Molecularly Imprinted Polymer Nanoparticles: An Emerging Versatile Platform for Cancer Therapy

Abstract Molecularly imprinted polymers (MIPs) are chemically synthesized affinity materials with tailor‐made binding cavities complementary to the template molecules in shape, size, and functionality. Recently, engineering MIP‐based nanomedicines to improve cancer therapy has become a rapidly growing field and future research direction. Because of the unique properties and functions of MIPs, MIP‐based nanoparticles (nanoMIPs) are not only alternatives to current nanomaterials for cancer therapy, but also hold the potential to fill gaps associated with biological ligand‐based nanomedicines, such as immunogenicity, stability, applicability, and economic viability. Here, we survey recent advances in the design and fabrication of nanoMIPs for cancer therapy and highlight their distinct features. In addition, how to use these features to achieve desired performance, including extended circulation, active targeting, controlled drug release and anti‐tumor efficacy, is discussed and summarized. We expect that this minireview will inspire more advanced studies in MIP‐based nanomedicines for cancer therapy.


Introduction
Cancer is amajor public health problem and has been one of the leading causes of death worldwide. [1] Nanomedicines based on polymers,l iposomes,i norganic particles,c arbon materials,m etallic materials,a nd their composites have emerged as promising platforms for the diagnosis and treatment of malignant tumors over the past decades,contributing to decreased mortality and prolonged lifetime. [2] Thee laborate design of nanomedicines for cancer therapy is expected to overcome as eries of biological barriers and pitfalls, improve the pharmacokinetic and pharmacodynamic profiles of conventional therapeutics,a nd optimize the therapeutic efficacyw ith minimal side effects. [3] Thea dvances in smart nanomaterials are one of the driving forces for the development of cancer nanomedicine. [4] Molecularly imprinted polymers (MIPs), also referred to as plastic antibodies or artificial antibodies,a re chemically synthesized affinity materials with tailor-made binding cavities complementary to the template molecules in shape,size, and functionality. [5] Owing to the presence of imprinted cavities,M IPs can specifically recognize and bind the template molecules.Compared to traditional bioligands such as antibodies,a ptamers,a nd lectins,M IPs feature several advantages,s uch as easy preparation, excellent stability,a nd low cost. [6] In recent years,s ignificant progress in the performance of MIPs including specificity,a ffinity,a nd targeting scope has motivated researchers to expand their traditional application fields,such as separation, sensing,and assays,t omore challenging biomedical applications,s uch as disease diagnosis,bioimaging,and cancer targeting. [7] MIP-based nanoparticles (nanoMIPs) have attracted great attention due to their easy combination of the biomimetic recognition of MIPs with the unique properties of nanoparticles (nanoscale size,h igh surface-to-volume ratio, and optical, acoustic,t hermal, magnetic, and electric properties). [8] Particularly,a sanew platform in cancer nanomedi-cine,n anoMIPs have exhibited great potential for cancer therapy. [9] The unique features of nanoMIPs bring versatile and desired functions to cancer nanomedicine.Moreover,the realization of these properties and functions is independent of bioligands. Thus,n anoMIPs provide not only an alternative to the currently existing nanomaterials for cancer therapy,b ut also hold great potential to fill the gaps in nanomedicines associated with the dependence on biological ligands,f or example immunogenicity,stability,applicability,a nd cost.
Recently,several excellent reviews on the fundamental chemical design, preparation strategies,a nd biomedical applications of nanoMIPs (e.g.b ioimaging,d isease diagnosis,a nd drug delivery) have been published. [10] MIPs as synthetic antibodies for cancer therapy have also been highlighted. [11] However,a no verview specifically focused on the recent advances in nanoMIPs for cancer therapy is apparently lacking. To fill this gap,i nt his minireview,w es urvey recent advances in the rational design of MIP-based nanomedicines for cancer therapy.Specifically, we highlight the distinct features of MIPs that have been used to rationally design and construct cancer nanomedicines with desired performance.F or cancer therapy,i ntravenously administered nanomedicines must evade the reticuloendothelial system during systemic circulation, permeate the tumor site via tumor vessel leakage,p enetrate deep into the tumor tissue,b ei nternalized by tumor cells,a nd exert at herapeutic effect on predesigned targets. [3] Along these lines,wediscuss the functions of nanoMIPs for cancer therapy in four aspects:e xtended circulation, active targeting,c ontrolled drug release,a nd antitumor efficacy.I na ddition, we further introduce some nanoMIPs with dual functions.Finally, we briefly discuss the present challenges and future perspectives of nanoMIPs for cancer therapy.
Molecularly imprinted polymers (MIPs) are chemically synthesized affinity materials with tailor-made binding cavities complementary to the template molecules in shape,s ize, and functionality.Recently, engineering MIP-based nanomedicines to improve cancer therapyhas become arapidly growing field and future researchdirection. Because of the unique properties and functions of MIPs,M IP-based nanoparticles (nanoMIPs) are not only alternatives to current nanomaterials for cancer therapy, but also hold the potential to fill gaps associated with biological ligand-based nanomedicines,such as immunogenicity,stability,applicability,and economic viability.Here, we survey recent advances in the design and fabrication of nanoMIPs for cancer therapyand highlight their distinct features.Ina ddition, howt ou se these features to achieve desired performance,i ncluding extended circulation, active targeting,controlled drug release and antitumor efficacy,i sdiscussed and summarized. We expect that this minireview will inspire more advanced studies in MIP-based nanomedicines for cancer therapy.

Extended Circulation
Defective tumor vessels and impaired lymphatic drainage allow intravenously delivered nanomedicines to preferentially accumulate in tumors,which is termed the passive targeting of cancer nanomedicine. [12] Sufficiently prolonged circulation contributes to the increased accumulation of nanomedicines in tumors.T herefore,r ationally designed nanomedicines avoiding rapid clearance in vivo are essential for the efficiency of passive targeting. Poly(ethylene glycol) is the most widely employed polymer for the surface modification of nanoparticles,w hich can delay phagocytic clearance and extend circulation time by hindering the adsorption of opsonic proteins. [13] Another biologically inspired strategy to extend the circulation time of nanomedicine emerged recently.N anoparticles can be disguised by cell membrane ligands or bioligands as autogenous cells to escape immune system recognition and elimination, thus prolonging the circulation time in the blood. [14] However,t he modification of nanoparticles with ligands complicates their preparation and alters their surface properties and stability.
Inspired by the biomimetic principle,T oshifumi Takeuchi et al. engineered and prepared molecularly imprinted nanogels by emulsifier-free precipitation polymerization with human serum albumin (HSA) as an imprinting template and pyrrolidyl acrylate,i sopropyl acrylamide,m ethacryloyloxyethyl phosphorylcholine as functional monomers. [15] The enhanced binding activity and high selectivity of the HSAimprinted nanogels to HSA was confirmed with as urface plasmon resonance (SPR) sensor.After intravenous injection into mice,the HSA-imprinted nanogels interacted with HSA in blood, inducing the formation of an albumin-rich protein corona around the nanogels.A ni nv ivo imaging experiment demonstrated that HSA-imprinted nanogels acquired stealth capability with al onger retention time in blood vessels ( Figure 1) and little accumulation in liver.Importantly,HSAimprinted nanogels exhibited at ime-dependent and approximately sixfold higher accumulation in tumors than in normal organ regions.Without extra modification and with just "self" makers,t he nanoMIPs can use the intrinsic biomolecules to camouflage themselves to inhibit phagocytosis,l eading to extended circulation and tumor accumulation. [15] Following this concept, more new nanoMIPs can be expected and used for passive targeting in cancer nanomedicine.

Active Targeting of Tumor Cells
Malignant cells co-exist with non-malignant stromal cells including fibroblasts,p hagocytes,e ndothelial cells,a nd perivascular cells embedded in aprotein-rich extracellular matrix and interstitial fluid, all of which constitute acomplex tumor microenvironment. [16] After extravasation of nanomedicines from the systemic circulation to the tumor sites,s elective targeting of cancer cells is crucial. Thes pecifically or highly expressed biomolecules in cancer cells related to tumor growth, proliferation, angiogenesis,a nd metastasis serve as potent candidates for tumor targeting. [17] Va rious nanomedicines for active targeting have been developed to improve tumor localization and increase tumor retention. [3b, 18] The active targeting of nanomedicines usually relies on biological ligands including antibodies,aptamers,and peptides. [19] However,t here are an umber of unresolved issues for the use of biological ligands.T he screening and preparation of biological ligands with high quality to target biomolecules is costly, time-consuming,and sometimes impossible.Invivo instability and the inherent immunogenicity of biological ligands are  half-life of HSA-imprinted nanogels (MIP-NGs) and non-molecularly imprinted nanogels (NIP-NGs) in the blood stream (n = 5). Pvalues were calculated from atwo-tailed t-test. [15] Reprinted with permission.C opyright 2017, Wiley-VCH. other challenges in cancer nanomedicine.Incontrast, without modification of biological ligands,n anoMIPs exhibit active targeting owing to their intrinsic specific recognition. Thus, nanoMIPs provide potential alternatives to existing biological-ligand-based nanomedicines for active targeting.

Targeting Receptor Proteins
Specific surface receptor proteins overexpressed by tumor cells provide ideal targets for the recognition of tumor cells. Targeted therapies by targeting receptor proteins overexpressed on tumor cell surface (e.g. integrin or CD20) enhance tumor cell killing and/or reduce off-target effects,and some of these have been successfully translated to clinical cancer therapy such as Trastuzumab,P ertuzumab,a nd Ibritumomab. [20] Receptor protein-binding MIPs have been exploited for the development of active targeting nanomedicines.T he reported nanoMIPs that are able to recognize surface receptor proteins expressed on cancer cells as well as their performance are summarized in Table 1. Thed issociation constant (K d )isaparameter representing the affinity of aMIP towards the target molecule.S electivity represents the specificity of aM IP to the target molecule,w hich is often quantitatively described by the cross-reactivity towards nontarget compounds.
Using entire protein-templated MIPs is adirect strategy to construct active targeting nanomedicines.A sa ne xample, cancer-specific antigen 125 (CA 125) was used as at emplate and imprinted on the surface of graphene oxide with dopamine as am onomer. [21] TheC A125-imprinted graphene oxide showed specific binding capacity to CA 125 with an imprinting factor (IF) value of 6.4. IF is defined as the ratio of saturated adsorption of the template on aprepared MIP over an on-imprinted polymer (NIP) prepared with the same imprinting procedure in the absence of template,w hich reflects the binding capability of the MIP over the NIP.When the antitumor drug doxorubicin (DOX) was loaded, the CA 125-imprinted graphene oxide led to the increased drug uptake by tumor cells overexpressing CA 125, and enhanced cytotoxicity against tumor cells.H owever, to avoid the conformational change of proteins,aqueous or biocompatible polymerization conditions are required for the preparation of entire protein-imprinted polymers,which becomes achallenge and al imitation. In addition, due to their large molecular sizes,proteins are often difficult to remove from highly crosslinked polymer networks.Moreover,asthe imprinted cavities are associated with al arge surface decorated with aw ide variety of functional groups,e ntire protein-imprinted MIPs are inclined to cause higher cross-reactivity. [22] Therefore, entire protein-imprinted MIPs for active targeting have been limited in the field of cancer nanomedicine.
To circumvent the limitations of entire protein-templated MIPs in recognizing receptor proteins,e pitope-imprinted polymers have been developed to recognize receptor proteins on tumor cells.E pitope imprinting was initially proposed by Rachkov and Minoura [23] and further developed by Shea and co-workers. [24] Although only as hort peptide fragment (epitope) that can represent whole proteins or peptides was utilized as the imprinting template,the synthesized MIPs can recognize whole proteins or peptides carrying the epitope peptide. [25] Compared with entire protein-templated MIPs, the epitope-imprinted polymers feature significant advantages with improved affinity to target proteins,r educed nonspecific binding, easy template synthesis,a nd widely applicable synthesis conditions. [26] These features make epitopeimprinted polymers highly desirable for whole proteins/ peptides that are stable only under physiological conditions or rarely available.T he epitope imprinting techniques have been thoroughly reviewed. [7e] Based on their merits,epitopeimprinted polymers are now widely adopted for developing nanoMIPs in cancer therapy.
Al inear peptide epitope of vascular endothelial growth factor (VEGF) was used as an imprinting template to prepare nanoMIP against human VEGF. [9a] ThenanoMIP was further coupled with quantum dots (QDs) to generate af luorescent hybrid nanoprobe,n amed QD-MIP.T he QD-MIP exhibited excellent affinity for the epitope of VEGF with K d in the nanomolar range (1.6 nM), while showing low binding to nontarget molecules. [9a] Thebinding ability of the QD-MIP to the whole human VEGF was also demonstrated. Though no concrete imprinting parameter on the protein level was provided, the QD-MIP specifically targeted VEGF-overexpressed human melanoma cells xenografted in zebrafish embryos ( Figure 2A). In another study,using alinear peptide epitope of epidermal growth factor receptors (EGFR) as the imprinting template,e pitope-imprinted nanoMIP was prepared through as olid-phase imprinting strategy. [9b] The nanoMIP exhibited high affinity,w ith a K d of 7.7 nM for binding with the peptide and a K d of 3.6 nM for binding with the extracellular domain of EGFR protein. [9b] Consequently, the nanoMIP was capable of selectively recognizing tumor cells overexpressing EGFR, which was demonstrated by flow cytometric analysis and confocal fluorescence imaging (Figure 2B). Similarly,t he linear peptide epitopes of human fibroblast growth factor-inducible 14 (FN14), human epidermal growth factor receptor 2(HER2), and CD59 were used as imprinting templates to prepare nanoMIPs for specifically binding receptor proteins overexpressed in different solid tumors. [27] Since some peptide epitopes exhibit aspecific and spatial conformation when integrated in the entire protein, linear epitope imprinted polymers,e specially when applied in biological systems,can hardly recognize atarget epitope with as pecific conformation. Therefore,c onformational peptide epitope-imprinted polymers have emerged as ap otent compliment for linear peptide epitope-imprinted polymers in cancer nanomedicine.
Membrane protein p32 is overexpressed on the surface of avariety of tumor cells.The N-terminal with a-helix structure serves as ab inding site for p32 recognition. To recognize the N-terminal with a-helix structure,anovel polypeptide HAPPE was used as an imprinting template to prepare conformational epitope-templated nanoMIP. [9c] Through polymerization in deionized water and trifluoroethanol, the a-helix conformation of the peptide was preserved during the imprinting process.T he resulted nanoMIP was capable of specifically binding the conformational epitopes as well as the Angewandte Chemie linear epitopes.I mportantly,t he synthesized nanoMIP was capable of specifically binding recombinant p32 with a K d of 340 nM, leading to greater cellular uptake by p32-positive cancer cells than with control NIP nanoparticles.Asaresult, [a] Calculated from the original data in the research articles.

Angewandte
Chemie the nanoMIP exhibited enhanced accumulation in p32positive tumors in amouse model inoculated with 4T1 breast cancer cells ( Figure 3A). Combined with photodynamic treatment, the nanoMIP potently inhibited tumor growth in asubcutaneous mouse tumor model. By comparison, another nanoMIP using alinear analogue of HAPPE as the imprinting template exhibited poor uptake by p32-positive cancer cells and low accumulation in tumors. [9c] Thetargeting capability of ananoMIP in vitro and in vivo was also reported where an Nterminal epitope of p32 with a-helix structure was used as an imprinting template. [28] In other interesting research, an N-terminal peptide epitope of folate receptor-a (FR-a)w as forced to fold into an a-helix structure as an imprinting template.T he synthesized nanoMIP could bind the template peptide and change the target peptide from ad isordered to an ordered conformation, thus "creating" binding sites in target receptors. [9d] Then anoMIP specifically targeted FRa-overexpressing He-La cells without being affected by the natural ligand, folate,in vitro and in vivo ( Figure 3B). Apart from the extracellular domain, the transmembrane domain with the a-helix struc-ture of FN14 was also chosen as the imprinting template to synthesize nanoMIP. [29] Thep repared nanoMIP exhibited specific binding to template peptide,w hich induced greater cell endocytosisand superior tumor penetration capability in FN14-positive tumors.
In addition, conformational epitopes can also be designed by in silico methods according to the protein sequence.Hamid et al. calculated the epitopes of HER2, then selected and synthesized the most powerful epitope. [30] Theimprinting was conducted on the surface of silica nanoparticles with the conformational epitope of HER2 and DOXa sc o-templates and dopamine as the monomer.Though the specific targeting ability to HER2 was not validated at the molecular or cellular level in an subcutaneous tumor model, the synthesized nanoMIP could specifically target ovarian cancer,a se videnced by the higher DOX concentration in the tumor tissues, enhanced antitumor efficacy, and prolonged survival of mice.
Va rious nanoMIPs have been successfully constructed for the active targeting of tumor cells by recognizing the tumorassociated receptor proteins with encouraging results. [9a-d, 27-30] However,s everal challenges still remain in this direction. Entire protein-imprinted polymers are difficult to prepare, especially for unstable or rare proteins.T hough peptide epitope-imprinted polymers provide potent alternatives to entire protein-imprinted polymers for protein recognition, the effective epitopes need to be elaborately chosen to ensure Figure 2. A) Images of human melanomacells (green) and the fluorescent hVEGF epitope-imprinted nanoprobes (red), demonstrating the ability of hVEGF epitope-imprinted nanoprobes to localize cancer cells overexpressing hVEGF in zebrafish embryos. [9a] Reprinted with permission. Copyright2 017, American ChemicalS ociety.B)FACS analysis and confocal microscopyo fbinding specificity for fluorescentE GFR epitope-imprinted nanoparticles to two breast cancer cell lines expressing different levels of EGFR. SKBR-3 cells express low amounts of EGFR while MDA-MB-468 cells express high amounts of EGFR, confirming the specific binding of EGFR epitope-imprinted nanoparticles (green) to the target protein.  the specificity and functionality of the prepared nanoMIPs. Forp eptide epitopes with specific and spatial conformation, the conformational maintenance during the imprinting process requires strict control over the composition of the polymerization reaction system, limiting the available monomer as well as polymerization methods.C onformational epitopes usually have specific secondary structure,a nd they are thus easily restricted by the polymeric network of the obtained MIPs.T herefore,t emplate removal should be considered to make the imprinted cavities available for the targeted molecules to bind. More advancements in MIPs for protein recognition are expected to improve nanoMIPs for the active targeting of protein receptors in tumor cells.

Targeting Aberrant Glycans Expressed by Tumor Cells
Glycans are auniversal and essential component of living organisms and are abundant in either free form or in covalent complexes with proteins or lipids;t hey range from am onosaccharide to polysaccharides with thousands of units. [31] Glycans on the cell surface play av ital role in am yriad of biological events,i ncluding cellular adhesion and migration, organism development, disease progression, and immunological response. [32] Tu mor occurrence and progression are usually accompanied by abnormal changes in glycosylation, making aberrant expression of certain glycan structures potential markers for cancer cells. [33] Commonly observed changes in glycan structures during malignancye ncompass abnormal branching of N-linked glycans (N-glycans), truncated O-linked chains,d iverse fucosylated and sialylated terminal structures,a nd alterations in glycosphingolipid expression. [17a,34] Unfortunately,l igands that can specifically recognize or bind relevant glycan biomarkers are scarce, [35] which limits the development of glycan-targeting nanomedicine for cancer therapy.
Taking advantages of molecular imprinting techniques, researchers have developed MIP-based nanomedicines specifically targeting glycans on the cell surface.T ob ind oligosaccharides on the cell surface having ag lucuronic acid (GlcA) terminal, GlcA-imprinted polymers were synthesized. [36] TheG lcA-imprinted polymers showed binding capability to phenylglucuronic acid and acetic acid, which have structural similarity with GlcA, while exhibiting less than 1% cross-reactivity in binding glucose,g alactose, Nacetylglucosamine,and N-acetylgalactosamine.Furthermore, the GlcA-templated nanoMIP had ahigh association constant (K a )w ith GlcA in dimethyl sulfoxide (7.1 10 3 M À1 )d ue to the noncovalent binding with the functional monomers. [36b] Using carbon nanodots as ac ore,t he GlcA-imprinted nano-MIP was demonstrated to target human cervical cancer cells by binding hyaluronan. [36c] Initially,o nly noncovalent forces were used for binding the target monosaccharide in the design of monosaccharidebinding nanoMIPs. [36] To gain higher affinity to glycans, boronate affinity was introduced later in glycan-imprinted polymers.Boronate affinity materials have gained increasing attention in the field of glycan recognition since they can undergo reversible covalent binding with cis-diol-containing compounds controlled by the solution pH. [37] Despite the superior selectivity of boronate affinity materials to cis-diols, they fail to recognize specific cis-diol-containing compounds. Thecombination of boronate affinity and molecular imprinting provides an optimal solution for developing targeting reagents to glycans.B oronic acids can be used for facilely anchoring or releasing glycan templates during the imprinting process through control of the solution pH. Thei mprinted cavities,which preferentially fit the template glycans,and the affinity provided by boronic acids endow the fabricated MIPs with high selectivity and affinity to target glycans.Sellergren et al. designed and fabricated af luorescent sialic acid (SA)imprinted nanoMIP for selectively labeling cell-surface glycans. [38] Thet ernary complex imprinting approach including boronate ester,h ydrogen bonding, and electrostatic stabilization produced at ailor-made nanoMIP with exceptional affinity for SA (binding constant K = 6.6 10 5 M À1 in 2% water, 5.9 10 3 M À1 in 98 %w ater). In quantification assays of the expression level of SA, the nanoMIP exhibited asimilar staining pattern as lectin but asignificantly stronger staining for SA than for lection, indicating ahigher affinity to SA. In another study using af acile re-precipitation of phenylboronic acid modified poly(fluorene-alt-benzothiadiazole), the fluorescent SA-imprinted nanoMIP exhibited selective staining for cancer cells overexpressing SA. [39] Themonosaccharide-imprinting strategy for cell targeting was further explored by our group via construction of ag eneral toolbox for specifically recognizing cancer cells by monosaccharide-imprinted nanoMIPs. [40] SA, fucose (Fuc), and mannose (Man) have been reported to be overexpressed on certain cancer cell surfaces and can serve indicators for some cancers.Based on the boronate affinity oriented surface imprinting approach, [41] nanoMIPs were prepared with different templates including SA, Fuc, and Man. Theo ptimal IF reached up to 8.4, 6.7, and 6.9 for SA, Fuc, and Man, respectively,w hile the cross-reactivity toward nontarget monosaccharides was less than 7.1, 26.2, and 22.2 %f or SA-, Man-, and Fuc-imprinted nanoMIPs,r espectively.T he K d value of the SA-imprinted nanoMIP was estimated to be 2.0 10 À4 M. With high affinity and low cross-reactivity,t he fluorescent monosaccharide-templated nanoMIPs well reflected the monosaccharide expression level on cell surface (Figure 4). Compared with lectins,the monosaccharide-binding nanoMIPs provided comparable cancer cell targeting capability but better specificity. [40] Furthermore,the in vivo targeting performance of monosaccharide-templated nanoMIPs was investigated. [42] SAimprinted nanoMIP with gold nanorods (AuNRs) as ac ore was prepared. [42a] Due to the boronate affinity and imprinted cavities,t he SA-imprinted nanoMIP exhibited excellent specificity and high affinity to SA, leading to enhanced in vivo accumulation at the tumor site compared to NIP nanoparticles.C ombined with the plasmonic heating effect of AuNRs,the SA-binding nanoMIP completely ablated solid tumors in as ubcutaneous tumor model. In another study, ah ollow double-layer SA-imprinted nanoMIP was designed and prepared as achemotherapeutic for targeted nitric oxide release. Compared with monosaccharides,t he aberrant glycan chains hold more potential as cancer markers with high specificity.N anoMIPs for the recognition of specific glycan chains were developed by our group using the boronateaffinity controllable oriented surface imprinting approach. [41,43] Glycans digested from the target glycoprotein were immobilized onto the boronic acid functionalized nanoparticles through boronate-affinity interactions and served as imprinting templates.A lternatively,t he target glycoproteins could be directly immobilized on the boronic acid functionalized nanoparticles and then digested by peptidase,l eaving the glycan peptides as the imprinting templates.W ith tetraethyl orthosilicate as the monomer,t he thickness of the imprinting polymer layer could be controlled to appropriately cover the template,r esulting in cavities specific to the template glycans as well as to the intact glycoprotein or glycopeptide containing the template glycan. Using the imprinting approach, HER2-glycan-imprinted nanoMIPs were prepared, specifically recognizing HER2-positive breast cancer cells, [9e] which will be highlighted later in Section 5. Theglycan-imprinting approach provides awide and efficient avenue for the preparation of nanomedicines targeting glycans.
So far, few active glycan-targeting cancer nanomedicines have been developed owing to the shortage of glycan-specific ligands.However,more and more studies support the notion that the aberrant expression of glycans by malignant cells provides ag reat opportunity to explore glycans as specific targets for cancer cells. [17a,44] Regarding the targeting reagents to glycans,M IPs,e specially boronate-affinity-based MIPs, exhibit significant advantages over traditional biological ligands such as lectins and antibodies with respect to specificity,a ffinity,s tability,a nd accessibility.T herefore, glycan-imprinted polymers will become ap otent platform for constructing nanomedicines targeting glycans.

Controlled Release of Antitumor Drugs
Chemotherapy is the mainstay of cancer therapy.C ontrolling drug release to improve treatment efficacyh as been ahot topic in the field. MIPs have been constructed in various devices (e.g. bulk hydrogel, patches,c oatings) for ocular, dermal, intravenous,and oral drug delivery. [45] Themolecular imprinting technology elicits new polymer formats and shifts the material size to the nanoscale,p roviding av ersatile tool for cancer nanomedicines with controlled drug release.F or systemically administered cancer nanomedicines,p remature drug release during blood circulation would reduce the amount of the drug arriving at tumors and also cause side toxicity. [3b, 46] Upon reaching the target site,the nanomedicines are expected to fully release the loaded drugs to achieve an effective therapeutic concentration. In addition to reaching the correct site,the drug delivery system must also sustain an effective drug concentration in tumors to maximize therapeutic efficacy. Aiming at these release kinetics,c urrent nanoMIPs for controlled antitumor drug release emphasize prolonged release and stimuli-responsive release.
Usually,a ntitumor drugs serve as template molecules for imprinting.T he imprinted cavities and functional reactive residues left after template removal are utilized for drug loading. Ford rug release,t he drug molecules must escape from multiple noncovalent or covalent interactions and diffuse from cavities.D uring the release of drug molecules from MIPs,the high affinity of the MIPs to the drug molecules results in am ore controlled release kinetics. [47] This mechanism has been utilized for the prolonged release of antitumor drugs.O ne example is ac arbazole derivative released from am agnetic nanoMIP prepared with methacrylic acid as the functional monomer and 1,4-dimethyl-6-hydroxy-9H-carbazole as the template molecule for controlled antitumor drug delivery. [48] Compared to the NIP which displayed total drug release in about 6h,the nanoMIP prolonged the drug release for more than 48 h. Similar results were also reported by Neda et al.,where a5 -fluourouracil-templated nanoMIP exhibited asustained release for more than 96 h, four times longer than the NIP counterpart. [49] Noncovalent forces (including hydrogen bonding,van der Waals forces,a nd hydrophobic interactions) are generally responsible for the affinity of MIPs to templates.T he sensitivity of some interactions to the external environmental pH and temperature renders stimuli-responsive release kinetics from nanoMIPs, [47] which is favorable for the stimuli-responsive release of antitumor drugs.H ydrogen bonding is the most commonly employed noncovalent force for template anchoring as well as for drug loading in MIPs. Anti-neoplastic agents including paclitaxel, 5-fluorouracil, and DOX have been loaded in their corresponding nano-MIPs,w here hydrogen bonding provided the main binding force. [27b,28, 50] Fast drug release was observed under acid conditions because the disruption of hydrogen bonding weakened the affinity of the imprinted cavities to the drug molecules.E lectrostatic interactions were also utilized for stimuli-responsive release.I nw ork by Peng, hollow microcapsules with DOX-imprinted shells were designed and prepared to encapsulate and deliver DOX. [51] Thei mprinted cavities could be blocked by DOX molecules due to electrostatic interactions between amino group of DOXa nd carboxyl groups of the MIP.T he interaction could be weakened at low pH, resulting in ap H-responsive release. Theh ollow microcapsules exhibited as ustained release for more than 168 ha tp H6.5, while ab urst release delivered 75 %o fthe loaded drugs in 24 ha tp H5.0. Thep H-sensitive release kinetics contribute to overall drug stability during systematic circulation in aneutral pH environment and rapid release in the acidic environment of the tumor, which is beneficial in cancer therapy.I na ddition, temperatureinduced fast release was also observed, where the binding force was affected by temperature change. [52] Apart from noncovalent forces,covalent forces were also used by nanoMIPs for stimuli-responsive release.Afunctional monomer containing boronic acid derivative was used to prepare an anoMIP. [27a] Through the pH-sensitive interaction between the boronic acid group and cis-diols,t he prepared nanoMIP effectively loaded cis-diol-containing bleomycin under alkaline conditions,w hile quickly and effectively releasing the loaded drug in an acidic environment. [27a] Though imprinted cavities and reactive forces endow molecule-imprinted nanoMIPs with the controlled release of antitumor drugs,a mple room remains for further improvement, with critical barriers that need to be addressed. Almost all nanoMIPs suffer from the initial burst release of their cargo caused by nonspecific absorption on the surface of nanoparticles,aresult of the inherent high ratio of surface to volume.T odate,the controlled release of antitumor drugs in vivo has not been completely investigated, though several nanoMIPs have been used for antitumor drug delivery in vivo. In addition to the small-molecule antitumor drugs,t he potential of nanoMIPs for the delivery of antitumor macromolecules needs to be explored in the future.

Antitumor Efficacy
Nanomaterials are generally employed as vehicles to deliver therapeutics to tumor sites in cancer nanomedicine. NanoMIPs,h owever,n ot only provide drug delivery platforms but have also been functionally used as therapeutic antagonists to target biologically functional molecules in cancer. With the increasing understanding of tumor occurrence,p rogression, invasion, and metastasis,c onsiderable tumor-related signaling pathways have been identified. Activating or blocking the interaction between ligands and receptors can effectively induce apoptosis of tumor cells. [53] Theb iomimetic recognition of MIPs has been utilized in the design of nanoMIPs for cancer therapy without loading extra drugs.
Catalase (CAT)-imprinted fibrous SiO 2 nanoparticles were prepared as an ovel nanotrapper for CATt ot rigger tumor cell apoptosis. [54] Thep repared CAT-templated nano-MIP efficiently trapped CATa nd inhibited its catalytic activity of decomposing hydrogen peroxide (H 2 O 2 ). When internalized by tumor cells,t he CAT-imprinted nanoMIP selectively recognized and captured CATmolecules,inducing intracellular accumulation of H 2 O 2 ,one of the major reactive oxygen species (ROS). Theh igh ROSl evel naturally triggered apoptosis of most tumor cells. [54] Asimilar example was reported by Tang et al. [55] Te stosterone (TSTO)-imprinted nanoMIPs were developed to block the TSTO-androgen receptor (AR) pathway.T hese nanoMIPs specifically absorbed intracellular testosterone and then inhibited the cascade of TSTO-AR pathway related functions as well as the growth of androgen-dependent prostate cancer cells through suppression of the cell cycle progression.
In addition to intracellular signal pathways,i ntercellular interactions could also be regulated by nanoMIPs.C adherin plays ac rucial role in mediating cell-cell adhesion. TheNterminal peptide epitope of cadherin was used as at emplate to prepare an anoMIP by solid-phase synthesis. [9f] The nanoMIP specifically bound to cadherin on the cell surface, disrupting cell-cell adhesion ( Figure 5). Furthermore,t he nanoMIP has been shown to disrupt preformed tumor spheroids and inhibit cancer cell invasiveness of HeLa cells in vitro,e xhibiting biological activity of modulating cellular function.
Thespatial organization of cell-surface receptors plays an important role in controlling cellular signaling cascades. [56] Controlling the nanoscale distribution of cell binding ligands has been shown to regulate cell behavior, and these can serve as at arget for cancer therapy. [57] Heterodimerization of HER2 with other EGFRs at the extracellular subdomain induces phosphorylation of tyrosine residues within the intracellular tyrosine-kinase domain and triggers several downstream signaling events,r esulting in cell proliferation, survival, migration, angiogenesis,a nd metastasis. [58] Preventing the dimerization of HER2 with EGFRs by monoclonal antibodies provides an effective treatment for HER2-positive breast cancer. [59] As an alternative to antibodies,t he HER2glycans-templated nanoMIP was employed to bind HER2 glycans.T his will suppress the dimerization of HER2 with other HER family members to finally block the downstream signaling pathways and inhibit HER2-positive breast cancer growth ( Figure 6). [9e] Then anoMIP specifically inhibited HER2 phosphorylation and proliferation of HER2-positive tumor cells in ad ose-dependent fashion. Thea ntitumor efficiency was validated in nude mice implanted with human HER2-positive breast cancer cells. [9e] Without the loading of Angewandte Chemie additional therapeutic agents,t he HER2-binding nanoMIP significantly inhibited tumor growth without showing obvious biological toxicity compared to control groups. [9e] By specifically recognizing and capturing pivotal biomolecules,n anoMIPs could effectively block signaling pathways to directly or indirectly induce antitumor effects.D ifferent from classical nanomaterials serving as vehicles for the targeted delivery of therapeutics,n anoMIPs themselves can exert therapeutic efficacyi nab iomimetic and molecularly selective way without the further loading of therapeutic agents.T his paradigm is similar to some antibody therapeutics.H owever,c ompared with antibodies,n anoMIPs have lower immunogenicity,higher stability,and lower toxicity and side effects.Additionally,nanoMIPs readily penetrate the cell membrane by cell endocytosis to regulate intracellular signals. Although the use of nanoMIPs as cancer therapeutic agents is still at avery early stage,nanoMIPs as drug-free therapeutics have strong potential for cancer therapy.

Dual Functions
Owing to the versatile functions of imprinted cavities, as ingle nanoMIP system can be used to integrate different functions for more effective cancer therapy.T he most widely studied dual-functional nanoMIPs are capable of simultaneously combining target molecules and antitumor drugs as cotemplates.The imprinted cavities provide specific recognition sites as well as drug loading sites.T he constructed nanoMIPs could specifically deliver antitumor drugs to tumor cells with controlled kinetics.I nl ine with this concept, several dual templates of nanoMIPs have been constructed and validated to specifically deliver cytotoxic drugs to tumors cells overexpressing ag iven receptor,l eading to am ore profound tumor cell toxicity compared to that of NIP nanoparticles or free drugs.T his was attributed to the specific recognition of the nanoMIPs as well as controlled drug release. [9b,27, 28, 30] Furthermore,t he in vivo antitumor efficacyw as explored in mouse tumor models,d emonstrating dual-functional nano-MIPs as an efficient platform for tumor-targeted therapy. [27a, 28, 30] To further advance cancer therapy,i ti se xpected that more dual-functional and even multi-functional nanoMIPs will be developed in the near future.For example,toeliminate the insufficient efficacyo fs ingle targeting, dual-targeting nanoMIPs can be constructed to improve tumor cell selectivity and uptake.T or educe drug resistance,c ombined drug delivery can be designed by using dual-or multi-templated nanoMIPs.

Summary and Outlook
In this minireview,w eh ave surveyed and highlighted recent advances in the rational design of MIP-based nanomedicines for cancer therapy.A sapolymer material, MIPs feature unique molecular recognition capability with high affinity and specificity.I mportantly,M IPs have excellent stability,e asy preparation, and low cost compared to traditional bioligands.G iven these merits,M IP-based nanomedicines have exhibited remarkable performance in systemic circulation, specific targeting,a nd controlled drug release when serving as carriers.O fn ote,n anoMIPs themselves can also serve as therapeutics by regulating cell signal pathways without loading any drug. This approach holds great potential to provide alternatives to classical chemical and biochemical drugs.F urthermore,n anoMIPs can integrate multiple functions into asingle system, presenting apowerful and versatile platform for synergistic therapeutics.T his is considered to be am ore effective strategy to achieve complete remission and even cures for patients with cancer.
Thea rea of MIP-based nanomedicines remains in its infancywith many exciting but unexplored areas.Inthe past,  most MIP-based nanomedicines aimed to address only one issue.H owever,t he successful application of nanomedicine must overcome ac ascade of biological barriers.T oa chieve optimal performance for cancer therapy,amore exquisite and comprehensive design of MIP-based nanomedicines is critical. Forf urther clinical translation of MIP-based nanomedicines,t he biocompatibility and biodegradability of the nano-MIPs as well as appropriate animal models for in vivo experiments must also be considered and investigated.
Although they faced many challenges at the initial stage, nanoMIPs have opened an ew research avenue and exhibit great potentials in cancer nanomedicine.With the progress in the molecular imprinting technology,n anoMIPs and their superior properties will significantly advance cancer therapy.