Nanocomposites Based on Magnetic Nanoparticles and Metal–Organic Frameworks for Therapy, Diagnosis, and Theragnostics

Abstract

In the last two decades, metal–organic frameworks (MOFs) with highly tunable structure and porosity, have emerged as drug nanocarriers in the biomedical field. In particular, nanoscaled MOFs (nanoMOFs) have been widely investigated because of their potential biocompatibility, high drug loadings, and progressive release. To enhance their properties, MOFs have been combined with magnetic nanoparticles (MNPs) to form magnetic nanocomposites (MNP@MOF) with additional functionalities. Due to the magnetic properties of the MNPs, their presence in the nanosystems enables potential combinatorial magnetic targeted therapy and diagnosis. In this Review, we analyze the four main synthetic strategies currently employed for the fabrication of MNP@MOF nanocomposites, namely, mixing, in situ formation of MNPs in presynthesized MOF, in situ formation of MOFs in the presence of MNPs, and layer-by-layer methods. Additionally, we discuss the current progress in bioapplications, focusing on drug delivery systems (DDSs), magnetic resonance imaging (MRI), magnetic hyperthermia (MHT), and theragnostic systems. Overall, we provide a comprehensive overview of the recent advances in the development and bioapplications of MNP@MOF nanocomposites, highlighting their potential for future biomedical applications with a critical analysis of the challenges and limitations of these nanocomposites in terms of their synthesis, characterization, biocompatibility, and applicability.

1. Introduction

Since 2006, metal–organic frameworks (MOFs) have been exponentially investigated in the biomedical field,¹⁻⁶ where they have been proposed as imaging agents⁷ and drug carriers.⁸ These materials are coordinative networks, based on metallic inorganic subunits (e.g., chains, clusters, atoms) and organic bridging ligands (Figure 1), with crystalline architectures, presenting a high and tunable porosity and structural versatility.⁹⁻¹¹ In this sense, their properties make them promising candidates in several biomedical areas such as drug delivery systems (DDSs),^5,12,13 biosensing,^14,15 antimicrobial therapy,^16,17 biomedical imaging,² phototherapy,^18,19 and theragnostics,^20,21 among others.^11,22

Figure 1.

Schematic representation of MOF structure.

An important consideration for the biomedical application of MOFs is their safety; it is crucial to ensure and evaluate the synthetic route and the final chemical composition taking into account that the solvents, metal ions, and organic ligand precursors could possess potential toxicity.²³ Thus, the selection of safe and biocompatible metal and ligand precursors is very important.²⁴ Furthermore, the size, morphology, and surface properties of the MOF are of pivotal relevance both for its biosafety and efficacy.^20,24 Indeed, the main features in MOFs as DDSs are the biocompatibility, porosity available for drug loading, and controlled release at the targeted site, which are closely related to its physicochemical properties and dimensions. For instance, intravenous administration generally necessitates a size below 200 nm; therefore, the nanoscale design, encompassing factors such as size, shape, and surface functionalization, can influence the capacity for cell-specific targeting and subsequent cellular uptake.²⁵ Furthermore, particles smaller than 250 nm have been reported to exhibit a higher likelihood of extravasation through leaky endothelium via the enhanced permeability and retention (EPR) effect, a characteristic fundamental for deposition in the tumor targeting site.^25,26 Additionally, the clearance route is affected by the size: nanoparticles larger than 200 nm in diameter are preferentially cleared by the reticuloendothelial system, whereas those smaller than 10 nm are eliminated by renal filtration.²⁵ In this context, nanoscaled MOFs (nanoMOFs) have attracted great attention in this matter because of their optimal size. Moreover, they can be tuned not only to target a particular administration route and safety but also the biodistribution.²⁷⁻³⁰

In addition to the advantageous properties of nanoMOFs, incorporating guest materials into MOFs, thereby forming MOF nanocomposites, presents a promising avenue for enhancing the performance of these structures in the realm of nanomedicine. Currently, there is a growing body of reviews that recognize the importance of improving MOF properties with the integration of organic polymers,³¹⁻³³ enzymes,^34,35 metals and metal oxides,^36,37 silica,³⁸ polyoxometalates,³⁹ quantum dots,^40,41 and carbons,^42,43 among others.⁴⁴ In this regard, MOF nanocomposites are becoming particularly promising as DDSs, by associating nanoMOFs with different nanometric (inorganic, organic) species^44,45 that will provide them with additional relevant properties (targeting, furtivity, therapeutic effect, imaging, etc.).^20,21

Of particular relevance are the composites based on nanoMOFs and magnetic nanoparticles (MNPs).⁴⁶⁻⁴⁸ In the biomedical field, MNPs are mainly represented by magnetite (Fe₃O₄), maghemite (γ-Fe₂O₃), and some ferrite colloids which typically have a hydrodynamic size below 100 nm.⁴⁹ Given the enormous variability in their synthesis, their physicochemical properties (e.g., size, shape, structure, surface charge, magnetism) are tuned for a multitude of applications, such as magnetically guided nanoparticles for drug delivery,^50,51 magnetofection (gene delivery),^50,52 magnetic hyperthermia,^53,54 magnetic resonance imaging (MRI)⁵⁵⁻⁵⁷ or magnetic particle imaging (MPI),⁵⁸ and magnetic separation (cell separation, cell sensing, biosensing, etc.).⁵⁹ Combining these relevant properties with those of MOFs (e.g., porosity, versatility) makes the resulting composites excellent candidates to be used in advanced drug delivery, nanothermometry, biosensing, bioimaging, and MRI contrast agents. A series of iron oxide nanoparticles with biocompatibility and nontoxicity are currently commercialized.^60,61 Despite the noteworthy advancements in MNPs and their significant clinical implications,⁶¹⁻⁶³ applying a nanocomposite system that harnesses the combined properties of MNPs and nanoMOFs introduces a compelling approach for augmenting the properties of both components while ensuring enhanced biocompatibility and potential efficacy for theragnostic applications.

Thus, here we will review the recent progress in the development of magnetic composites (MNP@MOF) based on MOFs and MNPs, specifically Fe-MNPs, discussing the main synthetic approaches and their challenging characterization with special attention to biomedical-related considerations, considering mainly nanocomposites (<500 nm) and the most significant submicron-sized composites (>500 nm). Further, the most relevant bioapplications of the MNP@MOF composites reported recently will be critically described, identifying their main advantages and limitations.

2. Synthetic Approaches for Magnetic Metal–Organic Framework Composites

To combine MNPs with MOFs, four main strategies have been traditionally employed depending on the synthetic route of the final composite: (1) mixing, where the composite is formed by simply putting in contact previously synthesized magnetic particles and MOF crystals,⁶⁴ (2) in situ formation of MNPs in the presence of preformed MOFs, (3) in situ synthesis of the MOF in the presence of magnetic particles, and (4) layer-by-layer, using functionalized nucleation sites to grow step-by-step the MOF by repeated cycles.⁶⁵ Furthermore, it is important to highlight that different morphologies can be obtained such as core–shell and non-core–shell structures (Figure 2). In a core–shell composite, the MOF acts as the shell surrounding a single nanoparticle core material. Therefore, the core material in the case of MNP@MOF nanocomposites is the MNP and this configuration provides unique properties and synergistic effects between the core and shell components. In contrast, a non-core–shell MOF composite refers to a structure where the MNPs are randomly distributed, resulting in a heterogeneous composite with the MOF and the additional MNPs allocated throughout the composite structure, either within the framework and/or on the outer surface.

Figure 2.

Schematic representation of the non-core–shell and core–shell MNP@MOF composites.

Overall, in the following section, the synthetic procedures are illustrated following the above classification, considering their architectural configuration.

2.1. Mixing

The simplest approach is the mixing method, which involves the interaction of both presynthesized MNPs and MOFs (Figure 3). The final composite is thus exclusively based on the stability of the interactions between the two components.⁶⁴

Figure 3.

Schematic representation of the mixing strategy.

Only a few examples are reported so far, mainly involving mesoporous iron(III) trimesate MIL-100(Fe)⁶⁶ or chromium(III) terephthalate MIL-101(Cr)⁶⁷ (MIL, Material Institut Lavoisier) as MOF, selected by their high porosity (up to S_BET ∼ 3000 m²·g^–1) and chemical robustness. Despite the in vivo proven biosafe character of MIL-100(Fe),⁶⁸ chromium-based materials (even if based on Cr(III)) are considered potentially toxic,⁶⁶ ruling out its interest in the biomedical field. However, MIL-101(Cr) is a benchmarked MOF widely proposed as a model in many fields due to its exceptional chemical stability. Indeed, the first synthesis of a magnetic MIL-101(Cr) composite by the mixing approach was reported in 2012, when Huo et al. described the formation of a Fe₃O₄@SiO₂-MIL-101(Cr) composite.⁶⁹ For the synthesis, silica-coated iron oxide microparticles (Fe₃O₄@SiO₂, ∼ 600 nm with about 30 nm of silica shell) and MIL-101(Cr) submicrometric crystals (∼650 nm) were dispersed in an aqueous solution under ultrasonication for 20 min. The negatively charged silica-coated Fe₃O₄ favored the electrostatic interactions with the positively charged MIL-101(Cr), leading to a microsized Fe₃O₄@SiO₂ assembled onto the external surface of the MIL-101(Cr) crystals. The silica coating, with an average thickness of about 30 nm, is required not only for preventing iron oxide corrosion and oxidation but also to favor the static electric interactions with the MOF. The saturation magnetization (M_S) value for Fe₃O₄@SiO₂-MIL-101(Cr) was 21 emu·g^–1 (vs 76 and 38 emu·g^–1 for Fe₃O₄ and Fe₃O₄@SiO₂, respectively), keeping the magnetic property for the desired magnetic solid-phase extraction aim. Critically, in this first work, the composite has unsteady characteristics based on too weak interactions between two enormous components attracted to each other in an insufficiently stable structure. In consideration of the pore size and the substantial dimensions of the MNPs, it is evident that these entities are predominantly located on the surface. The concept behind this work was further improved by reducing the size of the MNPs, which will interact more effectively with the MOF surface. Thus, Qian and co-workers⁷⁰ promoted the interaction between the here biocompatible iron version of MIL-101(Fe) (∼700 nm) and Fe₃O₄ in deionized water by increasing the pH up to 8 with a NaOH solution, which switches the ζ-potential of the MNPs from positive to negative. Even if the Fe₃O₄ nanoparticles presented an average size of ∼10–30 nm, through this method, the MNPs were restricted to the outermost layer. The resulting magnetic composite (M_S ∼ 26 vs 46 emu·g^–1 for Fe₃O₄) exhibited a powder X-ray diffraction (PXRD) pattern matching well with the indexed peaks of the Fe₃O₄, and some other peaks consistent with the MIL-101(Fe) structure. However, the majority of the diffraction peaks of the MOF were indistinguishable from the background. This result was attributed to the cover effect of the MNPs, but one could also consider a potential degradation of the Fe carboxylate MOF under basic pH (pK_a carboxylic acids ∼3–5 vs pH = 7.4), as previously reported for MIL-100(Fe).⁷¹

Similarly, magnetic composites γ-Fe₃O₄@MIL-100(Fe) and cit-γ-Fe₃O₄@MIL-100(Fe) with maghemite (γ-Fe₃O₄, ∼7–10 nm) and citrate-functionalized maghemite (cit-γ-Fe₃O₄, ∼15–20 nm) were synthesized with a mean hydrodynamic diameter of about 160 nm from dynamic light scattering (DLS) analysis.⁷² The nanocomposite, showing PXRD peaks of the MIL-100(Fe) and the maghemite, was achieved by mixing the previously microwave (MW)-synthesized MIL-100(Fe) nanoparticles⁷³ with γ-Fe₃O₄ or cit-γ-Fe₃O₄ at pH 4.2 and 3.5, respectively. The high-resolution transmission electron microscopy (HR-TEM) images revealed MIL-100(Fe) nanoparticles (∼130 nm) with a decoration on the outer surface of small aggregates of MNPs. The magnetometry experiments showed a superparamagnetic behavior. Among all the samples the cit-γ-Fe₃O₄@MIL-100(Fe) presented the highest magnetic moment with Brunauer–Emmett–Teller surface areas (S_BET) about 1180–1310 m²·g^–1, not significantly reduced compared to S_BET ∼ 1330 m²·g^–1 of the initial MIL-100(Fe).

In conclusion, the mixing synthetic method is however underdeveloped because, in spite of its simplicity, it has poor control over the final properties of the composite and the association is exclusively based on the formation of weak interactions (mainly electrostatic attraction). Therefore, it results in a lack of preference for MNPs on their directionality within pores or the outer surface of MOFs. Even if the crucial size of the MNPs may be compatible with the pore dimensions of the MOF, then this would be not sufficient to avoid the partial or total presence of the MNPs on the outer surface of the MOF, leading to less stable composites with MNPs’ leaching.

2.2. In Situ Formation of MNPs in the Presence of the MOF

Another notable pathway to obtain MNP@MOF composites is the “ship in a bottle” strategy, which consists of the in situ formation of the MNPs in the presence of the preformed MOF.⁷⁴ The magnetic composite is obtained by first incorporating iron ions or precursors of the MNPs in the MOF (mainly via chemical vapor infiltration, solution impregnation, and incipient wetness infiltration⁷⁵) and then, forming the MNPs through a transformation (e.g., phase and/or topotactic transformations, dehydration, reduction)⁷⁶ to iron oxide nanoparticles formation (see Figure 4). The MNPs are usually located on the MOF external surface or within the porosity, partially destroying the structure (creating defects) in case of particle size larger than the accessible pore dimension. However, this insertion of defect points in the framework is a hard equilibrium to control, in order to avoid the structure collapse. Otherwise, in the optimiztic and more desired case, the MNPs are well-dispersed inside the MOF,⁷⁵ being advantageously protected from degradation or leaching.

Figure 4.

Schematic representation of the in situ formation of the MNPs in the presence of presynthesized MOF for non-core–shell MNP@MOF composites.

In this sense, Wu et al.⁷⁷ originally proposed this “ship in a bottle” method to prepare γ-Fe₂O₃@ZIF-8 and γ-Fe₂O₃@MIL-53(Al) composites using rigid microporous zinc(II) 2-methylimidazolate ZIF-8 (ZIF = Zeolite Imidazolate Framework; ∼200 nm and S_BET ∼ 1800 m²·g^–1)⁷⁸ and flexible microporous aluminum(III) terephthalate MIL-53(Al) (S_BET ∼ 1500 m²·g^–1 in the open form structure).⁷⁹ The Fe(acac)₃ metal precursor was infiltrated in the MOF by the incipient wetness method, and γ-Fe₂O₃ MNPs were formed by pyrolysis (300 °C under N₂ for 1 h). Then, a reduction treatment was performed to form Fe⁰ (CO atmosphere at 550 °C for 1 h) and proved the formation of MNPs. The crystallinity of MOFs was retained, as shown by PXRD, and then, the magnetic composites were further investigated for drug encapsulation with the anti-inflammatory and analgesic ibuprofen. Under these conditions, the MNPs conferred a significantly lower M_S than that of pure γ-Fe₂O₃ (33.5 vs 1.8 and 6.1 emu·g^–1 for γ-Fe₂O₃@ZIF-8 and γ-Fe₂O₃@MIL-53(Al), respectively). The γ-Fe₂O₃ particles were mainly located on the outer surface of the ZIF-8 as large agglomerates; instead, ultrafine MNPs and clusters were observed within the crystalline structure, creating a hollow structure with defects during the pyrolysis process. This fact highlights that the location of the MNPs depends on the method to insert the precursor and its motion under the pyrolysis process, which is challenging to control. In addition, the pyrolysis process could be limited by the MOF thermal stability.

The in situ formation of MNPs has also been employed for the preparation of Fe₃O₄@MIL-101(Cr) composite by partial reduction.^80,81 MIL-101(Cr) micrometric crystals were dispersed and sonicated for 30 min at room temperature into a FeCl₃ solution to favor the impregnation. The in situ formation of MNPs (∼10–20 nm) occurred, first, with the addition of Na₂SO₃ solution dropwise. Then, NH₄OH solution was slowly added under an inert atmosphere, leading to the formation of Fe₃O₄ nanoparticles, as supported by PXRD. The final magnetic composite (M_S ∼ 15.6 emu g^–1), with a particle size of 600–700 nm, was successfully applied in magnetic solid phase extraction (MSPE) combined with ultrahigh performance liquid chromatography-tandem mass spectrometry (UPLC-MS/MS) for the quantitation of eight nonsteroidal anti-inflammatory drugs (NSAIDs) in wastewater and environmental water samples.⁸⁰

As reflected in the examples shown above, the difficulties of this method lay in the homogeneous diffusion of the metal precursors into the MOF and the challenging control of the in situ process. Moreover, the MOF stability under the in situ conditions is fundamental for the successful formation of the nanocomposite, being this procedure limited to high thermally and chemically robust MOFs.

2.3. In Situ Formation of MOF in the Presence of MNPs

One of the main approaches followed for the synthesis of magnetic nanocomposites, known as “bottle around ship”, consists of the in situ formation of the MOF in the presence of preformed MNPs (Figure 5).⁷⁴ The MNPs could be coated with different functional groups, polymers, or capping agents.⁷⁴ In the following sections, the synthetic methods will be distinguished considering the final architecture (non-core–shell or core–shell) of the composite as well as the MNPs functionalization (uncoated MNP, amine or acidic-coating, etc.).

Figure 5.

Schematic representation of the in situ formation of the MOF in the presence of presynthesized MNPs for non-core–shell and core–shell MNP@MOF composites.

2.3.1. Non-core–Shell Magnetic Composites

In the non-core–shell magnetic architecture, the MOF grows in the presence of preformed MNPs, leading to a composite where MNPs are often randomly distributed in/on the MOF.

2.3.1.1. Synthesis of Magnetic Composites with Nonfunctionalized MNPs

The complex structure of a magnetic nanocomposite can be obtained from the simplest uncoated MNPs, by interacting with the MOF precursors or being incorporated into the porosity that arises as the MOF grows.^82,83

In this matter, Lohe and collaborators⁸⁴ explored the benchmarked microporous copper(II) trimesate HKUST-1 (HKUST = Hong Kong University of Science and Technology; S_BET ∼ 600–1600 m²·g^–1)^85,86 together with other two MOFs, aluminum(III) 2,6-naphthalenedicarboxylate (DUT-4; DUT = Dresden University of Technolog; S_BET ∼ 1000–1300 m²·g^–1) and aluminum(III) biphenyl-4,4′-dicarboxylate (DUT-5; S_BET ∼ 1200–1700 m²·g^–1),⁸⁷⁻⁸⁹ as an efficient method for magnetically controlled catalyst separation. In this work, the magnetic composites were also studied as model systems for heating-trigger desorption of drug molecules by an external alternating magnetic field (AMF). In the synthesis, MNPs (spherical, ∼ 10–20 nm) were added to the organic ligand solution in N,N-dimethylformamide (DMF). Then, to obtain DUT-4 and DUT-5, an aluminum precursor (Al(NO₃)₃·9H₂O) was added also in DMF and the mixtures were heated at 180 °C for 24 h in an autoclave. For the preparation of magnetic HKUST-1, Cu(OAc)₂·H₂O was mixed with the previous MNP/ligand solution and altogether refluxed for 12 h under stirring. The composites exhibited high surface areas within the range of reported values (1394, 1346, and 1248 m²·g^–1 for the HKUST-1, DUT-4, and DUT-5 composites, respectively). Remarkably, all the composites showed magnetic properties, since they were collected through an external magnet. In the case of the magnetic HKUST-1 composite, it was characterized under a magnetic field of 1.7 kA·m^–1 at a frequency of 183 kHz, presenting a specific absorption rate (SAR) of 11.1 W·g^–1. Despite the relatively low SAR value, the magnetic HKUST-1 composite showed an accelerated release rate of ibuprofen as the temperature increased under the AMF, representing the first proof of concept of a MOF magnetic nanocomposite demonstrating improved drug release under heating by AMF.

Another example of catalytic application was recently reported by Zamani and colleagues,⁹⁰ developing a magnetic porphyrin-loaded MOF. The Fe₃O₄@CoTHPP@UiO-66 composite, based on the robust microporous zirconium(IV) terephthalate UiO-66 (UiO = University of Oslo; CoTHPP(OAC) = meso-tetrakis(4-hydroxyphenylporphyrinato) cobalt(II)).⁹¹ The resulting nanocomposite (∼400 nm) was prepared by a one-step solvothermal route, first dispersing the MNPs in DMF, then the MOF precursors (ZrCl₄ and terephthalic acid) and the desired porphyrin (CoTHPP), for finally carrying out the solvothermal reaction at 120 °C for 24 h in the presence of glacial acetic acid as the modulator. The magnetic nanocomposite formation was confirmed by PXRD and by the magnetization curve (10 emu·g^–1), presenting a surface area reduced from for the parent UiO-66 (S_BET 732 vs 1380 m²·g^–1). The MNPs permitted an easy, low energy and short-time consumption recovery magnetic separation of the composite that improved its reusability as a catalyst, using it for the epoxidation of olefins and allylic alcohols with a yield of the reaction up to 95% and 5 cycles-reusability.

As shown in these examples, the non-core–shell structures based on uncoated MNPs present a priori an absence of specific MNPs-MOF interactions, leading to some associated issues. For instance, it could lead to the segregation of the MOF formation, obtaining mixtures of pure components and not composites. Also, the uncoated MNPs usually tend to aggregate in the reaction mixture, leading to the formation of heterogeneous composites. In this sense, solvothermal reactions are usually carried out without stirring, generally preventing a good MNPs dispersion. Finally, the reaction conditions for the preparation of the MOFs should be compatible with the MNPs’ stability, limiting the number of suitable MOF structures (generally synthesized under acidic conditions and highly complexant species that could dissolve the MNPs).⁹²

2.3.1.2. Synthesis of Magnetic Composites Using with Acid-Functionalized MNPs

An interesting approach to prevent MNPs aggregation and promote specific MNPs-MOF interactions is the MNPs’ surface functionalization with acidic or other groups, which can be also regulated in terms of the length of the hydrocarbon chain.

In this sense, Schejn et al.⁹³ proposed the addition of citrate-capped Fe₃O₄ nanoparticles to form Fe₃O₄@ZIF-8. The citrate-capped Fe₃O₄ nanoparticles and the 2-methyl-1H-imidazole ligand (HmIM) were dispersed in water. Then, an aqueous solution of the Zn(NO₃)₂ was added, forming the composite at room temperature after only 10 min with a particle size of ∼250 nm. The specific surface area and the pore volume (S_BET ∼ 1856 vs 871 m²·g^–1; 0.71 vs 0.35 m³·g^–1)⁷⁸ decreased for the Fe₃O₄@ZIF-8 with respect to ZIF-8, and was justified by the TEM-based location of the MNPs (∼10 nm) at the MOF surface, blocking the cavities. The PXRD analysis confirmed the formation of the ZIF-8 and the presence of the Fe₃O₄ nanoparticles. The formation of the composite was favored by the presence of carboxylate groups, which can interact weakly with the MOF precursors (imidazolate and metal ions). Additionally, a more complex composite based on ZIF-8 and ZIF 67 (>450 nm) was synthesized based on citrate-capped Fe₃O₄.⁹⁴ The authors introduced the MNPs in the solution of HmIM, adding then the Zn²⁺ solution. After 6 h, HmIM solution was added again. In this way, first, the MOF shell of ZIF-8 was produced, then with Co(NO₃)₂·6H₂O, HmIM, and folic acid (FA) was obtained a second MOF shell (ZIF-67) functionalized with FA (Fe₃O₄@ZIF-8@ZIF-67/FA). TEM images showed that MNPs (∼6 nm) were dispersed in the ZIF-8 structure (∼400 nm), and the ZIF-67 shell (∼50 nm) maintained the same morphology. In the final composite Fe₃O₄@ZIF-8@ZIF-67/FA, the FA produced agglomerated flower-like structures, desired for the DD purpose. The observed PXRD diffraction peaks for the composites were attributed to the MNPs, ZIF-8, and ZIF-67. However, after the drug encapsulation of a model antitumoral drug (quercetin, Q), the PXRD pattern of Fe₃O₄@ZIF-8@ZIF-67/FA/Q was broadened. The surface areas varied from 42 m²·g^–1 for Fe₃O₄ to 1994, 1203, and 259 m²·g^–1 for Fe₃O₄@ZIF-8, Fe₃O₄@ZIF-8@ZIF-67/FA, and Fe₃O₄@ZIF-8@ZIF-67/FA/Q, respectively, indicating first the MOF porosity contribution, then, the presence of the FA functionalization, and, finally, the drug loading.

Certainly, these examples evidence the challenging control of the MNPs position (e.g., core–shell) in the MOF. Indeed, the functionalization over the MNPs surface may not only establish bonds with the MOF precursors but also the MNPs have to promote the growth of the MOF over them as nucleation seeds.

2.3.1.3. Synthesis of Magnetic Composites Using Polymer-Functionalized MNPs

Another interesting approach to accomplish the composite formation is using surface-engineered MNPs in order to avoid their aggregation, favoring their easy dispersion in solution. The surface functionalization has been mostly based on polymers, such as polyvinylpyrrolidone (PVP) or polystyrenesulfonate (PSS), among others, which could present additional functionalities (e.g., −COOH, poly(dopamine)-PDA) that might help the MOF formation.

On this matter, Fang and colleagues⁹⁵ synthesized magnetic MOF nanoparticles based on the ZIF-90 structure, based on zinc and the imidazolate-2-carboxyaldehyde (2-ICA) ligand (S_BET ∼ 1103–1297 m²·g^–1).⁹⁶ The composite was prepared by mixing polyvinylpyrrolidone (PVP)-coated MNPs (Fe₃O₄@PVP, 12 nm) with 2-ICA,= and then immediately pouring the Zn(NO₃)₂ into the solution for finally heating at 90 °C for 18 h. The dried MNP@MOF nanocomposite (∼64 nm) was attracted by an external magnet, determining a decrease in saturation magnetization from 32 to 7 emu·g^–1 in the magnetic nanocomposite due to the MOF presence. Interestingly, the nanocomposites indexed ZIF-90 and Fe₃O₄ peaks in the PXRD pattern and presented a kind of core–shell structure, being several MNPs located in the center of the ZIF-90 particles, as evidenced by TEM.

Another non-core–shell structure was also developed using ZIF-90 and (PDA)-coated MNPs (Fe₃O₄@PDA@ZIF-90).⁹⁷ Briefly, an ethanolic Zn(NO₃)₂ solution was put in contact first with an aqueous solution of Fe₃O₄@PDA nanoparticles, recovering magnetically the MNPs, to then, add the 2-ICA solution. Then, Zn(NO₃), trioctylamine, and 2-ICA solutions were poured in regular intervals until all the precursors’ solutions were added and continuously reacted. Trioctylamine, a tertiary amine with large alkyl substituents, can act as both a deprotonating agent and a surfactant, facilitating the nucleation and growth of the MOFs.⁹⁶ The product was recovered magnetically, having a saturation magnetization of 9.3 emu·g^–1 (vs 22.5 and 17.3 emu·g^–1 for Fe₃O₄ and Fe₃O₄@PDA, respectively). Fe₃O₄@PDA@ZIF-90 exhibited the characteristic peaks of ZIF-90 in PXRD, with however almost no visible peaks corresponding to the MNPs, probably due to the small proportion within the composite. The TEM images clearly showed a non-core–shell structure of about 200 nm, with narrow size distribution and well-dispersion evidenced by DLS, with an agglomeration of the PDA-coated MNPs (∼170 nm) and about 20–30 nm of ZIF-90 growth.

ZIF-8 was also extensively studied here because of its simple and versatile synthesis which allows a fine-tuning of its properties.^78,98 In this regard, Pang and co-workers⁹⁹ proposed a Fe₃O₄@ZIF-8 composite (∼150 nm) based on poly(acrylic acid) (PAA) grafted MNPs (Fe₃O₄@PAA, ∼10 nm). The reaction at room temperature involved an aqueous solution of HmIM and Fe₃O₄@PAA nanoparticles, bearing carboxylate groups on their outer surface to prevent aggregation. Then, Zn²⁺ was added, coordinating with these carboxylate groups and subsequently forming the desired MOF composite, with a saturation magnetization in the range of 0.56–4.35 emu·g^–1, depending on the Fe₃O₄ content. While the high purity of the resulting magnetite-based nanocomposite was supported by PXRD, the TEM images showed the presence of several MNPs grafted at the surface and also embedded into the MOF.

A noteworthy alternative approach based on ZIF-8 was developed by Zhong et al.,¹⁰⁰ presenting a composite called Void nFe₃O₄@Pd@ZIF-8@ZIF-8, featuring an empty internal cavity. In the multistep synthesis, MNPs were initially combined with polystyrene-co-acrylic acid (PS-co-AA) nanospheres to produce Fe₃O₄/PS nanospheres. Then, these nanospheres were dispersed in a MeOH solution containing ZIF-8 precursors at room temperature for 3 h. Subsequently, Pd nanoparticles were encapsulated within Fe₃O₄/PS@ZIF-8 using an impregnation method, resulting in Fe₃O₄/PS@Pd@ZIF-8. An additional shell of ZIF-8 was created using a similar procedure. Finally, the PS core was removed from Fe₃O₄/PS@Pd@ZIF-8@ZIF-8 through DMF treatment to create the internal cavity, leading to the final nanocomposite, Void nFe₃O₄@Pd@ZIF-8@ZIF-8. The crystal structure of the composite was confirmed through PXRD. TEM and SEM images revealed a non-core–shell structure with multiple MNPs homogeneously dispersed within the MOF internal surface rather than forming a single core. Notably, the synthesis method facilitated the controlled distribution of MNPs and subsequent Pd nanoparticles. The sizes of Fe₃O₄/PS, Fe₃O₄/PS@Pd, and Void nFe₃O₄@Pd@ZIF-8@ZIF-8 were approximately 410, 450, and 520 nm, respectively. The porous nature of the material exhibited variations in S_BET, with values of 192, 62, 306, and 523 m²·g^–1 for Fe₃O₄/PS@ZIF-8, Fe₃O₄/PS@Pd@ZIF-8, Fe₃O₄/PS@Pd@ZIF-8@ZIF-8, and Void nFe₃O₄@Pd@ZIF-8@ZIF-8. Indeed, the presence of Pd nanoparticles in the final 30 nm layer affected its porosity, which is partially restored by reducing the solid phase contribution of PS nanospheres. This example represents a novel approach, utilizing a removable template that does not compromise the nanocomposite porosity. The advantage of this template lies in the controlled positioning of MNPs, a feature often absents in other cases, making it an interesting solution.

Another example based on a non-core–shell morphology was reported by Chowdhuri and co-workers,^101,102 developing the Fe₃O₄@IRMOF-3 nanocomposite. The highly porous (S_BET = 1568 m²·g^–1 and pore volume = 1.07 cm³·g^–1)¹⁰³ zinc aminoterephthalate IRMOF-3 (Iso Reticular MOF)¹⁰⁴ was solvothermally formed on MNPs. Briefly, Fe₃O₄ nanoparticles (∼10 nm) were well dispersed in a PVP solution (1:1, DMF and absolute ethanol). Then, Zn(NO₃)₂ and 2-aminobenzene-1,4-dicarboxylic acid (NH₂–H₂BDC) dissolved in DMF were added to the previous Fe₃O₄ solution, heating at 100 °C for 4 h. Nanoparticles of Fe₃O₄@IRMOF-3 (∼65 nm) were observed through field emission scanning electron microscopy (FESEM), although the average particle size in DLS was around 200 nm, probably associated with slight aggregation in solution.¹⁰¹ Furthermore, the normalized saturation magnetization values of the synthesized bare Fe₃O₄ and the nanocomposites were observed to be ∼80 and ∼50 emu·g^–1, respectively.^101,102 More recently, Taghavi et al.¹⁰⁵ reported a similar non-core–shell Fe₃O₄@IRMOF-3 (∼150 nm with MNPs below 50 nm) with a saturation magnetization of about 60 emu·g^–1, making this composite highly promising.

2.3.2. Core–Shell Magnetic Composites

So far, this Review has shown the formation of the MNP@MOF nanocomposites where the particle location is not controlled. However, it is now necessary to explain the great interest in a well-defined core–shell architecture (Figure 2), since locating a single MNP inside (core) a single MOF nanoparticle (shell) is a nice chemical challenge to guarantee the homogeneity of the system, porosity, and intimate interaction between MNPs and MOFs, avoiding the MNPs–MNPs ones.¹⁰⁶ Therefore, it would be ideal for the manipulation of the nanoparticles insertion and the ratio between the MNPs and the MOF precursors content to systematically tune the properties for a precise control over the size, shape, and composition of the nanocomposites.¹⁰⁶ In this section, the more recent examples of core–shell composites will be discussed classifying them as a function of the magnetic core nature (e.g., uncoated, acidic functionalized, amino functionalized, polymer functionalized).

2.3.2.1. Synthesis of Magnetic Composite with Nonfunctionalized MNPs

Despite the issues related to the uncoated MNPs, their use is still considered for the synthesis of core–shell nanocomposites. The reason is evident in the following examples, where the choice of the uncoated-MNPs is related to their easy fabrication via a simple, and low-cost coprecipitation method. The central objective of the forthcoming analysis centers around the optimization of MOF growth surrounding a magnetic core. To this end, two recent examples have been selected, both adhering to a size criterion of approximately 300 nm because it is significant for biomedical applications.

In this regard, a magnetic nanocomposite based on uncoated-Fe₃O₄ nanoparticles (∼260 nm) and magnesium 2,5-dihydroxyterephthalate Mg-MOF-74 was reported.¹⁰⁷ The nanocomposite was solvothermally (125 °C for 5 h) prepared from a suspension of Fe₃O₄ nanoparticles and Mg(NO₃)₂ in a mixture of DMF-ethanol-H₂O (15:1:1), adding then the ligand. The final nanocomposite (∼320 nm) was magnetically collected, identifying both the MOF and the Fe₃O₄ structures by PXRD. However, the BET surface area was much lower than the expected one (265 vs 1250 m²·g^–1 for Mg-MOF-74), which could be related to the presence of dense MNPs and the low shell thickness of the MOF.

On the other hand, a more complex structure, Fe₃O₄@UiO-66@UiO-67/CTAB, was achieved with the cationic surfactant hexadecyltrimethylammonium bromide (CTAB) surface modification.¹⁰⁸ In the first step, ZrCl₄, 1,4-benzene dicarboxylic acid (H₂BDC), and DMF were mixed with uncoated-Fe₃O₄ nanoparticles, adding then acid modulators (HCl and acetic acid) and heated at 120 °C for 24 h. Second, after the recovery, UiO-67 was formed over the Fe₃O₄@UiO-66 composite following the same procedure but with the biphenyl-4-4′-dicarboxylic acid (H₂BPDC) as the organic linker. Finally, the Fe₃O₄@UiO-66@UiO-67/CTAB composite was obtained by introducing it in a CTAB solution. The final material exhibited an irregular morphology that differs from UiO-66 and UiO-67, presenting a dimension of about 60–130 nm. The magnetization loops indicated a superparamagnetic feature and a high saturation magnetization of about 36 emu·g^–1 for Fe₃O₄@UiO-66@UiO-67/CTAB, considering the values of saturation magnetization at room temperature of bulk magnetite (92 emu·g^–1) and maghemite (76 emu·g^–1).^109,110

As was pointed out in the introduction to this subsection, the limited number of reported works of core–shell nanocomposites emphasizes the difficulties for nanoparticles to be the seed for MOF formation. However, with nanoparticles of suitable size, even if uncoated, a synthesis protocol of a small nanoMOF can lead to a single shell of MOF around a single MNP.

2.3.2.2. Synthesis of Magnetic Composite with Acid- or Amino-Functionalized MNPs

As previously explained, carboxylic groups on the surface of nanoparticles are employed to stabilize them in a well-dispersed solution and to increase the affinity of the MOF in order to grow the crystal on the MNP surface. Another alternative, improving stabilization and preventing agglomeration, is the use of amino groups to increase again the chance of a core–shell architecture. In the following subchapter, some examples are reported with both functionalization over MNPs.

In this matter, Chen et al.¹¹¹ prepared magnetic core–shell Fe₃O₄@HKUST-1 (∼50–100 nm) with carboxyl functionalized Fe₃O₄ cores (∼20 nm). The MNPs were dispersed in a mixed solution of DMF/EtOH/H₂O (1:1:1), and PVP was then added as a surface capping agent to promote the core–shell growth together with Cu(OAc)₂·H₂O. Finally, with further metal precursor and trimesic acid, the reaction proceeded for 12 h. The BET surface area of the Fe₃O₄ core (S_BET ∼ 10 m²·g^–1) increased up to 738 m²·g^–1 in the Fe₃O₄@HKUST-1 nanostructure, within the range of the MOF itself (S_BET = 600–1600 m²·g^–1).⁸⁶ Furthermore, they achieved a magnetic fluid composite with high particle content (25.0–45.4 wt %) by introducing Fe₃O₄@HKUST-1 core–shell nanoparticles into a carboxymethylcellulose (CMC) solution.¹¹² According to the literature,^111,112 other Fe₃O₄@HKUST-1 composites were proposed for their good catalytic activities.^113,114 In all the cases, PXRD patterns demonstrated that both phases of Fe₃O₄ and HKUST-1 were present in the composites. The same synthetic protocol, with carboxyl functionalized Fe₃O₄ cores and PVP as a surface capping agent, guaranteed the core–shell Fe₃O₄@HKUST-1 nanocomposite formation, being later further improved with a DMF-free modified version.^113,114

Notably, for the synthesis of several MNP@ZIF-8 nanocomposites, citric acid (CA) has garnered significant interest. For instance, Hou et al.¹¹⁵ prepared a core–shell magnetic ZIF-8 via a solvothermal method with glucose oxidase (GOx) embedded into the composite (Fe₃O₄@ZIF-8@GOx). For the synthesis, CA-modified Fe₃O₄ nanoparticles (CA-Fe₃O₄, ∼100 nm) were suspended and sonicated in an EtOH/H₂O (1:1) solution containing Zn(NO₃)₂ and HCl. Lastly, an EtOH/H₂O (1:1) solution containing HmIM and PVP was added and stirred. The magnetic Fe₃O₄@ZIF-8 (M_S ∼ 48.2 vs 82.2 emu·g^–1 for Fe₃O₄) was easily collected with a magnet, exhibiting a core–shell spherical morphology. Additionally, in the PXRD pattern, the diffraction peaks were consistent with the Fe₃O₄ and ZIF-8 patterns.

Slightly different, Lin and colleagues¹¹⁶ proposed the synthesis of Fe₃O₄@ZIF-8 (∼120 nm) using MNPs (∼6 nm) coated with 3,4-dihydroxyhydrocinnamic acid (DHCA). Prior to the synthesis of the nanocomposite, the MNPs underwent a ligand exchange procedure to substitute oleylamine ligands with DHCA molecules. In this way, their solubility in H₂O was improved, providing also terminal −COOH on the surface of MNPs for coordinating the Zn²⁺ ions. Afterward, in a solution of HmIM, PVP, and Fe₃O₄ nanoparticles, the Zn(NO₃)₂·6H₂O aqueous solution was rapidly poured to afford the final composite. Both FTIR and PXRD determined the presence of Fe₃O₄ and ZIF-8 in the magnetic nanocomposite (M_S ∼ 18.9 vs 44.3 emu·g^–1 for Fe₃O₄). The hydrothermal method maintains the aforementioned advantages; however, the ligand exchange introduced a slighter modification which permitted the synthesis without ethanol.

Recently, further ZIF-8-based nanocomposites were produced, incorporating Fe₃O₄-nanorods (NRs) as the core material and subsequent decoration with Pt for catalytic purposes (Fe₃O₄-NR@ZIF-8/Pt).¹¹⁷ During the synthesis process, oleylamine-functionalized Fe₃O₄-NRs were mixed with ZIF-8 precursors in a MeOH solution at room temperature. The resulting nanocomposite was characterized through the analysis of PXRD and FTIR, confirming the successful growth of the MOF and the presence of the Fe₃O₄-NRs. TEM images provided further insights, revealing the length of the Fe₃O₄-NRs to be approximately 700 nm with a width of around 50 nm. Additionally, the synthesis of ZIF-8 led to the formation of a 30 nm MOF shell surrounding the NRs. Instead, Pt content was determined using Energy-dispersive X-ray spectroscopy (EDS), and the surface area as measured by the BET method decreased from 620 (Fe₃O₄-NR@ZIF-8) to 265 m²·g^–1 (Fe₃O₄-NR@ZIF-8/Pt), as a consequence of the Pt decoration. Notably, this fascinating material, primarily synthesized for catalytic purposes, also exhibited magnetic characteristics, with M_S of 73.9, 58.9, and 50.2 emu·g^–1 for Fe₃O₄-NR, Fe₃O₄-NR@ZIF-8, and Fe₃O₄-NR@ZIF-8/Pt, respectively. Thus, this example once again emphasizes the remarkable versatility of ZIF-8 synthesis, even when there are variations in the composition of the magnetic core. Despite the ZIF-8 advantageous synthetic properties (fast, simple, and versatile), its utilization for bioapplications is accompanied by several limitations. Specifically, it demonstrates low to medium stability in aqueous environments. To enhance its long-term performance, stability, and biocompatibility, as well as to improve its targeting capabilities, additional modifications such as functionalization and surface engineering steps (e.g., GOx) become imperative to address these challenges to fully exploit the potential of ZIF-8 for bioapplications. For the simplicity of the synthesis, in the same way, other magnetic core–shell Fe₃O₄@ZIF-8 composites with size ranges from 180 to 400 nm were synthesized with minor modifications but finalized for different applications, such as water treatment,¹¹⁸ protein separations,¹¹⁹ bioimaging,¹²⁰ drug delivery,¹²¹ and potential theragnostic agents.¹²²

Similarly, an Au@Pt nanoparticle-decorated magnetic Fe₃O₄@UiO-66 composite¹²³ (∼100–300 nm) was constructed. PXRD and FTIR of the Fe₃O₄@UiO-66 confirmed the formation of the composite, exhibiting Fe₃O₄ nanoparticles (∼90 nm) interacting with the MOF in the SEM and TEM images. Recently, in a similar manner, the same composites were also synthesized, reaching particle size of the final product about 200 nm,¹²⁴ smaller than 100 nm,¹²⁵ or even about 16–17 nm.¹²⁶ The focus is on achieving a size below 200 nm, targeting suitable sizes for intravenous nanocarriers.¹²⁷

In another study, core–shell Fe₃O₄@MIL-100(Fe) spheres (∼350 nm) were fabricated using citrate-capped Fe₃O₄ magnetic particles.¹²⁸ The citrate-functionalized Fe₃O₄ particles (∼250 nm) were dispersed in an ethanol solution, first of FeCl₃•6H₂O and then, of H₃BTC at 70 °C for 15 and 30 min, respectively. This mixture acted as a precursor of Fe₃O₄@MIL-100(Fe), which was dispersed into a solution of FeCl₃•6H₂O and H₃BTC under stirring and heated at 70 °C for 24 h for the MOF growing. The magnetic composite exhibited a size of 350 nm, with PXRD and FTIR analyses establishing the presence of both Fe₃O₄ and MIL-100(Fe) phases. In a reduced particle size range, Tregubov et al.¹²⁹ prepared as well Fe₃O₄@MIL-100(Fe) nanocomposite (∼100 nm) just increasing the temperature to 95 °C and keeping it for 12 h. In these cases, the solvothermal synthesis was anticipated with a short step, which created a precursor for the next step. The precursor acted as a seed for the MOF growth, and the MNPs (∼80 nm) were in contact with the metal ions and the organic linkers for 30 min to create the first interactions for the further shell framework growth in the solvothermal procedure. The reasoning behind this is to facilitate the interaction of the MNPs with the precursors of the MOF under stirring, enhancing the dispersion of the Fe₃O₄ nanoparticles. Therefore, MNPs in the static condition of the solvothermal reaction were less encouraged to aggregate.

Instead, Li and co-workers¹³⁰ used Fe₃O₄-NH₂ nanoparticles (∼158 nm) to MW-assisted synthesize Fe₃O₄-NH₂@MIL-101(Fe)-NH₂ core–shell nanocomposites (∼268 nm) since the amino group of MNPs could improve the interaction with the carboxylic groups of the organic linker. The obtained magnetic MOFs were recovered by a magnet (M_S ∼ 20–21 emu·g^–1 vs Twenty-seven emu·g^–1 for Fe₃O₄-NH₂). The efficient MW heating gave rise to a highly homogeneous Fe₃O₄-NH₂@MIL-101(Fe)-NH₂ nanocomposite. Actually, this strategy has been employed in MOFs synthesis in the past decade not only for the short reaction times but also because the high yield and the properties (tuning crystal size) are affected by the specific and almost instantaneous and homogeneous heating.¹³¹ The microwave-assisted method is almost unexplored for the synthesis of MNP@MOF nanocomposites; however, in the near future, we expect an improvement in this method. One of the issues around this approach is the temperature, and consequently, the pressure, in the microwave vessels, because generally fast reactions are correlated to higher temperature with respect to the solvothermal synthesis. Currently, there is a dearth of evidence concerning the influence of radiation on MNPs. Nevertheless, it is worth noting that certain studies have reported evidence of a decrystallization effect in powder magnetite when exposed to microwave radiation for a brief duration of only 5 min near its Curie temperature.¹³² Although this finding does not provide insights into the behavior of MNPs under specific conditions, it could suggest potential limitations for the synthesis of nanocomposites involving MOFs.

2.3.2.3. Synthesis of Magnetic Composite with Polymer-Functionalized MNPs

As previously stated, the core–shell architecture does certainly offer superior control over the shape, morphology, and size of the nanocomposites.¹³³ However, it depends on the addition of a mediator between the magnetic core and the shell growth.¹³³ Examples of these capping agents are once again polymers, facilitating the MOF overgrowth¹³⁴ and also affecting the magnetic properties of the final magnetic nanocomposite.

Based on PVP-coated iron oxide particles, Zhuang et al.¹³⁵ synthesized a 70 nm core–shell composite with a PVP-Fe₃O₄ core (∼50 nm) and a ZIF-8 shell with encapsulated fluorescein. Zhang et al.¹³⁶ also prepared Fe₃O₄@ZIF-8 magnetic core–shell composites, but as microspheres (∼800 nm, M_S = 54.6 vs 63.2 emu·g^–1 of Fe₃O₄) starting from poly(styrenesulfonate, sodium salt) (PSS) functionalized Fe₃O₄ microspheres (∼600 nm) mixed with methanol, Zn(NO₃)₂·6H₂O, and HmIM. Upon heating (50 °C/3h), the resulting composite showed diffraction peaks corresponding to both components, and the core–shell structure was clearly evidenced by microscopy. The aforementioned examples depict two opposing extremities regarding their dimensions. The former exemplifies dimensions that hold greater relevance to biomedical applications, albeit the study fails to thoroughly investigate the material’s magnetic properties and its potential application in biomedical contexts. Conversely, the latter serves as an illustration of a scenario wherein the magnetic core is employed merely as a means of material retrieval in catalytic applications. In addition, the variability in the selection of polymer to facilitate the growth of MOF shell is noteworthy. Specifically, in the former instance, the commonly employed amphiphilic and nonionic PVP was utilized and, in the latter case, the anionic PSS was opted for due to its established efficacy in reducing the surface charge to increasingly negative values (up to −26.9 mV from an initial value of −7.56 mV for unmodified MNPs). This reduction might promote interaction between the metal cation for deposition and subsequent MOF nucleation.

An alternative approach for forming polymer coatings over the MNPs is the use of PDA. For instance, the CoFe₂O₄@PDA@ZIF-8 nanocomposite¹³⁷ evidenced the chelating effect of the PDA. Moreover, a different CoFe₂O₄magnetic core (∼70 nm) was employed, which exhibited a mesoporous structure. The magnetic nanocomposite (36.4 vs 68.5 and 50.3 emu·g^–1 for CoFe₂O₄ and CoFe₂O₄@PDA, respectively) was obtained at room temperature after 30 min, starting from CoFe₂O₄@PDA (∼100 nm) dispersed in a Zn(NO₃)₂ methanol solution and, subsequently, with the continuous dripping of HmIM solution. In the PXRD patterns, the characteristic peaks of the cubic spinel phase of CoFe₂O₄ and the crystalline ZIF-8 could be found. Furthermore, the porosity of the magnetic core increased with respect to the MNPs from S_BET ∼164 to 349.6 m²·g^–1 due to the presence of the MOF. Finally, the architecture and morphology of the composite in TEM revealed a core–shell structure of around 150 nm. The PDA presence was proved to be essential for the formation of the ZIF-8 because, in the absence of the polymer, the interactions between the MNPs and the ZIF-8 do not form a core–shell architecture. Moreover, the negatively charged magnetic core advantageously presents hydrophilic open voids, which has extended the encapsulation capability of the hydrophobic ZIF-8 positively charged shell presented in the other nanocomposites.

A further example is the production of spheres consisting of Fe₃O₄@PDA@Cu₃ (BTC)₂ (∼300–500 nm; S_BET ∼ 161 m²·g^–1).¹³⁸ Specifically, Fe₃O₄@PDA (∼280 nm) were dispersed in an ethanol solution of Cu(OAc)₂·H₂O and trimesic acid and altogether was heated at 70 °C. The composite formation was proved by PXRD. In this interesting work, the interface between the magnetic core and the MOF overgrowth was evidenced by TEM. The PDA was a homogeneous coating over the Fe₃O₄ surface, which stimulate the crystallization of the Cu₃ (BTC)₂ because the metal ions could be coordinated by the phenolic hydroxyl and amino groups of the PDA.¹³⁸ From polydopamine-modified Fe₃O₄ particles, also Deng and colleagues¹³⁹ described a solvothermal method for synthesizing a magnetic MIL-101(Fe) composite. A solution of FeCl₃·6H₂O and H₂BDC in DMF was added to the magnetic Fe₃O₄@PDA particles (∼270 nm), then, and heated at 110 °C for 24 h. The TEM images clearly exhibited a core–shell structure with about 30 nm-thick shell of MIL-101(Fe) on the Fe₃O₄@PDA core. The thin coating of PDA once again acted as an interface between the MNPs and the framework structure.

Similarly, this strategy has been exploited also for Zr-MOF composites.^140,141 In a solvothermal synthesis, Fe₃O₄@PDA particles (∼250–300 nm) were dissolved in a DMF solution containing ZrCl₄ and the different dicarboxylate organic linkers (H₂BDC,^140,141 H₂BPDC,¹⁴¹ and [2,2′-bipyridine]-5,5′-dicarboxylic acid (H₂BPYDC)¹⁴¹). The functional groups (−OH, −NH₂) of PDA can chelate to Zr⁴⁺, enabling MOF growth onto the MNP surface.¹⁴¹ In the H₂BDC case,¹⁴⁰ upon heating at 140 °C for 20 min, a porous core–shell nanocomposite (∼400 nm and S_BET ∼ 216.14 m²·g^–1) with an increase in the thickness of the PDA shell from 40 nm on the MNPs to an additional 47 nm of MOF was observed. The nanocomposite exhibited diffraction peaks consistent with the MOF growth and the MNPs presence. In the other study,¹⁴¹ the mixture solution of Fe₃O₄@PDA and Zr-MOF precursors was heated to 140 °C under stirring for 8 h. The PXRD patterns of all the nanocomposites exhibited peaks for Fe₃O₄ and the MOF’s crystalline structure, confirming the synthesis of Fe₃O₄@UiO-66, Fe₃O₄@UiO-66-PYDC, Fe₃O₄@UiO-67 and Fe₃O₄@UiO-67-BPYDC. The TEM images of all magnetic nanocomposites showed a core–shell composite with a shell thickness of the MOF of about 40–75 nm, and an average size of about 280–300 nm for the nanoparticles.¹⁴¹

As shown in the described studies, PDA coating of MNPs has provided functional groups (−OH, −NH₂) as anchoring points for Cu-, Fe-, and Zr-MOFs. However, the polymer coating also affects the Fe₃O₄ core’s magnetic properties together with the MOF shell. Indeed, in general, the bare MNPs have a saturation magnetization which decreases with the surface coating and, even more, with the MOF shell. Nevertheless, the magnetic properties are tunable with an optimized synthesis. In fact, the in situ core–shell method relies on control of the size, as well as the shape and morphology control. Therefore, a control of the thickness of the MOF shell as well as the starting magnetic nanocore can tune the nanocomposites’ properties.

In recent years, other researchers proposed the synthesis of Fe₃O₄@UiO-66 or Fe₃O₄@UiO-66-NH₂ nanocomposites employing PAA. Zhao and colleagues¹⁴² directly dispersed Fe₃O₄ nanoparticles (treated with PAA and urea, with a diameter size of ∼150 nm) into a DMF solution of the synthetic precursors of UiO-66 (ZrCl₄, NH₂-H₂BDC). The core–shell Fe₃O₄@UiO-66 showed a UiO-66 shell of about 25 nm thickness, while the composite size was about 240 nm.¹⁴² The M_S of both Fe₃O₄@UiO-67-BPYDC¹⁴¹ and Fe₃O₄@UiO-66¹⁴² exhibited similar values within the range of approximately ∼50 emu·g^–1 (vs ∼70–75 emu·g^–1 for Fe₃O₄). However, it was noteworthy that only the latter has undergone adequate dimension to facilitate its investigation as a theragnostic system, as elaborated on subsequently (see section 3.3: Theragnostics).

2.3.2.4. Synthesis of Magnetic Composite with Carbon- or SiO₂-Functionalized MNPs

In addition to the aforementioned methods, there are two other possibilities. In the first case, MNP@MOFs can be synthesized employing Fe₃O₄@carbon (Fe₃O₄@C) nanoparticles. Besides the role in the stabilization of the magnetic core in the reaction mixture, the porous carbon shell can improve the imaging properties of the nanocomposites due to the carbon dots fluorescence. On this matter, He and co-workers¹⁴³ synthesized ZIF-8 nanocomposites using Fe₃O₄@C nanospheres as the core. The nanocomposite (∼220 nm) was simply synthesized at 60 °C for 1 h by adding Fe₃O₄@C nanospheres (∼190 nm) in a methanol solution containing Zn(NO₃)₂·6H₂O and HmIM.

In the second strategy, silica-capped MNPs (Fe₃O₄@SiO₂) are employed, with the SiO₂ capping conferring lower cytotoxicity to the MNPs.¹⁴⁴ Moreover, the thickness of the silica surface over the Fe₃O₄ core can be easily tuned,¹⁴⁵ and then, the magnetic responsivity varied as well as the porosity properties related to the mesoporous silica.¹⁴⁴ Additionally, through different silanization agents, different functionalities can be included to promote the interactions with the MOF precursors. In the field of MNP@MOF nanocomposites, core–shell structures of Fe₃O₄@SiO₂@MIL-100(Fe) were fabricated^146−148 by the reaction mixture of Fe₃O₄@SiO₂ (∼360,¹⁴⁶ 50,¹⁴⁷ 20¹⁴⁸ nm) with the MIL-100(Fe) precursors under reflux at 100 °C for 8 h. In all these composites (∼440,¹⁴⁶ 50,¹⁴⁷ 50¹⁴⁸ nm), the PXRD patterns confirmed the MIL-100(Fe) structure and the Fe₃O₄ phase.¹⁴⁶ In this case, the detection of the magnetic component into MIL-100(Fe) could be recognized. It was observed a magnetic saturation decrease from 82.5 emu·g^–1 for the MNPs to 30 emu·g^–1 in the nanocomposite due to the SiO₂ and MIL-100(Fe) shells, permitting an easy magnetic separation. However, the identification of the iron content by EDS did not allow to discriminate between the Fe₃O₄ and the MIL-100(Fe), distinguishing the Fe₃O₄ core from the MOF shell only by TEM.¹⁴⁶

Instead, Jia et al.¹⁴⁹ presented a thermoresponsive polymer, poly(N-isopropylacrylamide) (PNIPAM), tethered to Fe₃O₄@SiO₂@MOF core–shell magnetic nanospheres. The magnetic composite Fe₃O₄@SiO₂@UiO-66-NH₂ was fabricated from Fe₃O₄@SiO₂ nanospheres (∼200–300 nm) obtained by carboxylate-terminated reaction with succinic anhydride and (3-aminopropyl)triethoxysilane (APTES). The magnetic nanosphere were ultrasonically mixed with a DMF solution containing the UiO-66-NH₂ precursors (ZrCl₄, NH₂-H₂BDC and acetic acid). Fe₃O₄@SiO₂@UiO-66-NH₂ (MOF shell ∼ 30–50 nm) was obtained at 130 °C for 4 h under stirring, and after drying, the composite underwent to a subsequent reaction in chloroform using PNIPAM-NHS at 60 °C for 24 h. The Fe₃O₄@SiO₂@UiO-66-NH₂–PNIPAM nanospheres reached dimensions around 350–450 nm, presenting magnetic properties (M_S ∼ 45.60 emu·g^–1) with S_BET ∼ 262 m²·g^–1 and indexing the PXRD pattern to UiO-66-NH₂ and Fe₃O₄. In the same manner, Yang et al.¹⁵⁰ fabricated Fe₃O₄@SiO₂@UiO-67 by dispersing Fe₃O₄@SiO₂ in a ZrCl₄ solution for the complexation of Zr⁴⁺ and adding then H₂BDC and glacial acetic acid (heating at 120 °C for 24 h). The resulting Fe₃O₄@SiO₂@UiO-67 nanocomposites (224–258 nm; MOF shell thickness ∼ 20 nm) were characterized by SEM and TEM, and compared with Fe₃O₄ (∼170 nm) and Fe₃O₄@SiO₂ (∼208–232 nm). The composite exhibited a saturated magnetization lower than that of Fe₃O₄ (61.0 emu·g^–1) and Fe₃O₄@SiO₂ (47.3 emu·g^–1), specifically 20.9 emu·g^–1. Although, as expected, the shell reduced the magnetic properties, the MOF presence was there to influence its positive features, such as the porosity and biocompatibility.

The most evident “bottle around ship” strategy was reported by Huang and colleagues,¹⁵¹ Fe₃O₄@SiO₂@Cu(OH)₂, where the self-template shell of Cu(OH)₂ over Fe₃O₄@SiO₂ nanoparticles (∼15 nm) promoted the conversion of Cu(OH)₂ into HKUST-1. Briefly, Fe₃O₄@SiO₂@Cu(OH)₂ in a water–ethanol solution of the organic linker (H₃BTC) gave rise to core–shell nanostructures (MOF shell ∼ 5–10 nm) of Fe₃O₄@SiO₂@HKUST-1 at room temperature after stirring for 12 h. This method was not so different from what was previously reported, the Cu ions being able to interact with the magnetic core before the MOF growth. The Fe₃O₄@SiO₂@Cu(OH)₂ precursor acted as a seed for the crystallization of the MOF, favoring a core–shell architecture.

2.4. Layer-by-Layer Strategy

The layer-by-layer (LbL) strategy has the purpose to control the crystal growth of the MOF over the MNPs. The synthetic strategy is a step-by-step sequential repeated cycle of immersion in solutions of the metal precursor and solutions of organic ligand¹⁵² or MOF precursors solutions, first proposed in 2007 for HKUST-1.⁶⁵ The protocol proposed by Shekhah was et al. extended to different MOF structures.¹⁵³ Herein, we will present the extension of this strategy for the synthesis of magnetic composites, where the first layer of the MOF is over a MNP and the thickness of the MOF shell is controlled by the number of repeating cycles performed, as represented in Figure 6.

Figure 6.

Schematic representation of the layer-by-layer (LbL) strategy.

2.4.1. Synthesis of Magnetic Composites with Acid- Or Thiol-Functionalized MNPs

As was pointed out earlier, MNP functionalization has been employed for several advantages. In the LbL method, Ke and co-workers¹⁵⁴ proposed magnetic core–shell spheres of Fe₃O₄@HKUST-1 or Fe₃O₄@MIL-100(Fe), subsequently, numerous works^155−169 followed the same synthetic protocol with mainly a carboxylic functionalization over the magnetic core or thiol groups. In general, COOH-functionalized Fe₃O₄ MNPs ranging from 20 to 500 nm were dispersed in solutions of Cu(CH₃COO)₂.·H₂O or FeCl₃·6H₂O precursors, respectively for Fe₃O₄@HKUST-1 or Fe₃O₄@MIL-100(Fe) composites. After 15–30 min and a magnetic recovery of the Fe₃O₄ nanoparticles, the organic linker solution (H₃BTC) was mixed with the MNPs for 30 min at 25, 40, or 70 °C. These two steps formed a cycle to coat Fe₃O₄ with a layer of MOFs and each cycle was subsequently repeated several times. Generally, the diffraction peaks for the samples Fe₃O₄@HKUST-1 in the PXRD pattern matched well with those of both Fe₃O₄ and crystalline HKUST-1.^{154−156,162} In the case of SEM and TEM investigation for Fe₃O₄@HKUST-1, the images showed a spherical-shaped morphology with a core–shell structure with narrow size distribution and uniform dispersion (ranging from 210 nm to 1.5 μm depending on the number of assembling layers).^{154−156,162} The magnetic Fe₃O₄@HKUST-1 composites exhibited BET surface area variable depending on the number of cycles employed, with values from 57 to 668 m²·g^–1; as expected the surface area increases with a larger number of MOF layers.^{154−156,162} Moreover, the magnetic properties and separability were also tested. For instance, the core–shell microspheres desired in the selective removal of Hg²⁺ and Pb²⁺ were separated in a few seconds in an aqueous solution by placing a permanent magnet near the glass bottle.¹⁶² The saturation magnetization varied from ∼14 to 43 emu·g^–1 and all core–shell MNPs exhibited superparamagnetic behavior at room temperature, which also depicted the strong magnetic response to an AMF.

In the case of Fe₃O₄@MIL-100(Fe),^{154,157−161,163−169} the diffraction peaks for the samples were consistent with the crystalline phases of Fe₃O₄ and MIL-100(Fe). However, the intensities of MIL-100(Fe) patterns in some diffractograms were very weak, which was correlated to the low thickness of the MOF shell.^158,161,166 By SEM and TEM, the magnetic composites reported a core–shell structure, with an average size dependent on the diameter of the MNP core and also on the number of layers, varying from 150 nm up to 1 μm.^{154,157−161,163−169} The BET surface areas of Fe₃O₄@MIL-100(Fe) increased with increasing assembly cycles, ranging from ∼36 to 899 m²·g^–1 in the different composites reported.^{154,158−161,164−169} This general trend was due to the decreasing contribution of nonporous Fe₃O₄ MNPs to the total mass of the magnetic core–shell nanoparticles.¹⁵⁸ Furthermore, the majority of these Fe₃O₄@MIL-100(Fe) presented magnetization saturation values in the range of ∼20–56 emu·g^–1.^{154,158,160,163−169} It was observed either the presence of a magnetic hysteresis loop or no obvious remanence or coercivity at 25 °C. Therefore, the magnetic composites possessed ferromagnetic or superparamagnetic features and, for the majority of them, it was observed a simple magnetic separation in the solution media through a magnet in a few seconds. The great number of studies on MIL-100(Fe) based composite was related to the versatility of MIL-100(Fe), which had good in vitro and in vivo biocompatibility as well as important drug loadings for biomedical applications, and, moreover, a potential application for separations through strong coordination with guest molecules.¹⁶⁰

Furthermore, the LbL strategy has also been proposed for other composites. For instance, Zheng et al.¹⁷⁰ developed a Fe₃O₄@ZIF-8 core–shell structure (∼530 nm). The synthetic protocol consisted of citrate-Fe₃O₄ (∼390 nm) solution mixed with Zn(NO₃)₂·6H₂O and HmIM, heating later at 70 °C for 20 min and, separating the product with a magnet. The thickness of the ZIF-8 shell was increased by repeating the above process several times. Likewise, also Liu et al.¹⁷¹ reported the fabrication of core–shell Fe₃O₄@ZIF-8. Both studies confirmed the Fe₃O₄@ZIF-8 formation by PXRD, indexing the diffraction peaks of Fe₃O₄ and crystalline ZIF-8. The specific surface areas were about 1075 m²·g^–1, lower than that of the isolated ZIF-8 (S_BET ∼ 1709 m²·g^–1)¹⁷¹ as an effect of the Fe₃O₄ core on the formation of ordered microporous. On the other hand, the effect of the shell on the magnetic core was translated into a saturation magnetization value of 14.38 emu·g^–1.¹⁷⁰ A similar procedure was followed also for the complex multifunctional system, Fe₃O₄@PAA/AuNCs/ZIF-8 (∼130 nm).¹⁷² In this case, oleic acid (OA)-capped MNPs (∼20 nm) underwent, first, a polymer coating process with PAA, and then the gold nanoclusters (AuNCs) were integrated into the synthetic step for the MOF growth to obtain a theragnostic agent that combines multiple capabilities for cancer treatment.

Another MOF widely used for this strategy was UiO-66-NH_2. In this sense, Chen and co-workers¹⁷³ reported a core–shell Fe₃O₄@UiO-66-NH₂. During the synthesis, Fe₃O₄ nanoparticles (∼200–300 nm) functionalized with carboxylic moieties were immersed, first, in the metal node [Zr₆O₄ (OH)₄]¹²⁺ precursor solution, and then, in the organic linkers (NH₂–H₂BDC) solution, respectively for 15 and 20 min (Figure 7). Before each subsequent immersion step, MNPs were recovered with a magnet and washed. After 20 cycles, the MOF shell was only around 13 nm, leading to a composite size of about 300–350 nm. The surface area increased for the magnetic Fe₃O₄ core with the MOF shell (S_BET ∼ 11 vs 76 m²·g^–1, respectively), far from the high value of the isolated UiO-66-NH₂ (S_BET ∼ 735 m²·g^–1). The crystal growth was however not uniform throughout the magnetic iron oxide core. Therefore, the final morphology and structure may not justify the enormous effort and highly time-consuming of the method.

Figure 7.

Schematic illustration for the layer-by-layer fabrication of core–shell Fe₃O₄@UiO-66-NH₂. Reproduced from ref (173). Copyright 2019 American Chemical Society.

Thus, the synthesis of materials using the LbL technique offers significant advantages in terms of adaptability in the final structure and consequent properties; therefore, in the optimization of the nanocomposite performance in diverse fields, offering a highly customizable approach to achieve desired applications. This versatility arises from the control of the number of synthetic cycles, directly correlated to the thickness of the resulting MOF, enabling adjustments to the porosity and magnetization characteristics. However, in contrast to the core–shell in situ formation of the MOF in the presence of MNPs, the LbL method is a more time-consuming procedure, with complex scalability.

2.4.2. Synthesis of Magnetic Composites with Polymer-Functionalized MNPs

In the LbL approach, for instance, Li et al.¹⁷⁴ presented a core–shell Fe₃O₄@IRMOF-3, through the growth promotion of the MOF modifying the surface of the MNPs with PVP. In brief, the synthesis included the mixture of a dissolution of Zn(NO₃)₂ and NH₂–H₂BDC in DMF, with PVP in DMF:EtOH (3:2) and Fe₃O₄ nanoparticles, heating then at 100 °C for 4 h under vigorous stirring. All the previous steps were repeated several times. The PXRD patterns possessed diffraction peaks assigned to both Fe₃O₄ and the MOF. Also, it was demonstrated that a certain dosage of the polymer not only stabilized the MNP but also favored the crystalline growth of the MOF. In TEM images, after three cycles, the spherically shaped Fe₃O₄ nanoparticles (∼200–500 nm) were embedded in micrometric IRMOF-3 crystals, with a BET surface area and pore volume of 238 m²·g^–1 and 0.31 cm³·g^–1. The saturation magnetization values of Fe₃O₄ decreased from 78.5 to 13.5 emu·g^–1, in the case of the composite formation due to the presence of the MOF layer. In this example, the synthesis involved uncoated-MNPs, where the presence of a polymer was necessary to favor the dispersion of the magnetic core and the consequent MOF growth.

Instead, Miao and colleagues¹⁷⁵ synthesized a core–shell magnetic Fe₃O₄@P4VP(poly(4-vinylpyridine))@MIL-100(Fe) composite. In this example, the PAA-functionalized Fe₃O₄ where involved in a polymer-shell formation, Fe₃O₄@P4VP, to favor the interactions between pyridine and Fe³⁺. Then, Fe₃O₄@P4VP nanospheres (magnetic core ∼200 nm and polymer shell ∼38 nm) were dispersed in an ethanol solution, first, of FeCl₃·6H₂O for 15 min and collected with a magnet; subsequently, they were dispersed in an H₃BTC ethanolic solution and stirred at 70 °C for 30 min. These steps were repeated, giving a composite confirmed in the PXRD pattern. In the HR-TEM, the images clearly showed a 200 nm-diameter magnetic core with a shell thickness of the polymer of 38 nm and the outer MIL-100(Fe) thickness ranging between 15 and 90 nm, depending on the number of assembling cycles (from 5 to 20 cycles). The magnetization saturation values decreased from 73.90 emu·g^–1 for Fe₃O₄-(PAA) to 47.35 and 28.21 emu·g^–1, respectively for Fe₃O₄@P4VP and Fe₃O₄@P4VP@MIL-100(Fe). These results indicated that the materials exhibited a strong magnetic response.

2.4.3. Synthesis of Magnetic Composites with SiO₂-Functionalized MNPs

The versatility of the SiO₂-functionalized MNPs to obtain a core–shell structure via the LbL method was only proposed in the work conducted by Jiang et al.¹⁷⁶ A sophisticated nanocomposite consisting of two distinct 3D MOF structures based on the same ligand, namely, SiO₂@Fe₃O₄@Yb-MOF@Nd-MOF, was proposed, involving the SiO₂@Fe₃O₄ as a core template for the successive synthesis of multiple MOF layers. In the first synthetic step, the SiO₂-functionalized MNPs were added to a solution containing ytterbium(III) acetate at 80 °C for 5 min in a DMF/H₂O mixture. Subsequently, the resulting mixture was centrifuged to separate the precipitate from the supernatant. In the second step, the separated precipitate was once again suspended in a solution containing terphenyl-3,4″,5-tricarboxylic acid (H₃L) ligand. This step followed a similar procedure to that of the first step. Moving on to the third step, a solution of neodymium(III) trichloride was employed, following the same sequence as in step 1. Finally, in the fourth step, the process from step 2 was repeated. All these complex steps were then cyclically repeated 14 times. Based on TEM images, the MNPs’ core could be identified, enveloped by a silica layer with a thickness < 6.5 nm. Additionally, sequential layering of MOFs ranging from 9.86 to 21.63 nm was observed, resulting in a total nanocomposite size of approximately 150–200 nm. This innovative example introduces in the LbL method the alternating arrangement of two MOFs that share the same ligand. The resulting structure appears to be relatively complex and warrants further investigation, particularly concerning the magnetic properties of the core and the surface area characteristics, as well as the variability of the properties increasing the cycle number.

3. Applications of Magnetic Metal–Organic Framework Composites in the Biomedical Field

As mentioned in the Introduction, combining MOFs as promising nanocarriers and MNPs to provide potential imaging, targeted release, and hyperthermia, makes the MNP@MOF nanocomposites particularly interesting in diagnosis, therapy, and theragnostics. The most recent developments will be covered in this section, categorized in therapy, MRI, and theragnostics, comprising nanocomposites with sizes below 500 nm, and analyzing the potential of these systems for their real application.

3.1. Therapy

The administration of therapeutic agents by nanocarriers has been developed to minimize toxicity and side effects, increase the efficacy avoiding early clearance, and ensuring a progressive and located drug release within the active sites.^177,178 Since 2006 and 2010, when micrometric⁸ and nanoscaled⁶ MOFs were originally proposed as DDSs, great advances have been achieved in this exciting topic.^179,180 In particular, this section will describe in detail the use of MNP@MOF nanocomposites as DDS, providing representative examples. Finally, as far as we know, the only work describing magnetic nanocomposite as magnetic hyperthermia (MHT) agents will be presented, as a nice combination of MHT and DDS to further improve the therapeutic efficacy.

Based on one of the most used families of MOFs for bio applications, as a proof of concept, the anti-inflammatory drug ibuprofen (IBU) was successfully encapsulated in a γ-Fe₂O₃@MIL-53(Al) nanocomposite, reaching a drug loading of 110 mg·g^–1 with a long progressive release in PBS at 37 °C (7 days).⁷⁷ Considering the potential toxicity of this Al-based MOF, the IBU was successively loaded in a magnetic LbL composite (M_S ∼ 50.69 and 20.42 emu·g^–1, after 40 and 20 cycles, respectively) based on the biocompatible MIL-100(Fe) (Fe₃O₄@MIL-100(Fe)),¹⁵⁸ achieving higher IBU loadings (310 mg·g^–1) associated with a greater porosity. In this exploratory research, the IBU release under simulated physiological conditions (PBS, pH 7.4) took place for about 35% in the first 2 h, then it slowed down and was completed in 70 h. These works evidenced the possibility to encapsulate drugs into magnetic nanocomposites. However, these preliminary studies only tested the drug encapsulation and the release without further practical considerations (e.g. biocompatibility, in vitro and in vivo therapies).

In recent times, significant effort has been predominantly placed on the encapsulation of more challenging antitumoral drugs. As far as cancer therapy is concerned, a large number of studies have reported the use of the convenient fluorescent antitumoral drug doxorubicin (DOX), widely used in clinics to treat a variety of human diseases, including Hodgkin’s lymphoma, leukemia, multiple myeloma, breast cancer, osteosarcoma, ovarian cancer, and lung cancer.¹⁸¹ This chemotherapeutic product was granted FDA approval as the first nanodrug in 1995 and, at present, is known under the brand name Doxil, a liposomal formulation.¹⁸² Even if the DOX hydrochloride salt formulations on the market have a high therapeutic index and high efficacy against a variety of solid tumors, they are also associated with significant side effects including heart damage, typhlitis, cardiac arrhythmias, nausea, and vomiting.¹⁸¹ Thus, DOX delivery through nanocarriers has garnered considerable interest. Within the magnetic MOF-based nanocomposites, DOX was encapsulated in Fe₃O₄-NH₂@MIL-101(Fe)-NH₂ (∼140–330 nm; M_S ∼ 20.47–21.32 emu·g^–1 vs 27.67 emu·g^–1 for Fe₃O₄-NH₂).¹³⁰ Its DOX loading capacity was optimized by using different MNP:MOF ratios, reaching the highest loading (360 mg·g^–1) at 1:1 molar ratio due to the improved porosity (S_BET = 96 m²·g^–1 vs 12–88 m²·g^–1). The relative amount of dense MNPs in comparison to the pure MOF elucidates the significantly diminished porosity of the nanocomposite (pure MOF reaching 1800 m²·g^–1).¹⁸³ The drug release, dependent on pH, was faster under acidic media due to MOF degradation, simplifying targeted DOX release in acidic cancer cells. Nevertheless, 37–61% of DOX release from the composite in simulated body fluid (SBF) at pH 7.4 occurred in 48 h, more gradually than pristine MIL-101(Fe)-NH₂. Furthermore, cytotoxicity tests pointed out the biocompatibility of the nanocomposite and the applicability of the DOX-loaded composite as a DDS. In this sense, the biocompatibility and controlled DOX release in an acidic environment were further guaranteed by coating a Fe₃O₄@Fe-MOF composite with hydroxyapatite (HAp),¹⁸⁴ although HAp is present in the body at the hard tissue level. In this case, the DOX capacity was 53 and 75 mg·g^–1 in Fe₃O₄@Fe-MOF and Fe₃O₄@Fe-MOF@HAp, respectively. Indeed, the HAp-coating not only increased the drug cargo but also contributed to a more gradual release (at pH 7.4, 25 vs 46% after around 11.7 h from the HAp-coated and noncoated systems, respectively). Finally, cytotoxic assays confirmed the biocompatibility of both composites and the selective DOX effect on HeLa cells.

Another nanocomposite that was employed for DOX release was based on ZIF-8. In this study, DOX was encapsulated in the Fe₃O₄@ZIF-8 nanocomposite (∼70–100 nm) with a content of about 120 mg·g^–1 and a loading efficiency of 12%.¹⁸⁵ The release was proven to be controlled over time without a burst effect, and its in vitro biosafety was demonstrated on the hepatocarcinoma cell line (MHCC97H). Similarly, in the biocompatible Fe₃O₄@ZIF-8 nanocomposite (∼180 nm, M_S ∼ 18.6–37.2 vs 48.1 emu·g^–1 for Fe₃O₄), developed by Chen and co-workers,¹²¹ the DOX encapsulation was enhanced (330 mg·g^–1, 76.6% of loading efficiency). In this latest example, an elevated concentration of DOX during encapsulation may have increased loading efficiency. Plausibly, it is due to a surplus of DOX adsorbed in the outer surface due to the electrostatic interaction between the negatively charged carboxylate groups and positively charged DOX molecules.¹⁸⁶ Additionally, the strong coordinative affinity of the C–O and C=O groups may contribute to the formation of coordination bonding of Zn²⁺-DOX in aqueous solutions.¹⁸⁷ The drug release in PBS after 48 h was higher at pH 5.5 (63%) than at pH 7.4 (33%), showing again a pH dependence that was in line with the acidic environment in the cancer cell. Moreover, the resulting Fe₃O₄@ZIF-8 nanocomposite evidenced photothermal effects under laser irradiation (808 nm), showing a selective cancer cell death for the Fe₃O₄@ZIF-8-enriched area. This study investigated the magnet-targeted photothermal effect of the composites in a preliminary manner. Prior studies indicated that Fe₃O₄ nanoparticles exhibit an outstanding photothermal effect,¹⁸⁸ hence, this work showed that the formation of composites does not impede cancer treatment under these laser conditions. Furthermore, the possibility of evaluating promising combined antitumoral therapies, considering DOX-encapsulated Fe₃O₄@ZIF-8, warrants further investigation.

Likewise, a magnetic composite based on the UiO-66-NH₂ material was studied as DDS for the anticancer DOX, proposing Fe₃O₄@UiO-66-NH₂ (∼150 nm) with hybridization over a layered 2D material, the graphdiyne (GDY).¹⁸⁹ In the resulting complex nanocomposite (∼250 nm; Fe₃O₄@UiO-66-NH₂/GDY), the macroporous structure of the GDY offered an ulterior surface for drug uptake, resulting in a high DOX loading content of 438 mg·g^–1. Once more, the composite demonstrated a pH-dependent DOX release (after 36 h, 49 vs 34% released at pH 5 and pH 7.4, respectively), negligible cytotoxicity, and efficient endocytosis-mediated drug carrier uptake in HeLa cells. Notably, the antitumor activity of the DOX-loaded Fe₃O₄@UiO-66-NH₂/GDY nanocomposite was evaluated in vivo in BALB/c-nu mice, displaying no obvious toxicity with promising tumor-targeting and -inhibition when compared with the free DOX (77.8 vs 27.7%).

As anticipated in the introduction of this section, magnetic nanocomposites present also a strong potential in the development of MHT oncological therapy. MHT consists of increasing the temperature (42–46 °C) in a target tumoral tissue by using nanoheating probes (i.e., MNPs) under an AMF in the kHz radiofrequency range.^190,191 In clinical use, for patient safety, a frequency of 100 kHz and a magnetic field amplitude of 15 kA·m^–1 is generally employed.^190,191 In this regard, the only reported example so far was Fe₃O₄@PDA@ZIF-90 with an average particle size of about 200 nm, originally reported for combined MHT and chemotherapy.⁹⁷ This magnetic nanocomposite (M_S ∼ 9.2 vs 22.5 and 17.3 emu·g^–1 for Fe₃O₄ and Fe₃O₄@PDA, respectively) exhibited a good increase of temperature from 30 to 45.6 °C under an AMF at 409 kHz and 14.3 kA·m^–1 for 20 min. In this experiment, the frequency of the AMF is 4-fold greater than the value for clinical use. For comparison, under the same conditions, Fe₃O₄ and Fe₃O₄@PDA controls led to a temperature increase up to 77.5 and 49.4 °C, not adapted to safe physiological values. Furthermore, the DOX-loaded Fe₃O₄@PDA@ZIF-90 composite (160 mg·g^–1, with a loading efficiency of 80%) demonstrated a faster release as the acidity increases (PBS pH 7.4, 6.0, and 4.5). This fact, together with the magnetic heating, could favor the DOX release at the tumor level, combining chemotherapy and MHT. In fact, for DOX-loaded Fe₃O₄@PDA@ZIF-90, cell death was enhanced under an AMF, demonstrating a hyperthermia therapy in combination with chemotherapy for only the drug-loaded nanocomposites (i.e., cell viability without AMF, with AMF once and twice, respectively, for the composite = 110, 60 and 40% and DOX-loaded composite = 80, 30 and 10%).

Table 1 summarizes all the examples developed in this section. Note here that DOX has been mainly selected as the active ingredient, mostly as a proof of concept due to (i) its dimensions (around 15.4 Å), compatible with the pore size of various MOFs;¹⁹² (ii) its fluorescence, facilitating its quantification and intracellular tracking (confocal microscopy); and (iii) its high efficacy against several tumors, despite severe drawbacks (e.g. cardiotoxicity,^193,194 self-association tendency in aqueous solution,^195,196 drug resistance¹⁹⁷). High drug loading ranging from 75 to 438 mg·g^–1 (loading efficiency 12–96%) was obtained with a pH-dependent release, emphasizing its promoted delivery under an acidic microenvironment favored in tumor cells, lysosomes, or endosomes (pH ∼ 4.5–7.8) as opposed to the pH of normal cells or blood (pH ∼ 7.4).³⁰ The stimuli-responsive drug delivery as an acidic response has been correlated to the degradation of the framework, where pH-sensitive bonds (amine, imide, or carboxylates) may lead to the release of the cargo.^179,198

Table 1. Comparative Overview of Magnetic Nanocomposites for Therapy^a.

MNP@MOF	OL	MP for MOF	MNP@MOF size [MNPs] (nm)	MS (emu·g–1)	drug	DL (mg·g–1)	LE (%)	TCL	ref
γ-Fe2O3@MIL-53(Al)	H2BDC	Al3+	not reported (NR)	6.1	IBU	110	NR	–	(77)
Fe3O4@MIL-100(Fe)	H3BTC	Fe3+	90–150 [20]	50.7 (20 layers) and 20.4 (40 layers)		310	NR	–	(158)
Fe3O4-NH2@MIL-101(Fe)-NH2	NH2-H2BDC	Fe3+	140—330 [140]	20.5–21.3	DOX	360	NR	HeLa	(130)
Fe3O4@Fe-MOF@HAp	H3BTC	Fe3+	400 [300]	34		75	NR	HeLa	(184)
Fe3O4@ZIF −8	HmIM	Zn2+	70–100 [9]	–		120	12	MHCC97H	(185)
Fe3O4@ZIF-8	HmIM	Zn2+	180 [120]	18.6–48.1		330	77	HeLa	(121)
Fe3O4@UiO-66-NH2/GDY	NH2-H2BDC	Zr2+	250 [<150]	21.6		438	NR	HeLa	(189)
Fe3O4@PDA@ZIF-90	2-ICA	Zn2+	200 [140–160]	9.2		160	80	HeLa	(97)

^aOL = organic linker, MP = metal precursor, DL = drug loading, LE = loading efficiency, TCL = targeted cell line.

Furthermore, the mean values of the M_S are disclosed in Table 1, encompassing a range from 6.1 to 51 emu·g^–1. The M_S plays a crucial role in the context of MHT, as the heating efficiency, affected by the specific loss power (SLP), is directly proportional to its square,¹⁹⁹ requiring high M_S values to achieve elevated SLP.¹⁹⁹ In this sense, M_S is influenced by factors such as the size and shape of the MNP.¹⁹⁹ Although also depending on experimental conditions (e.g. strength and frequency of applied AMF), few theoretical²⁰⁰ and experimental²⁰¹ studies determined an optimal size of MNPs for a maximum heating efficiency of around 15–20 nm. The presented examples have predominantly larger MNPs size than the ideal range for MHT. The most favorable outcome is observed in the nanocomposite with MNP size of approximately 20 nm,¹⁵⁸ yet a considerable number of them still exhibits a commendable MS value. Indeed, the value of MNPs generally tends to reach the bulk value and the typical range of 30–50 emu·g^–1 can be regarded as a favorable outcome.²⁰² Notably, MHT was solely validated in one single case, indirectly determining the nanocomposites effect on the cell viability.⁹⁷ Even further, this nanocomposite reported a low M_S (9.12 emu·g^–1), indicating that this parameter is not sufficient to evaluate the MHT efficacy. As a matter of fact, it is crucial to acknowledge that determining the optimal conditions for MHT implementation can be a complex endeavor, necessitating intricate instruments (e.g., superconducting quantum interference device (SQUID), alternating current (AC) magnetometer, and magnetic nanoheating device) and methodologies (e.g., combined magnetic and in vitro techniques). It is worth noting that most of the referenced papers are confined to in vitro investigations, with merely one study presenting in vivo assessments, raising pertinent inquiries concerning the clinical feasibility and applicability of these magnetic nanocomposites.

3.2. MRI

This section will present the principal findings of the current investigation on magnetic MOF nanocomposites as contrast agents for MRI, as an interesting noninvasive diagnostic technique in soft tissues, characterized by unlimited signal penetration depth, lack of ionizing radiation, and wide clinical applicability.^203,204 The contrast agents improve the image contrast because they decrease the longitudinal or transverse proton relaxation times (T₁ and T₂, respectively) of the hydrogen nuclei of the water molecules present in our tissues.^203,204 In the last years, it has been proposed to replace the traditionally used T₁-weighted MRI contrast agent gadolinium (Gd³⁺) chelates with other inorganic nanoparticles with enhanced contrast, sensitivity, and MRI detection capability.²⁰³ One of those with remarkable performances is the magnetic iron oxide, also reported as T₁-weighted MRI contrast agents when they are small and isolated, depending on its magnetic effect on its composition, size, shape, and assembly.²⁰⁵ The preparation of MNP@MOF composites can be seen as an efficient method to ensure the dispersion of the MNPs, improve their contrast agent properties, and provide additional relevant properties associated with the MOF (e.g., porosity, versatile composition, drug loading, inherent therapeutic effect, targeting; see section 3. Theragnostic). Even more, although not reported so far, MOFs could provide intrinsic additional imaging properties by using MRI-active MOFs based on Gd(III), Mn(II), and Fe(III).^206,207 Thus, to date, only two reports have described the MRI-activity of MOF composites, by integrating MNPs (Table 2).

Table 2. Comparative Review of Magnetic Nanocomposites for MRI^a.

MNP@MOF	OL	MP for MOF	MNP@MOF size (nm)	MS (emu·g–1)	B (T)	r1 (mM–1·s–1)	r2 (mM–1·s–1)	r2/r1	TCL	ref
Fe3O4@ZIF-8	HmIM	Zn2+	120	18.9	0.5	15.1*b	372.0**c	24.6***d	HUVEC, 4T1	(116)
Fe3O4@MIL-100(Fe)@CDM	H3BTC	Fe3+	250		0.23	NR	157	NR		(129)

^aOL = organic linker, MP = metal precursor, B = magnetic field, TCL = targeted cell line.

^bFor pH 7.4, 28.4 (pH 6.2 and 4 mM of GSH), and 30 (pH 5 and 4 mM of GSH).

^cFor pH 7.4, 238.9 (pH 6.2 and 4 mM of GSH), and 176.3 (pH 5 and 4 mM of GSH).

^dFor pH 7.4, 8.4 (pH 6.2 and 4 mM of GSH), and 5.7 (pH 5 and 4 mM of GSH).

The first work was reported by Tregubov et al.,¹²⁹ preparing Fe₃O₄@MIL-100(Fe) with a coating of carboxymethyl dextran sodium salt (CMD), resulting in Fe₃O₄@MIL-100(Fe)@CMD with a hydrodynamic diameter in PBS of about 250 nm. The Fe₃O₄@MIL-100(Fe) and Fe₃O₄@MIL-100(Fe)@CMD nanocomposites possessed a transverse relaxivity of about 140 and 157 mM^–1·s^–1, respectively, considered interesting values close to the highest value reported between the commercialized contrast agent, in particular, the Feridex/Endorem, with a transverse relaxivity of about 152 mM^–1·s^–1.²⁰⁸ In further in vivo MRI investigations, a darkening effect was concentrated mainly in the spleen and liver, reticuloendothelial organs in charge of removing foreign compounds. This outcome serves however as a clear demonstration of the nanocomposite’s potential suitability for imaging applications. Further surface functionalization of the composites could facilitate evasion of the reticuloendothelial system or specific targeting.²⁰⁹

The second example, proposed by Lin and colleagues,¹¹⁶ took advantage of the magnetic properties of a Fe₃O₄@ZIF-8 nanocomposite (∼120 nm, M_S ∼ 18.9 vs 44.3 emu·g^–1 for Fe₃O₄) as a responsive T₂–T₁ switching contrast agent. The PBS-colloidally stable nanocomposite was however disassembled at acidic pH (6.2 and 5), and also in the presence of glutathione (GSH) at pH 7.4, 6.2, and 5, concluding that acidic conditions and GSH are degrading the system. However, further qualitative (i.e., PXRD) and quantitative characterization (i.e., inductively coupled plasma (ICP), high performance liquid chromatography (HPLC)) could better support the chemical and structural stability of this composite. Remarkably, the pH- and GHS-dependent degradation can be advantageously used in the acidic and GSH-overexpressed environment of cancer cells, exhibiting an inverse contrast behavior when the pH decreases and/or the GSH concentration increases. Indeed, under neutral conditions (pH = 7.4), Fe₃O₄@ZIF-8 was a T₂-contrast agent with a ratio of transverse and longitudinal relaxation rates (r₂/r₁) of 24.6, while by the composite disassembling at low pH and GSH, the system switched to a T₁-contrast agent with a r₂/r₁ ratio of 8.4 and 5.7 in case of 4 mM of GSH with pH 6.2 and pH 5, respectively. The r₂/r₁ ratios are higher than the range value (1.03–5.63) of the majority of the commercialized T₂ contrast agents such as Resovist (4.17) and Feridex/Endorem (5.63), at 0.47 T in water at 37 °C.²⁰⁸ After demonstrating good biocompatibility in human umbilical vein endothelial cells (HUVECs) and a mouse breast cancer cell line (4T1), the in vivo MRI detection in mice bearing a 4T1 breast tumor revealed a T₂-darkening effect in the liver and a trigger T₂–T₁ switching to brightening contrast in the tumor. This high proved the potential of these composites as MRI contrast agents.

3.3. Theragnostics

MOFs have garnered significant attention in the biomedical field also as theragnostic systems, with reports of several comprehensive reviews highlighting guest inclusions into MOFs to achieve combined therapeutic and diagnostic capabilities.^{20,48,210−216} Herein, we present a range of recent examples showcasing the use of MNP@MOFs for theragnostic purposes. These examples are organized according to our established criteria, progressing from the simplest to the most intricate nanocomposite configurations among each MOF’s family.

One of the first examples presented two composites based on the biocompatible MIL-100(Fe) MOF with maghemite (γ-Fe₃O₄) and citrate-functionalized maghemite (cit-γ-Fe₃O₄).⁷² These materials were prepared by a simple mixing method that easily provides composites although not very homogeneous. By varying the maghemite content (1 or 10 wt %), the transverse relaxivity for the cit-γ-Fe₃O₄@MIL-100(Fe)-1% and cit-γ-Fe₃O₄@MIL-100(Fe)-10% composites were about 93 and 21 mM^–1·s^–1, respectively, being lower than the pure cit-γ-Fe₃O₄ (∼180 mM^–1·s^–1). The best performing cit-γ-Fe₃O₄@MIL-100(Fe)-10% nanocomposite encapsulating DOX (14 wt %) was investigated both in vitro and in vivo, disclosing good biocompatibility and higher anticancer effect than the free DOX on prostatic cancer cells (PC3). Also, they demonstrated in vivo a good contrast for T₂-weighted images as well as T₁ contrast agent in a specific ultrashort echo time (UTE) sequence.

Subsequently, Wang et al.¹⁶⁰ also studied the MIL-100(Fe) by the in situ growth of the MOF over a carbon shell with carbon dots embedded and the iron oxide core (Fe₃O₄@C), proposing the resulting Fe₃O₄@C@MIL-100(Fe) composite for bioimaging and DDS of the anticancer dihydroartemisinin (DHA). The high DHA cargo (805 mg·g^–1; efficiency of 80.5%) was released in PBS (pH 6.2 and 5.0), being favored under an acidic environment. Concomitantly to the drug release, the MIL-100(Fe) structure collapsed and released Fe³⁺ to the medium, which could be reduced to Fe²⁺ by reductive molecules of the cells (e.g., ferric reductase) and the acidity of tumor cells. The Fe₃O₄@C@MIL-100(Fe) composite exhibited good in vitro biocompatibility; however, with an increased cytotoxicity of the DHA-loaded Fe₃O₄@C@MIL-100(Fe) by the generated reactive oxygen species (ROS) as a consequence of both the produced Fe²⁺ and the released DHA. Remarkably, promising in vivo anticancer therapy was evidenced, with low side effects and a higher tumor inhibition for DHA-loaded composites (M_S ∼ 21.2 vs 44.4 emu·g^–1 for Fe₃O₄@C), in particular, when an external magnetic field was applied. Additionally, the carbon-shell MNPs contributed with a double bioimaging effect, combining fluorescence optical imaging and MRI (r₂ ∼ 352.45 mM^–1·s^–1 at 3 T and 25 °C).

Both of the presented composites are based on the biocompatible MIL-100(Fe), although they differ primarily in the phase of the MNPs employed, specifically maghemite and magnetite in the former and latter cases, respectively. Unlike the first example, the second report extends beyond in vivo MRI studies, showcasing a significant antitumor effect in an in vivo animal model and attesting to the remarkable potential for engaging in future clinical experiments of great interest.

Additionally, another complex system based on MNP@MIL-100(Fe) was prepared based on upconversion nanoparticles (UPNPs) for PCT and PDT assisted by computed tomography (CT) and upconversion luminescence (UCL) imaging.¹⁶¹ Through the LbL method, the MOF was synthesized as a shell over the MNPs core and later UPNPs were added. The surface of the resulting composite (∼300 nm, Figure 8) was coated with carboxylate-terminal polyethylene glycol (PEG-COOH; Fe₃O₄@MIL-100(Fe)-UCNPs-PEG; FMUP), keeping a BET surface area of about 106 m²·g^–1, with this low surface being associated with the low MOF content (approximately 6%). Indeed, the much higher mass concentrations of Fe₃O₄ and UCNPs were estimated to be approximately 30 and 53%, respectively. Interestingly, this complex system was formed by a N–P heterojunction, where the Fe₃O₄ nanoparticle was an n-type semiconductor and MIL-100(Fe) was a p-type semiconductor. For the first time, this nanocomposite was observed from this perspective, noticing that Fe₃O₄@MIL-100(Fe) (FU) absorbs below 500 nm, with a band energy of about 2.16 eV in the UV–vis diffuse reflectance spectroscopy (DRS). Therefore, the UCNPs acted as a photosensitizer that absorb the near-infrared (NIR) excitation light and convert it to UV–vis light. The combination of them in this heterojunction system was able to produce hydroxyl radicals (^•OH), one of the toxic ROS, exhibiting a PDT activity. A higher ROS generation capability and, therefore, the highest cell death in HeLa cells, was observed for the Fe₃O₄@MIL-100(Fe)-UCNPs-PEG upon NIR excitation (980 nm laser) with clear detection of the cell death through the confocal laser scanning microscopy (CLSM). Moreover, the Fe₃O₄@MIL-100(Fe)-UCNPs with tested hemocompatibility and low toxicity were further studied for tumor inhibition in mice upon a subcutaneous injection of the carcinoma cells of the uterine cervix of U14, demonstrating a high antitumoral effect. Indeed, for the mice treated with Fe₃O₄@MIL-100(Fe)-UCNPs under the 980 nm NIR laser for 15 min (0.9 W/cm², 5 min break after 8 min excitation) at the tumor site, it was observed a slight decrease of the relative tumor volume with respect to the enormous relative volume increase of the control systems. In comparison to its predecessors, this nanocomposite undoubtedly embodies a more intricate system, encompassing the incorporation of UCNPs to enhance therapy efficiency and PEG functionalization to improve MOF stability and biocompatibility. Furthermore, the presence of UCNPs was exploited for in vitro and in vivo CT and UCL imaging, demonstrating good signals and, therefore, good results for a promising theragnostic agent. Nevertheless, further investigation can be conducted to explore the potential contributions of MNPs, such as MRI imaging or MHT. Undoubtedly, the nanocomposite possesses an untapped potential that warrants further exploration.

Figure 8.

Schematic illustration of the synthesis process of Fe₃O₄@MIL-100(Fe)-UCNPs-PEG (FMUP), and the intracellular photon-Fenton reaction of FMUP with intracellular H₂O₂ under the irradiation of 980 nm. Reproduced from ref (161) with permission from Copyright 2018 Elsevier B.V.

Apart from the MIL-100(Fe), there are also several reports based on the extensively studied ZIF-8. For instance, an amino-terminal polyethylene glycol (PEG-NH₂) coated Fe₃O₄@ZIF-8 nanocomposite (∼97 nm) with magnetic properties (M_S ∼ 6.6 vs 42.6 emu·g^–1 for Fe₃O₄) was investigated.¹²² The drug loading of the antitumoral arsenic trioxide (ATO, As(OH)₃) was performed before the polymer coating in an aqueous solution, encapsulating a relatively low amount (53 mg·g^–1). Note here that the PEG-NH₂ coating could promote a partial release of the drug cargo, with a decrease, indeed, in the determined As content by ICP from 14.35 ± 0.02 to 13.95 ± 0.03 wt %. Subsequently, the drug release in PBS at pH 6.0 or 7.4 revealed 27 or 17% of ATO released after 1 h, 80 or 44% after 1 day, and a maximum of 80% or 53% after 7 days, respectively. The difference in both pH values was related to the partial (pH 7.4) and total (pH 6.0) degradation of the ZIF-8 shell of the composite due to the low hydrolytic stability of this MOF. Furthermore, negligible cytotoxicity was detected in fibroblasts for both drug-loaded and drug-free composites, whereas in malignant atypical teratoid rhabdoid tumor cell lines (BT12 and BT16) the drug-loaded composite presented cytotoxicity. Besides, all the systems provided a good T₂-weighted contrast agent in MRI at 1.5 T in 0.1% agar solution at the temperature of 37 °C, resulting in r₂/r₁ ratios of 48.51, 10.67, and 12.39 for Fe₃O₄ nanoclusters, the drug-free, and the drug-loaded Fe₃O₄@ZIF-8@PEG-NH₂ nanocomposites, respectively. These results highlight the potential of this system because the r₂/r₁ ratio is higher than one of the highest reported between the commercialized T₂ contrast agent, such as Resovist (7.0) and Feridex/Endorem (8.7), at 1.5 T in water at 37 °C.²⁰⁸ We should however consider that the conditions (agar solution) are far from biological environments, in which the composite could exhibit a completely different chemical and colloidal stability.

Another example of the ZIF-8 nanocomposite was synthesized by He et al.¹⁴³ by growing the MOF on carbon shell magnetic cores (Fe₃O₄@C@ZIF-8; ∼ 190 nm) and encapsulating a very high content of DOX (730 mg·g^–1). As shown in previous cases, the drug release was favored under acidic conditions (after 200 h, 95 vs 38% under pH 5.5 and 7.4, respectively). The in vitro biocompatibility of Fe₃O₄@C@ZIF-8 and the higher toxicity of DOX-loaded Fe₃O₄@C@ZIF-8 rather than the free-DOX were evidenced on A549 cells. The magnetic nanocomposite (M_S ∼ 12 vs 28 emu·g^–1 for Fe₃O₄@C) reported a promising value of r₂ of about 331.8 mM^–1·s^–1 at 3.0 T in water solution at 25 °C. This specific relaxivity is more than double the value reported for the commercialized T₂ contrast agent Resovist, which is 160 mM^–1·s^–1 at 3.0 T in water at 37 °C.²⁰⁸ In the previous system, Fe₃O₄@ZIF-8@PEG-NH₂,¹²² the cytotoxicity tests were conducted only in vitro (fibroblasts and cancer cell lines), suggesting a target cell death for the cancer cells through the cargo release of the DOX-loaded nanocomposites. Instead, in the case of Fe₃O₄@C@ZIF-8, the study was extended to in vivo therapeutic efficacy studies upon intravenous administration in mice bearing established A549 cells (lung cancer model), exhibiting an accumulation in the tumor site due to the EPR effect and a clearance mechanism mainly directed to the liver through the T₂*-weighted MR images. Subsequent evaluation of the therapeutic effect demonstrated an average tumor suppression efficacy of the DOX-encapsulated nanocomposite of about 64.5% (vs 16.1% and 12.8% of the free DOX and the not-encapsulated Fe₃O₄@C@ZIF-8, respectively). However, in the first approach, the selection of the PEG-coating suggests that there are some concerns about stability and shelf life of a product based on ZIF-8, which is not considered in the second example. The second example impressively presents in vivo efficacy for the therapeutic aspect, coupled with promising MRI detection capabilities.

In an upgrading attempt, Bian et al.¹⁷² fabricated a multifunctional nanocomposite based on Fe₃O₄@ZIF-8 implemented by gold nanoclusters (AuNCs). The MNPs were initially PAA-coated (∼90 nm), adding then GSH capped AuNCs and finally, growing the ZIF-8 over the Fe₃O₄@PAA/AuNCs to obtain the Fe₃O₄@PAA/AuNCs/ZIF-8 nanocomposite (∼130 nm) with a moderate magnetic effect (M_S ∼ 8.2 vs 44.4 emu·g^–1 for Fe₃O₄). By associating the DOX with exceptional loadings (1540 mg·g^–1; therapy), this theragnostic agent combines a trimodal imaging by (i) MRI (Fe₃O₄ nanoparticles), with a r₂ value of about 53.8 mg^–1·mL·s^–1 at 1.2 T, revealing a r₂ value between the commercialized T₂ contrast agents Resovist (61 mM^–1·s^–1) and Feridex/Endorem (41 mM^–1·s^–1) at 1.5 T in water at 37 °C;²⁰⁸ (ii) computed X-ray tomography (Au), exhibiting the nanocomposite a CT imaging contrast behavior, with increasing intensity of CT signals with the nanocomposite concentration, and (iii) fluorescence optical imaging (FOI), evidencing a more emitting intensity at 609.6 nm of Fe₃O₄@PAA/AuNCs/ZIF-8 than discrete AuNCs at same concentration due to aggregation-enhanced fluorescence (AEF) effect. In comparison to the previous example, the drug loading value in this case demonstrates a remarkable increase (730 vs 1540 mg·g^–1) possibly attributed to adsorption on the outer surface of MOF. However, drug release studies in PBS pH 5.3 at 37 °C only showed an initial DOX release of 12.2% within the first 15 min, ruling out the weak association of a large drug amount to the composite. Subsequently, a sustained release was reached, a priori associated with a prolonged therapeutic effect. Additionally, in vitro studies in human liver cancer cells HepG-2 of this trimodal cancer imaging composite revealed the desired intensity signals of CT and MR (T₂) dependence on the composite concentration and the cell uptake (λ_em ∼ 609.6 nm) via endocytosis. For the therapeutic effect, the magnetic DOX-loaded composite was suitable for promising magnetically targeted drug delivery, displaying a recovery with a magnet in solution, and pH-sensitive drug release (PBS, pH 7.4 vs 5.3), consistent with the desired preferential release in the acidic cancer cells. The biocompatibility of the Fe₃O₄@PAA/AuNCs/ZIF-8 composite was supported in vitro and in vivo. Additionally, its intravenous administration for the in vivo tumor inhibition in hepatocarcinoma of a H-22 xenograft demonstrated a higher average inhibition rate (70%) in the DOX-loaded Fe₃O₄@PAA/AuNCs/ZIF-8, rather than DOX-free (39%). Respect to the Fe₃O₄@C@ZIF-8, the Fe₃O₄@PAA/AuNCs/ZIF-8 system exhibited a r₂ of about 53.8 mM^–1·s^–1 at 1.2 T, which is compatible with the commercialized T₂ systems at 1.5 T in water at 37 °C (i.e., Resovist and Feridex/Endorem with 41 and 61 mM^–1·s^–1, respectively).²⁰⁸ This system has demonstrated trimodal imaging capabilities in vitro, which, combined with the in vivo tumor suppression efficacy, holds promising potential for synergistic diagnosis and therapy.

Instead, the complex system Fe₃O₄@ZIF-8@ZIF-67/FA/Q⁹⁴ (quercetin, Q) exhibits a secondary shell composed of a distinct MOF (ZIF-67), in conjunction with a FA modification that distinguishes it from preceding examples. In the present intricate system, the antitumoral drug (Q) encapsulation and release reported the highest value of efficiency (∼93%) at pH 5 with a release of ∼79 and 90% of Q after 1 and 9 h in PBS solution at pH 5, respectively. Certainly, the distinct characteristics and attributes of the magnetic core are subject to modification due to the presence and interactions of these dissimilar and overlapping shells. Indeed, the phantom images confirmed the contrast enhancement, therefore, the increase in relaxivity with increasing Fe₃O₄ concentration and the value of r₂ relaxivity were 429.6, 112.0, 85.9 mM^–1 s^–1 for Fe₃O₄, Fe₃O₄@ZIF-8, and Fe₃O₄@ZIF-8@ZIF-67/FA nanoparticles at 3 T and 22.5 °C, respectively. The values suggested the possibility to use the nanocomposite as an efficient (T₂) MRI contrast agent. Indeed, the values of r₂ for Fe₃O₄@ZIF-8@ZIF-67/FA are within the range of the commercialized contrast agents measured in water at 37 °C at the same magnetic field.²⁰⁸ Therefore, the strong MRI signal and drug loading make promising this nanocomposite for theragnostic purposes. In fact, the viability of MCF-7 (FA receptor negative cell line) and MDA-MB-231 (FA receptor positive cell line) cells and the flow cytometric analysis showed a specific uptake by MDA-MB-231 due to the FA targeting, with a toxicity of ∼40, 60, and 80% in MDA-MB-231 cells for higher Fe₃O₄@ZIF-8@ZIF-67/FA/Q concentrations (27, 40.25, and 54 μg·mL^–1, respectively). The noteworthy aspect pertains to the safety associated with the chosen MOFs. The resolution of this inquiry can only be attained through additional in vivo testing.

Up to this point, we have seen a magnetic core based on Fe₃O₄. As an alternative, Yang et al.¹³⁷ developed a smart pH/NIR dual-stimulus-responsive CoFe₂O₄@PDA@ZIF-8 nanocomposite (∼150 nm, Figure 9) based on a mesoporous magnetic CoFe₂O₄ containing DOX. Briefly, a PDA layer was first incorporated into the DOX-loaded CoFe₂O₄ and, subsequently, a ZIF-8 shell. The latter favored a second encapsulation of the anticancer camptothecin (CPT), which was introduced during the MOF synthesis. The novelty of this DDS lies in both its unconventional magnetic core and its porosity, which translates into a dual loading capacity. Indeed, this system had a loading efficiency of about 98% for DOX and 46% for CPT (the loading value for DOX could be estimated around 98% with respect to the magnetic core, instead, it was not possible to extract in case of CPT due to lack in experimental details related to nanocomposites amount). The cargo release was consistent with the desired pH-response, with the release content of 61% for CPT and 37% for DOX at pH 5.0. Moreover, it had two-stage acidic-mediated processes associated with the degradation of ZIF-8 and PDA, favoring first CPT release in 12 h and, then, DOX in 40 h. The nanocomposite exhibited also a NIR-stimulation release profile, which evidenced a burst release of the two drugs under an 808 nm laser because the magnetic core had a thermal expansion, disintegrating both the PDA and ZIF-8. Furthermore, CoFe₂O₄@PDA@ZIF-8 exhibited good viability in HepG2 cells. In contrast, the nanocomposite showed significant cytotoxicity in the case of all the drug combinations (DOX, CPT, DOX+CPT) with/without NIR-stimulation, with the highest cell death for the CPT-loaded CoFe₂O₄@PDA@ZIF-8 with NIR, proving the in vitro efficacy of the system. In addition, the magnetic CoFe₂O₄@PDA@ZIF-8 (M_S ∼ 36.4 vs 68.5 emu·g^–1 for CoFe₂O₄) demonstrated a T₂-weighted imaging capacity at 1.2 T (r₂ = 38.3 vs 53.3 mM^–1·s^–1 CoFe₂O₄@PDA), in the range of the highest value for the commercialized T₂ contrast agent (at 1.5 T in water at 37 °C²⁰⁸). Interestingly, a significant darkening in vivo effect (HepG2 tumor-bearing mice) was observed in the liver upon 1 h postinjection of the drugs-loaded composite and, after 9 h, in the tumor region. Subsequently, considering the in vitro photothermal effect of CPT- and DOX-loaded CoFe₂O₄@PDA@ZIF-8 (T > 65 °C), the in vivo performance was studied in a HepG2 xenograft tumor-bearing mice model. After 9 h-injection, according to the highest MRI darkening signal at the tumor site, the light–heat conversion under NIR laser was tested, resulting in an increase from 32 to 50 °C in the cancer area after only 10 min exposure. In conclusion, the in vivo PTT and simultaneous CPT and DOX therapy showed a promising synergistic effect on tumor size inhibition (from relative tumor volume 5.5 ± 1.2 (PBS) to 0.6 ± 0.2 for nanosystems under NIR), with histologic coagulative tumor necrosis of 90%. In addition, upon the location of a magnet located in the proximity of the tumor site (magnetic guided therapy) without NIR, the relative tumor volume was about 0.9 ± 0.4 with a tumor necrosis of 80% (vs 1.2 ± 0.6 and 75% in the case of nonmagnetically guided CPT and DOX therapy). The dual drug therapy combined with the implementation of magnetically guided administration and NIR laser reached a relative tumor volume of 0.3 ± 0.1 (magnet + NIR), with histologic coagulative tumor necrosis of 97%. These results, associated with the absence of appreciable damages or inflammatory lesions on normal tissues after treatment, demonstrated safe and potential applications. This example serves as a notable demonstration of the potential to leverage the “tissue transparent window”, emphasizing the already-established significance of MOFs in PTT.²¹⁷ Further, this advanced research highlights an unconventional approach wherein the noninvasive placement of a magnet in proximity to the tumor site in mice contributes to enhance the formulation efficacy. This groundbreaking approach, first pioneered by Falcaro and collaborators, involved spatial control pertaining to MNPs@MOF within microfluidic circuits.⁸³ The potential implications of this concept are far-reaching, as it paved the way to extend and position control techniques in therapeutics within organisms, thereby enabling sophisticated control over drug delivery systems. Nonetheless, further studies dealing with the biodistribution and safety of both ZIF-8 and CoFe₂O₄ would complete this really nice piece of work.

Figure 9.

(A) Synthesis of CPT- and DOX-loaded CoFe₂O₄@PDA@ZIF-8 nanocarrier and (B) theragnostic strategy of CPT- and DOX-loaded CoFe₂O₄@PDA@ZIF-8 nanocarrier for magnetically guided multidrug chemotherapy and photothermal synergistic therapy with pH and NIR-stimulation release. Reproduced from ref (137). Copyright 2017 American Chemical Society.

Apart from the guided-therapy using the magnetic properties of MNPs, its therapeutic effect can be also explored under an external AMF. In discussing therapy involving MNPs, the primary focus is often on utilizing a high-frequency alternating magnetic field (HF-AMF, 50–400 kHz) to induce localized temperature elevation (or magnetic hyperthermia; MHT).²¹⁸ However, in the forthcoming example, the recent advancements in various magnetic nanocomposites have led to the exploration of an alternative approach utilizing extremely low-frequency alternating magnetic fields (ELF-AMF, 0.01–10 kHz).²¹⁸ Indeed, in several magnetic nanocomposites (e.g., magnetoliposomes),^218−223 extensive investigations have been conducted to examine drug release promotion under lower magnetic field strength. In the case of MNP@MOF, Fang et al.⁹⁵ proposed a magnetic Fe₃O₄@ZIF-90 nanocomposite (∼65 nm) conjugated with rat serum albumin (RSA; M_S ∼ 7 vs 49 emu·g^–1 for Fe₃O₄) as an AMF-triggered DDS, under potential MRI observation. Indeed, MRI studies reported a good transverse relaxivity (r₂) of about ∼133.7 mM^–1·s^–1 under 7 T and a T₂-weighted darkening effect in images. Further studies with different magnetic fields will permit to better compare it with commercial products, even if some recent studies are investigating also this magnetic field for human brain disorders.²²⁴ Additionally, Fe₃O₄@ZIF-90 was able to encapsulate the antitumoral drug 5-FU with a progressive release in PBS (50% released in 7 h), which could be accelerated by applying an ELF-AMF (50% released upon 1.5 h under ELF-AMF with a frequency of 20 Hz applied for 20 min every 1 h). An ELF-AMF applied for even 3 h did not provoke any MHT effect, which was detected only under an HF-AMF using 488 kHz for 10 min, with an increase of temperature of about 15 °C. Therefore, this study presents the initial instance of employing ELF-AMF in a magnetic nanocomposite based on MOFs. Undoubtedly, this system represents a potential alternative pathway for exploration in the coming years, extending this approach to in vitro and subsequently in vivo investigations.

In the family of UiO, three interesting works are reported. First, Zhao et al.¹⁴² proposed a biocompatible Fe₃O₄@UiO-66 nanocomposite (∼240 nm; M_S ∼ 51.6 vs 69.7 emu·g^–1 for Fe₃O₄) with an extremely high DOX loading capacity (2000 mg·g^–1 and efficiency of 66.3% wt). The authors attributed the substantial drug capacity to the extensive surface area of the shell and the UiO-66/DOX interactions; notably, the stable coordination bonding between the deprotonated hydroxyls in DOX and the numerous Zr⁴⁺, as confirmed by UV–vis spectroscopy and X-ray photoelectron spectroscopy (XPS)¹⁴² as well as other interactions (π–π stacking between the aromatic anthracycline and the aromatic ligand, hydrogen bonding, etc.). However, this extreme loading could be also related to surface adsorption or to the presence of a large number of defects in the UiO-66 structure, promoting the interactions and a larger porosity. Afterward, the drug was gradually delivered as a function of the pH in PBS, with a sustained release of DOX in 41 days of 36, 22, 17, and 14% at pH 4.0, 5.0, 6.0, and 7.4, respectively. As already stated, the release of drugs in buffer solutions under acidic pH is influenced by various factors. The stability of the MOF can be impacted by the phosphate groups, able to coordinate Zr replacing the interactions with both DOX and the organic linker. Additionally, the protonation of the amino group in DOX leads to a positively charged DOX molecule, thus weakening its interactions with the MOF, which presented a less negative surface zeta potential under acidic conditions. Even at a low dose (20 mg·L^–1), DOX-loaded Fe₃O₄@UiO-66 composites manifest 60% of cell death in HeLa cells, comparable with free DOX. The antitumor activity of DOX-loaded Fe₃O₄@UiO-66 increased either with the DOX loading or with longer incubation times, supporting a progressive release under these conditions. Furthermore, in vivo safety and biodistribution studies demonstrated the low toxicity of the magnetic composite with accumulations mainly in the spleen and liver, as expected for nonfunctionalized nanoparticles.²⁰⁹ In addition, Fe₃O₄@UiO-66 had a T₂ contrast agent behavior with an important r₂ ∼ 255.9 mM^–1·s^–1, revealing the in vitro MR images with a concentration-dependent darkening effect. Experimental conditions used a slightly higher magnetic field, closer to one of the generally applied ones (0.55 vs 0.47 T), and a low temperature (32 vs 37 °C). The determined r₂ is higher than the majority of the commercialized T₂ contrast agents, with maximum value for Resovist at 0.47 T in plasma at 37 °C.²⁰⁸In vivo studies in HeLa tumor-bearing mice intravenously injected with the nanocomposites showed good T₂-weighted MR signals in the tumor after 1 h postinjection with a maximum of darkness after 9 h, verifying a tumor growth inhibition after 21 days postinjection.

In another work dealing with a DOX-containing Fe₃O₄@UiO-66-NH₂ nanocomposite (drug loading content 62 wt %), Alijani and co-workers¹²⁶ proposed a tiny DOX-loaded core–shell nanostructure (∼16 nm) with the conjugation of highly fluorescent carbon dots (CDs) and a nucleolin-binding aptamer (Apt), AS1411. This system, denoted as Fe₃O₄@UiO-66-NH₂-DOX-CDs-Apt (Figure 10), was investigated as a stimuli-responsive antitumoral drug carrier and cellular bioimaging. Additionally, the cytotoxicity on normal HUVEC and human epithelial breast cancer cells (MDA-MB-231), characterized by nucleolin overexpression, was evaluated for Fe₃O₄@MOF, Fe₃O₄@MOF-DOX, Fe₃O₄@MOF-DOX-CDs, and Fe₃O₄@MOF-DOX-CDs-Apt. In the HUVEC cells, it was observed good cell viability (>90%) for all the composites. Meanwhile, in the MDA-MB-231, it was reported a concentration-dependent toxicity in the two DOX-loaded systems. At the same concentration, the highest toxicity was obtained with Fe₃O₄@MOF-DOX-CDs-Apt, with a 77% of cancer cells death by mainly apoptosis induction. Henceforth, a selective therapeutic effect was proven. Subsequently, cell internalization mediated by aptamer-nucleolin recognition was confirmed in the MDA-MB-231 cancer cell line. Regarding the in vitro drug release, Fe₃O₄@MOF-DOX-CDs-Apt at pH 5.5 and 7.4 exhibited a progressive release over time without burst effect and, after 4 days, of about 29.1% and 47.3% at pH 7.4 and pH 5.5, respectively.

Figure 10.

Schematic illustration of the synthesis of UiO-66-NH₂ on the as-synthesized magnetic Fe₃O₄ nanoparticles, and postmodification of the nanostructure with DOX, carbon dots, and nucleolin-binding aptamer conjugation. Subsequently, a schematic illustration of the cell internalization of the Fe₃O₄@MOF-DOX-CDs-Apt via a nucleolin-mediated interaction and pH-triggered DOX release in lysosome or endosome of the cancer cells. Reproduced from ref (126) with permission from Copyright 2020 Elsevier Inc.

Finally, based also on Fe₃O₄@UiO-66-NH₂, Wu et al.²²⁵ developed a theragnostic nanocomposite (∼40–60 nm) loaded with the 5-FU (efficiency 11%, the loading value was not possible to extract due to lack in experimental details related to 5-FU content) and a water-soluble carboxylatopillar[6]arene (WP6) coating, known as Fe₃O₄@UiO-66-NH₂@WP6. The drug loading proceeded before the WP6 functionalization, to ensure high drug loading and furthermore promote the release in cancer cells, although postfunctionalization could promote delivery of the active cargo. Indeed, the drug release was facilitated by acidic pH (7 vs 5), temperature, and divalent metal ions in the body (e.g. Ca²⁺ and Zn²⁺), weakening the interaction of the WP6 due to its carboxylate groups chelated by them. The 5-FU-loaded Fe₃O₄@UiO-66-NH₂@WP6 (M_S ∼ 46.8 vs 89.3 emu·g^–1 for Fe₃O₄) could be separated by an external magnetic field and, furthermore, the Fe₃O₄ in the nanoplatform imparted superparamagnetism. Also, 5-FU-loaded Fe₃O₄@UiO-66-NH₂@WP6 demonstrated good in vitro biocompatibility on normal HUVEC and an antitumoral activity in HeLa cells. Finally, the nanocomposite behaved as a T₂-weighted MRI contrast agent with an evident darkening effect of MRI signals in HeLa cells and a r₂ of about 72.2 mM^–1·s^–1 under a 7 T magnetic field. In contrast with commercial contrasts, generally employed at 1.5 T for clinical instruments and at 0.47 T for relaxation measurements,²⁰⁸ here the darkening effect is investigated at a high magnetic field strength of 7 T. Emerging research trends currently to explore higher magnetic fields, aiming to provide a rationale for its implementation since, as opposed to lower magnetic fields, their superior signal-to-noise and contrast-to-noise ratios. These enhanced ratios facilitate high-resolution imaging and improved contrast, thereby enabling easier identification of lesions and structural alterations associated with brain disorders.²²⁴ In the case of Fe₃O₄@UiO-66 and Fe₃O₄@UiO-66-NH₂@WP6, the obtained r₂ values are particularly compelling for a simpler nanocomposite. Conversely, despite the proven targeting in Fe₃O₄@UiO-66-NH₂-DOX-CDs-Apt and the bright green emission of CDs under UV light, further investigations are required to explore the potential of this system as fluorescence imaging or MRI agent.

Nanocomposites based on other MOF structures were also prepared. For instance, a magnetic core–shell Fe₃O₄@IRMOF-3 nanocomposite conjugated with FA (200 nm; M_S ∼ 48.9 vs 80.46 emu·g^–1 for Fe₃O₄) was proposed for theragnostic,¹⁰¹ associating the antitumoral paclitaxel (PTX; 123.2 mg·g^–1 and efficiency of 82%). The cytotoxicity assays in HeLa and murine fibroblast (NIH3T3) revealed a selective and concentration-dependent toxic effect only for PTX-loaded Fe₃O₄@IRMOF-3/FA, with a cell death higher for cancerous cells. Additionally, the fluorescence microscopy confirmed an endocytosis cell uptake of Fe₃O₄@IRMOF-3/FA and the MRI study in a clinical scanner (1.5 T) reported a darkening signal in HeLa cells. Even if the M_S presented a promising value, the system was not further investigated for MHT. Lastly, the PTX-loaded Fe₃O₄@IRMOF-3/FA system exhibited a controlled drug release over time in a simulated physiological media at pH 7.4, underlining its potentiality as a DDS and T₂-weighted MRI contrast agent. As an alternative, Chowdhuri and co-workers¹⁰² reported an interesting more complex IRMOF-3 composite (Figure 11) based on O-carboxymethyl chitosan (OCMC) functionalized MNPs (Fe₃O₄@OCMC) and fluorescence CDs, evaluating its optical imaging. The MOF growth occurred on the Fe₃O₄@OCMC nanoparticles in the presence of FA, showing a flower-like morphology (∼200 nm), and promoting targeted drug delivery toward the overexpression of folate receptor in various tumors. Finally, highly fluorescent CDs (∼10 nm) were conjugated to the FA-nanocomposite for optical imaging. Actually, the UV–vis absorbance spectrum of Fe₃O₄@OCMC@IRMOF-3/FA showed two bands, at 280 nm characteristic of n−π* transitions of FA and ∼345 nm due to IRMOF-3 (330 nm) and the π–π* transitions of FA (360 nm). In point of fact, the magnetic nanocomposite (M_S ∼ 51 vs 78 and 66.5 emu·g^–1 for Fe₃O₄ and Fe₃O₄@OCMC, respectively) exhibited also a fluorescent band, with intensities tuned by the excitation wavelength. Therefore, the CDs presence permitted the monitoring of the intracellular uptake into the folate-overexpressed HeLa cells, demonstrating an endocytosis process favored by FA recognition. Besides, the Fe₃O₄@OCMC@IRMOF-3/FA nanocomposite underwent a physical DOX encapsulation (163 mg·g^–1; efficiency 96%), where the loading efficiency may be correlated with the surface adsorption of the drug onto the complex nanocomposite.¹⁸⁶ In addition, the in vitro toxicity of the unloaded nanocomposite showed biocompatibility while the DOX-loaded nanocomposite reported a selective toxic effect toward cancer cells via apoptosis due to the FA-conjugation. Lastly, Fe₃O₄@OCMC@IRMOF-3/FA exhibited good stability under physiological media at pH 7.4, with a higher drug release at lower pH (48 and 55% of DOX released with respect to 22 and 27% at pH 5.5 and 7.4 after 12 and 24 h, respectively). In both IRMOF-3-based examinations, FA-conjugation was exploited to achieve selective endocytosis. Despite the promising MRI capabilities of these systems, even demonstrating fluorescence imaging when associated with CDs, it would be worthwhile to pursue in vivo studies investigating their biodistribution, elimination, safety, and efficacy.

Figure 11.

Schematic representation of the synthetic procedure for the FA encapsulated magnetic nanoscaled MOFs as targeted DOX carriers. Reproduced from ref (102). Copyright 2016 American Chemical Society.

Another nanocomposite, Fe₃O₄@C@P-MOF (∼95 nm, M_S ∼ 24.5 emu·g^–1) based on a Fe₃O₄@C core (M_S ∼ 39.8 emu·g^–1) and a porphyrin-MOF (PMOF) was prepared,²²⁶ proposing a dual-imaging and dual-therapy. An imaging-guided therapy was possible for this system, which presented a MRI darkening effect at 1.2 T with a r₂ ∼ 72.6 mM^–1·s^–1 and r₂/r₁ of about 59.0 (note here that the experimental temperature was set at 30 °C). Indeed, the r₂/r₁ ratio is higher than the best reported for commercialized T₂ contrast agents (Resovist 7.01 and Feridex/Endorem 8.72), with also a higher r₂ value (Resovist 61 and Feridex/Endorem 41) at 1.5 T in water at 37 °C.²⁰⁸ Moreover, the optical properties highlighted potential not only for FOI but also for PTT and PDT, with a strong absorption band at 416 nm and four peaks over an extended broad absorption band that covered all the spectrum until the NIR region. The emission capacity (λ_em ∼ 688 nm with λ_ex ∼ 553 nm) and the significant Stocks shifts extended the nanocomposite also as a fluorescent imaging system. Related to the therapy, the photothermal tests revealed a significant rise in temperature (>50 °C) under 808 nm NIR irradiation. This is a value maybe too high for safety, although more real tests will be required. Furthermore, the PTT efficiency was independent of the PDT one, which reported a ¹O₂ generation under 655 nm irradiation with a ¹O₂ quantum yield of ∼44%. From a solution to in vitro test, further interesting results showed, first, the nanocomposite biocompatibility in breast cancer MCF-7 cells. Then, upon irradiation for 10 min under 665 nm (PDT) or 808 nm (PTT), an increase in toxicity was reported by increasing the nanocomposite concentration and by cotreating with both PDT and PTT. Fixing a dose of 20 mg·kg^–1, then the biocompatibility was further confirmed in female BALB/c-nu mice, without evident side effects. The in vivo dual-imaging efficiency for intravenously injected Fe₃O₄@C@P-MOF was verified in healthy nude mice and MCF-7 tumor-bearing nude mice, qualitatively reporting after 22 h in both MR and fluorescence imaging an accumulation in the liver, with a clearance within 8 days through feces and urine. In the presence of a tumor, it was observed intense fluorescent and MRI signals in the cancer region after 26 h, assigned to an EPR effect. The high tumor uptake was in vivo studied for PTT, PDT, and PTT-PDT co-therapy, reporting the highest tumor growth inhibition in the synergistic dual-therapy compared to single PTT, single PDT, just composites and control. Additionally, ex vivo tests reported an efficient uptake of the nanocomposites in the tumor tissue, analyzing all the major organs dissected and imaged with both MRI and FOI. From our perspective, this particular example is undeniably highly biologically relevant, although may lack some MOF chemical information (structural, textural, etc.). The amalgamation of MRI properties offered by the MNPs, coupled with the diagnostic and therapeutic attributes contributed by the MOF, holds great promise within the nanocomposite.

Overall, Table 3 comprehensively presents the data pertaining to the selected theragnostic examples, evidencing once again that benchmarked MOF families are prevalent (MIL-100(Fe), UiO-66, ZIF-8), although here two new MOF structures are also introduced with interesting performances (e.g. IRMOF-3 and PMOF). Most of them with a particle size from 50 to 250 nm, compatible with an intravenous administration of course, dependent also on their biological chemical and colloidal stability, which will significantly influence their biodistribution, elimination, safety and efficacy. In terms of magnetic properties, M_S values range from 6.6 to 62 emu·g^–1. Although mostly with relatively low M_S values, few composites demonstrate higher values such as 51,¹⁰² 52,¹⁴² and 62,⁷² many of these systems combine magnetic properties with the incorporation of other types of nanoparticles to enhance their imaging and therapy capabilities. Some of these complex nanocomposites present noteworthy competitive MRI results, comparable with commercially available contrast agents. Nonetheless, it is evident that there remains limited exploration of the magnetic properties specifically for therapeutic aspects, especially within the realm of MHT.

Table 3. Comparative Review of Magnetic Nanocomposites for Theragnostics^a.

MNP@MOF	MNP	OL	MP for MOFs	MNP@MOF size (nm)	MS (emu·g–1)	D	B (T)	r1 (mM–1·s–1)	r2 (mM–1·s–1)	r2/r1	T	drug	DL(mg·g–1)	LE (%)	TCL	ref
γ-Fe3O4@MIL-100(Fe) and cit-γ-Fe3O4@MIL-100(Fe)	Fe3O4	NH2–H2BDC	Fe3+	150–170	35–62	MRI	7	NR	93*b	NR	DDS	DOX	140	NR	PC3	(72)
Fe3O4@C@MIL-100(Fe)	Fe3O4	NH2–H2BDC	Fe3+	190	21.2	FOI and MRI	3	NR	352.45	NR	DDS	DHA	804.9	80.5	HeLa and A549	(160)
Fe3O4@MIL-100(Fe)-UCNPs-PEG	Fe3O4	NH2–H2BDC	Fe3+	300	NR	CT and UCL	NR	NR	NR	NR	PCT and PDT	NR	NR	NR	L929 and HeLa	(161)
Fe3O4@ZIF-8@PEG-NH2	Fe3O4	HmIM	Zn2+	97	6.6	MRI	3	2.37	25.25	10.67	DDS	ATO	70	NR	fibroblast, BT12, and BT16	(122)
Fe3O4@C@ZIF-8	Fe3O4	HmIM	Zn2+	190	12	MRI	3	NR	331.79	NR	DDS	DOX	730	NR	A549	(143)
Fe3O4@PAA/AuNCs/ZIF-8	Fe3O4	HmIM	Zn2+	130	8.2	MRI, CT, and FOI	1.2	NR	53.79	NR	DDS	DOX	1540	81.1	HepG-2	(172)
Fe3O4@ZIF-8@ZIF-67/FA	Fe3O4	HmIM	Zn2+, Co2+	>450		MRI	3	2.042	85.86	NR	DDS	Q	NR	93	MCF-7 and MDA-MB-231	(94)
CoFe2O4@PDA@ZIF-8	CoFe2O4	HmIM	Zn2+	150	36.4	MRI	1.2	NR	38.3	NR	DDS	DOX and CPT	NR	98 (DOX) and 46 (CPT)	HepG2	(137)
Fe3O4@ZIF-90	Fe3O4	2-ICA	Zn2+	65	7	MRI	7	NR	133.7	NR	DDS under ELF-AMF	5-FU	NR	NR		(95)
Fe3O4@IRMOF-3/FA	Fe3O4	NH2–H2BDC	Zn2+	200	48.9	MRI	1.5	NR	NR	NR	DDS	PTX	123.2	82.04	HeLa and NIH3T3	(101)
Fe3O4@OCMC@IRMOF-3/FA	Fe3O4	NH2–H2BDC	Zn2+	250	51	FOI					DDS	DOX	163	96	L929 and HeLa	(102)
Fe3O4@UiO-66	Fe3O4	NH2–H2BDC	Zr4+	240	51.6	MRI	0.5	NR	255.87	NR	DDS	DOX	2000	NR	3T3 and HeLa	(142)
Fe3O4@UiO-66-NH2-DOX-CDs-Apt	Fe3O4	NH2–H2BDC	Zr2+	16**c		FOI					DDS	DOX	62	NR	HUVEC and MDA-MB-231	(126)
Fe3O4@UiO-66-NH2@WP6	Fe3O4	NH2–H2BDC	Zr4+	40–60	46.8	MRI	7	NR	72.23***d	NR	DDS	5-FU	NR	10.9	HUVEC and HeLa	(225)
Fe3O4@C@P-MOF	Fe3O4	TCPP****e	Fe3+	95	24.5	FOI and MRI	1.2	1.23	72.6	59.0	PTT and PDT	NR	NR	NR	MCF-7 cells	(226)

^aOL = organic linker, MP = metal precursor, D = diagnosis, B = magnetic field, T = therapy, DL = drug loading, LE = loading efficiency, TCL = targeted cell line.

^bcit-γ-Fe₃O₄@MIL-100(Fe)-10% (with a 10 wt % of maghemite content)

^cValue for just Fe₃O₄@UiO-66-NH₂

^dDrug-loaded nanoplatform

^eTCPP: [5, 10, 15, 20-Tetrakis (4-carboxyl)-21H, 23H-porphine]

4. Challenges and Perspectives

Studies over the last 15 years have achieved great development of MOFs and MNPs, as well as their composites, in the biomedical field, including drug delivery, bioimaging, and theragnostics. The main idea behind the development of this type of composites materials is to enhance and exploit the properties of the two components in an additive and/or synergistic manner, combining high biocompatibility (even an intrinsic therapeutic MOF activity) and efficient chemotherapy with complementary magnetic hyperthermia therapies, magnetically guided targeting and/or MRI, among others.

The current synthetic strategies based on magnetic MOF composites have been categorized into: mixing, in situ formation of MNPs in the presence of MOFs, in situ formation of MOFs in the presence of MNPs, and layer-by-layer (LbL) protocols. One of the main challenges is to favor the interaction between the MNP and the MOF. In this sense, strong interactions are required to promote (a) the encapsulation of MNPs within the framework (mixing), (b) the incorporation of MNP’s precursor (MNPs in situ formation), and (c) the growth of a MOF shell around the MNP (in situ MOF growth and LbL). Furthermore, the interactions should lead to a stable composite with the desired properties.

Another key factor is controlling the ratio of MNPs per MOF as well as the MNPs location in the final composite, leading to a lack of reproducibility and homogeneity in the composite synthesis. To date, there is no evidence in the literature on how to control the distribution of the MNPs within a porous matrix by mixing or in situ MNP growth method, promoting their formation exclusively onto the surface or, even, leading to the formation of outer aggregates (close to a physical mixture of MNPs and MOFs) that are difficult or impossible to be lately removed. This is the reason why in situ MOF formation methods with presynthesized MNPs and LbL strategies are much more developed in order to finely tune the morphology, although here the chemical and colloidal stability of the MNPs under the synthetic conditions of the MOF should be considered. In this sense, it remains still unclear the exact conditions necessary to obtain core–shell over non-core–shell structures. Undoubtedly, a more comprehensive understanding of the interactions that favor core–shell versus non-core–shell structures will potentially raise the interest in this topic.

At present, such a knowledge gap impedes its development, reproducibility, and subsequent scale-up. Thus, the current scientific literature is still far from achieving a marketable product suitable for biomedical purposes. To optimize synthetic procedures for such systems, various critical aspects need to be addressed, including: (i) the selection of precursors that are amenable to large-scale production, with particular attention given to their solubility; (ii) the use of water as the reaction media for environmentally friendly conditions, maintaining mild synthetic conditions (temperatures and pressures), compatible with MNPs’ stability; and (iii) the development of convenient washing processes for the large-scale implementation, allowing the nanocomposite recovery and avoiding aggregation phenomena.

As for their application, when employing them as DDSs, the main challenge lies in achieving adequate magnetic properties that can facilitate potential drug release under a magnetic field, such as in the case of MHT. The ultimate goal is to create a nanocomposite that enhances drug transport capabilities beyond what the individual components can achieve alone. While the current work holds great promise, it is important to note that it primarily focuses on a more or less model drug (DOX; with already commercialized nanoformulations) as the main encapsulated active ingredient. The size of the pores within the MOF nanocarrier may limit the range of drugs that can be investigated. One of the challenges toward new drugs for these nanocarriers is understanding the interactions that facilitate and promote drug encapsulation within the porous matrix, and the subsequent manipulation of these interactions to achieve controlled and potentially targeted drug release through the properties of magnetic nanoparticles. Furthermore, it is essential to conduct in vivo studies, not only to demonstrate controlled and targeted drug release but also to evaluate the potential advantages of these nanocomposites over existing drugs in terms of increasing their efficacy while minimizing their adverse effects. Multitude of challenges must be addressed for their widespread clinical use for theragnostic. Nevertheless, the benefits of magnetic nanocomposites for this application could be significant, including the ability to diagnose and treat diseases with greater precision and efficiency, leading to improved patient outcomes.. In this sense, the strategies to address the current limitations and challenges in the development of MNP@MOF composites in the biomedical field include: (1) investigating intrinsic activity of both MNPs and MOFs to promote the formation of highly effective and safe nanocomposites and evaluating the MNP-MOF interactions through advanced characterization techniques (e.g., NMR, HR-TEM, 3D tomography reconstruction); (2) extending synthesis optimization with standard operating procedures, prioritizing biocompatible and available precursors and solvents for a step forward large-scale production; (3) developing nanocomposites that enhance drug transport capabilities beyond what the individual components (MOF or MNPs) can achieve alone, with a focus on understanding the interactions that facilitate and promote drug encapsulation/release by using spectroscopic analysis or computational studies; (4) conducting comprehensive in vitro and in vivo studies to ensure the safety and efficacy of the nanocomposites, as well as to demonstrate controlled and targeted drug release, potentially leading to improved patient outcomes. These strategies aim to overcome the challenges and limitations associated with MNP@MOF composites, ultimately paving the way for their widespread clinical use in theragnostic applications.

In conclusion, the key point is to study the factors that promote the formation of MNP@MOF nanocomposites, subsequently, by investigating the stability, porosity, biocompatibility, and intrinsic activity of both MNPs and MOFs. In this sense, researchers can develop nanocomposites that are highly effective and safe for their use in nanomedicine. After synthesis optimization, such procedures must be extended for reproducible large-scale production while maintaining the desirable properties of the nanocomposites. Finally, both the safety and efficacy of the nanocomposites must be ensured through comprehensive in vitro and in vivo studies to propose them in the future for the biomedical market.

Author Contributions

CRediT: Conceptualization by P.H., D.F.P wrote the review, C.B. and P.H. reviewed it and support the writing. All the authors have accepted responsibility for the entire content of this submitted manuscript and approved submission. CRediT: Darina Francesca Picchi investigation, writing-original draft, writing-review & editing; Catalina Biglione investigation, supervision, writing-review & editing; Patricia Horcajada conceptualization, funding acquisition, project administration, resources, supervision, writing-review & editing.

The authors declare no competing financial interest.