the LOD ranging from nM to fM.¹⁹  Notably, there are several optical modalities that can be implemented to these systems: (i) fluorescence, fluorescence quenching and chemiluminescence, (ii) UV-vis and infrared spectroscopy; (iii) surface plasmon resonance (SPR) spectroscopy, and (iv) SERS⁶⁵  (Scheme 1.4). An MIP-based optical chemosensor can be designed in two main ways: (i) imprinting the polymer with fluorescent molecules, and (ii) rendering the MIP fluorescent by using fluorescent functional monomers or with the incorporation of a fluorescent agent.^19,64  In the case of the fluorescent target, the advantage lies in the straightforwardness of the strategy.^66,67  However, there is an intrinsic limit related to the availability of naturally fluorescent molecules.⁶⁸  Therefore, fluorescent MIPs appear to be more generic in nature and, therefore, suitable for this application.⁶⁹  Notably, since the 2000s several fluorescent monomers have been exploited to produce fluorescent MIPs for the detection of certain analytes in liquid samples by monitoring changes in the fluorescence intensity as well as quenching events upon binding.^69,70  Very recently, fluorescent nanoMIPs imprinted for tetracycline were used for quantification in real samples (bovine and pig serum).⁷¹  An anthracene-derivative monomer was first synthesised, and then used as signalling functionality capable of being quenched upon addition of the drug, down to sub-micromolar concentrations. Additionally, MIP-based optical chemosensors have been recently exploited for the detection of protein biomarkers, with a detection limit of 10⁻⁸ M.⁷²  Quantum dots (QDs) appear to be the most suitable fluorescent agents for the development of fluorescent MIPs. Indeed, thanks to their small size together with their remarkable chemical and physical properties, QDs have been widely employed to develop molecularly imprinted optosensing materials (MIOMs), showing LODs down to 10⁻¹⁰ M.⁷³  Additionally, the incorporation of noble metals, particularly platinum, gold, and silver ions or their nanostructures may be a clever trick to enhance fluorescence variations. In a recent review, Wackerlig and Lieberzeit¹⁹  explored the advantages and disadvantages of MIP-based sensors coupled to QDs and noble metal structures. Spectroscopy has been exploited for multiple applications, particularly in high-throughput assays as well as for studying the interaction with the analytes. Notably, (UV-vis spectroscopy)-coupled MIP-based chemosensors were used for high-throughput screening by using microtiter plates functionalised with MIPs.⁷⁴  On the other hand, infrared spectroscopy has been used to study in deep the interaction between the MIP recognition element and the target molecule upon binding. For instance, the thermal profile⁷⁵  or the changes in the frequencies of the vibration of NH bond stretching⁷⁶  before and after the polymerisation were employed.

Optical sensors based on either SPR spectroscopy or SERS appear to be even more appealing compared to other transducer modalities explored so far. They have gained increasing importance in the field of chemosensors, not only for bearing intrinsic elegance related to the measurement modality but also for the ability to detect the analyte at femtomolar concentrations.⁷⁷  SPR chemosensors typically couple a recognition element (i.e. MIPs) to either gold or silver; upon binding the target molecule, the chemosensor registers changes in the refractive index. Notably, variations in the stiffness after analyte binding leads to measurement of the amount of protein accumulated onto the chemosensor by detecting changes in the Bragg shift of the MIP. The great advantage of these chemosensors lies in the very low biofouling observed after several measurements and, therefore, the reliability of the system.

The first SPR MIP-based chemosensors were devised in the 90s. During the past years, SPR chemosensors exploiting MIPs have undergone several improvements and now they have reached LODs in the order of tens of femtomolar for the detection of small molecules,⁷⁸  whereas macromolecules have been detected in the nM range.⁷⁹  In particular, nanoMIPs increased the sensitivity of the system compared to MIP films, likely due to the higher number of recognition cavities. While SPR MIP chemosensors have been optimised since the 90s, SERS systems are relatively novel.⁸⁰  The big advantage of SERS MIP-based chemosensors is their independence from variations of the size and stiffness of the polymer. In fact, these chemosensors allow observing changes in the vibrational spectrum upon target binding. For these reasons, the sensitivity of SERS chemosensors is expected to increase with further optimisation, especially in the case of nanoMIPs. Nevertheless, SERS based chemosensors still suffer from some issues, such as the long exposure time to the sample before starting to detect changes in Raman spectra as well as the robustness of the system itself, especially related to the integration of MIPs to the transducer,⁶⁵  which may be overcome by improving the design of the sensor. Indeed, Kamra and co-workers⁸¹  designed a new approach for the devising an MIP optical chemosensor for label-free detection by exploiting SERS. The idea was to immobilise core–shell nanoMIPs onto the gold substrate, and subsequently couple the gold nanoparticles for SERS. They obtained and demonstrated the efficiency of this Au@MIP@AuNPs Raman system. Notably, the chemosensor showed reproducibility and very high recognition efficiency without the presence of a catalyst or an increase of temperature, ease of handling and most importantly robustness, that way opening the gates to further optimisation strategies.⁸¹ 

Other MIP-based systems worth mentioning are: (i) calorimetric, (ii) SAW, and (iii) QCM chemosensors. Calorimetric sensors are systems capable of detecting heat released upon binding of the analyte of interest,^51,82  whereas SAW chemosensors are acoustic systems. The latter are particularly interesting because they detect variations in the frequency of the acoustic waves upon analyte binding.⁵¹  Very recently, Tretjakov and colleagues⁸³  deposited MIPs onto the surface of a SAW chip via electropolymerization. This innovative system demonstrated binding of the target protein with high affinity, i.e. three times higher than that of the corresponding NIP sensor, while also having the great advantage of being a label-free approach. However, the most popular among these systems are those of QCM chemosensors using sensing of the mass change. QCM-based chemosensors use a thin disc of a piezoelectric material, usually “sandwiched” between two metal film electrodes to complete the transducer, and integrated with the recognition element (i.e. an MIP film). The quartz crystal is a piezoelectric material capable to oscillate mechanically when a.c. voltage is applied to the electrodes. Therefore, these chemosensors are able to detect variations in the resonance frequency of the piezoelectric material. In particular, changes in this frequency are opposite to the mass change because of the analyte accumulation in the MIP film. There are three principal applications of piezoelectric MIP-based chemosensors: (i) diagnostics, (ii) environmental, and (iii) purification.⁸⁴  Since the technology for the production of MIPs was developed initially for purification of small-molecule compounds, it is not surprising to find in the literature a large number of these chemosensors for the detection of drugs and contaminants, some even in human serum.⁸⁵  It is worth mentioning the recent work published by Naklua and co-workers,⁸⁶  who devised a piezoelectric chemosensor coupling an MIP-QCM system to a dopaminergic receptor D1R for the detection of dopamine.⁸⁶  Additionally, MIP-based piezoelectric chemosensors have been employed for the detection of macromolecular compounds, achieving a sensitivity in the pM range.⁸⁷ 

In the last decade, nanoMIPs have been successfully applied in diagnostic assays for analyte quantification. In one such example, nanoMIPs imprinted for vancomycin were used in the first ELISA-like assay, where the use of Abs was replaced by the said nanoparticles (Scheme 1.5).⁸⁸  In this assay, a horseradish peroxidase (HRP)-template conjugate was first prepared, and then employed in competitive binding experiments. That is, several solutions of free vancomycin (between 1 pM and 70 nM) were added to the wells together with the HRP conjugate. After incubation and washing, the substrate (3,3′,5,5′-tetramethylbenzidine, TMB) was added, leading to colour generation in solution due to reaction with HRP. The assay showed signal linearity from 1 pM to 70 nM and an LOD of 2.5 pM. Interestingly, this sensitivity is much higher than that of other ELISA tests reported in the literature, whose LOD was merely 0.1 nM.⁸⁹ 

In a similar example, multifunctional nanoMIPs were recently used in a novel ELISA-like assay with no biomolecules involved.⁹⁰  Imprinted NPs, produced by solid-phase synthesis and embedding an iron oxide core with catalytic properties, act simultaneously as recognition and signalling elements. In light of the intrinsic peroxidase-like activity, iron oxide can be employed in different assays.⁹¹  The iron oxide particles were first modified with carbon double bonds, and then used as “reactive seeds” for further polymerisation in the presence of a solid phase bearing vancomycin as the template. The assay was developed by initially conjugating vancomycin onto the well surface. Then, upon addition of free vancomycin and magnetic MIPs, competition occurred, in a similar manner as the aforementioned pseudo-ELISA. After addition of the TMB substrate, a blue colour was detected due to the catalytic activity of the core–shell MIPs, thus allowing detection of vancomycin in the nanomolar concentration range.

Very recently, a novel approach for synthesising nanoMIPs by using proteins as macro-functional monomers has been developed.⁹²  For proof-of-concept, HRP was chosen as a model and cross-linked using glutaraldehyde in the presence of a solid phase bearing immobilised templates (vancomycin and ampicillin). The cross-linking process in the presence of templates led to the formation of target-selective recognition cavities, without any significant adverse effects on the enzymatic activity. In contrast to complex protein engineering methods typically used to generate affinity proteins, this approach can be employed to prepare protein-based ligands in a short time. Since these affinity materials reveal both catalytic and molecular recognition properties, they are potentially useful in assays, for instance in an ELISA format where cross-linked HRP imprinted with a given template can replace traditional enzyme–antibody conjugates. However, this approach appears to work effectively only down to micromolar concentrations and, therefore, can be potentially a rapid alternative to raising Abs for targets that do not require high assay sensitivities.

Very recently, several groups have started applying the molecular imprinting technology for the recognition of selected targets on the surface of cells. For instance, fluorescent nanoMIPs for molecular imaging of cells and tissues were devised for the first time by imprinting glucuronic acid, a monosaccharide present as the terminal unit on larger oligosaccharides.⁹³  The produced nanoMIPs were then employed to image the hyaluronan on human keratinocytes and on adult skin specimens (Figure 1.1a). Interestingly, molecules of other potentially interfering compounds, such as galactose, N-acetylglucosamine, N-acetylgalactosamine and glucose did not bind to the nanoMIPs, possibly because of the lack of the charged carboxyl group. Another target recently exploited by other groups for cellular targeting is sialic acid. Sellergren’s group fabricated nanoMIPs targeting cell surface glycans, via sialic acid imprinting.⁹⁴  The overexpression of glycans terminating with sialic acid (SA) residues on the surface of cells correlated with various diseases including cancer. Two-hundred-nanometre silica nanoparticles were used as cores for the subsequent grafting of the SA-imprinted layer. The authors speculate that a certain mixture of functional monomers is ideal for the nanoMIPs to bind strongly to the template. In particular, hydroxyl groups of SA would interact with a boronate-based monomer, while the carboxyl group would be targeted by a urea-based fluorescent monomer synthesised in-house. The produced fluorescent nanoMIPs were capable of selective staining different cell lines depending on the SA expression level (Figure 1.1b). Similarly, Yin et al. prepared SA-imprinted nanoMIPs for selective imaging of cancer cells.⁹⁵  In this case, the authors employed Raman-active nanotags (silver nanoparticles) as signaling cores. By surface imprinting based on silanes, an imprinted layer based on boronate, used as the functional monomer, was then created around the silver core. Raman spectroscopy provides significant advantages, such as high photostability and sensitivity, as well as multiplexing capacity. Healthy human hepatic cells and hepatocarcinoma cells (HepG-2) were used as a model to test the selectivity of the nanoMIPs. HepG-2 cells showed an evident SERS signals at 1435 cm⁻¹, much stronger than that for healthy cells, proving the selective binding of the nanoMIPs to SA on the cell surface.

An intriguing application of nanoMIPs involves their use as a therapeutic tool for selective protein/enzyme sequestration in cells. Once the nanoMIPs capture the protein/enzyme, the latter cannot carry out its physiological function, thus altering the cellular metabolism. Very recently, this principle was introduced by imprinting silica-coated iron oxide nanoparticles with DNAase I, a cytoplasmic enzyme involved in cell apoptosis.⁹⁶  The nanoparticles were also fluorescently tagged to visualize their distribution inside the cell. As done by several other groups, a silica layer was inserted between the magnetic core and the fluorescent reporter to minimise potential quenching issues. The nanoMIPs successfully inhibited the DNAase activity without affecting the short-term cell viability. These examples prove that nanoMIPs hold great potential as molecular recognition and imaging tools since, in contrast with Abs where multistep staining processes are required, multiple labelling can be easily achieved by using a panel of nanoMIPs, each one incorporating different dye for a given target. Furthermore, it is possible to link drugs or add magnetic functionalities to the nanoMIPs, thus allowing their use as drug delivery systems or in hyperthermia therapy. Therefore, it is expected that the number and complexity of these multifunctionalities in nanoMIPs will grow in the next years. Despite the aforementioned successful applications of nanoMIPs within sensors or assays, such nanoparticles have not been widely applied for in vivo diagnosis/therapy. The first in vivo application of imprinted nanoparticles was reported by Hoshino et al.,⁹⁷  who employed nanoMIPs imprinted with melittin (a peptide that is the principal component of bee venom) to remove the said molecule from the bloodstream of living mice (Scheme 1.6). The mice were intravenously injected with melittin and, afterwards, the nanoMIPs were administered via the tail vein. The MIPs successfully cleared melittin, improving the survival rate of mice over 24 h and decreasing the melittin toxic effects (e.g. weight loss and peritoneal phlogosis). This study demonstrates the potential of the nanoMIP for selective recognition of molecules in vivo.

The only other in vivo study has been recently performed by Wu and colleagues⁹⁸  who prepared amoxicillin loaded nanoMIPs imprinted with an epitope of Lpp20, a membrane lipoprotein expressed in Helicobacter pylori. In vivo imaging demonstrated a prolonged permanence of the nanoMIPs in stomachs of H. pylori infected mice. In particular, the nanoMIPs showed a more pronounced antibacterial effect than that of free amoxicillin, after intragastric administration.

Although currently Abs and enzymes are widely employed in diagnostics, they suffer from some drawbacks, such as high manufacturing costs, relatively poor stability (especially at extreme values of temperature and pH), short shelf life at room temperature, and long lead times. The use of nanoMIPs can overcome these problems. Compared with Abs, the synthesis of MIPs is simpler and more cost-effective, and it does not involve the use of animals. In addition, MIPs show high stability and excellent mechanical properties, and they can be prepared virtually for any target. No cold chain is required and, therefore, MIPs can be applied in chemosensors or assays to be deployed in remote geographical areas where storage at room temperature might be the only option available. NanoMIPs, sometimes called ″plastic or synthetic antibodies″, are nanostructured polymer particles capable of selective recognition of the target molecule. Thanks to their size, they represent a viable alternative to Abs, as demonstrated by their recent use in several diagnostic fields. The latest successful applications of nanoMIPs in cells and in vivo demonstrate their potential in diagnostics, and several other MIP-based composite nanosystems are expected to be fabricated in the coming years. However, improvements in the synthesis of nanoMIPs are required, in particular in relation with their size distribution and large scale production. In this regard, the herein described solid-phase approach – fully automatable – can become a viable option for large scale manufacturing of nanoMIPs to be applied for chemosensor devising and assay developing. Molecular imprinting of polymers by solid-phase synthesis represents the most generic, versatile and cost-effective approach to formation of synthetic molecular receptors/binders to date with antibody-like features. Plastic antibodies represent an entirely new compound class, which can be deployed to address both extracellular protein targets and, thanks to their capability to cross the cell membrane, potentially to currently intractable intracellular proteins. In fact, both Abs and aptamers are not cell-permeable. Furthermore, depending on the imprinted template, the polymers can be used repeatedly without loss of their “memory effect”.³²  Moreover, MIPs are not proteins and, therefore, are not susceptible to proteolysis or bacterial degradation. Undoubtedly, nanoMIPs will play a crucial role in the development of novel robust assays/sensors, possibly in a home environment, where expensive and sensitive components like Abs would require extra care by the end user. Due to the robust nature of MIPs and their possibility to be reused, MIP-based systems able to work in continuous and/or under harsh conditions will soon be commercialised. Moreover, lateral flow tests and microfluidics based on nanoMIPs are likely to be deployed in the next years, potentially allowing multiplexed analysis due to the possibility to embed different types of coloured/fluorescent cores within a single nanoMIP.

Ab

Antibody

ATRP

Atom transfer radical polymerisation

CRS

Cytokine release syndrome

D1R

Dopaminergic receptor 1

DAPI

4′,6-Diamidino-2-phenylindole

DiO

3,3′-Dioctadecyloxacarbocyanine perchlorate

DU145 cells

Human prostatic carcinoma cells

ECS

Electrochemical sensor

EDC

N-(3-Dimethylaminopropyl)-N′-ethylcarbodiimide hydrochloride

ELISA

Enzyme-linked immunosorbent assay

HaCaT

Human keratinocytes

HepG-2 cells

Human liver hepatocellular carcinoma cells

HRP

Horseradish peroxidase

LOD

Limit of detection

Lpp20 membrane

Lipoprotein expressed in Helicobacter pylori

mAb

Monoclonal antibody

MIOM

Molecularly imprinted optosensing material

MIP

Molecularly imprinted polymer

nanoMIP

MIP nanoparticle

NHS

N-Hydroxysuccinimide

NIR

Near-infrared

NMP

Nitroxide-mediated polymerisation

PD

Polydispersity

PEG

Polyethylene glycol

QCM

Quartz crystal microbalance

QD

Quantum dot

RAFT

Reversible addition–fragmentation chain transfer polymerisation

SA

Sialic acid

SAW

Surface-acoustic wave

SDS

Sodium dodecyl sulphate

SERS

Spectroscopy and surface enhanced Raman spectroscopy

SPR

Surface plasmon resonance

TBM

3,3′,5,5′-Tetramethylbenzidine

UV

Ultra violet