1 Introduction

Metal–organic frameworks (MOFs) represent a promising class of highly ordered crystalline porous coordination polymers (PCPs) [1,2,3]. The extended infinite one-/two-/three-dimensional networks of MOFs are formed by the linkage of inorganic metal (e.g., transition metal and lanthanide metal) ions/clusters as the node and organic ligands (e.g., carboxylates, phosphonates, imidazolates, and phenolates) as the strut. In 1995, the Yaghi group studied selective binding and removal of guest molecules in a microporous MOF composed of 1,3,5-benzenetricarboxylate (BTC) and cobalt cation [4]. In 1999, the same group reported the design and synthesis of MOF-5, which contains 1,4-benzenedicarboxylate (BDC) and Zn4O clusters [5]. MOF-5 showed exceptionally high Langmuir surface area of 2900 m2 g−1. Over the past two decades, owing to extremely high surface area and pore volume, as well as tunable pore size and chemical composition, MOFs have been studied for various applications, including, for example, gas storage and separation [6,7,8,9], chemical separation [10b). The pEGFP-C1@ZIF-8-PEI 25 kD nanoparticles showed a higher transfection efficacy (above 10%) in every dosage. These results were also confirmed by the confocal images (Fig. 10c).

Fig. 10
figure 10

Reproduced with permission from Ref. [127]. Copyright 2019, WILEY-VCH Verlag GmbH & Co. KGaA, Weinheim

a Schematic illustration of the synthesis of pEGFP-C1@ZIF-8 via biomimetic mineralization and pEGFP-C1@ZIF-8-polymer via co-precipitation followed by cellular delivery and expression. b Transfection efficacy of pEGFP-C1@ZIF-8, pEGFP-C1@ZIF-8-PEI 25 kD, and lipofectamine-2000 at different concentrations. c Representative CLSM images of pEGFP-C1 expression in MCF-7 cells.

Full size image

Small interfering RNA (siRNA) was discovered in 1998, offering a new way to combat resistant cancers [128]. MOFs have been proved as effective nanocarriers for siRNA delivery to protect it against clearance or degradation before taking effect in the target cells. The Lin group reported the first use of MOF nanocarriers for the co-delivery of cisplatin and pooled siRNAs to enhance chemotherapeutic efficacy in drug-resistant ovarian cancer cells (SKOV-3 cells) [129]. siRNA was loaded on the surface of UiO-type Zr-MOF nanoparticles through coordination to Zr6 clusters with high loading efficiency (81.6%), while cisplatin prodrug was efficiently encapsulated into the MOF nanoparticles (12.3 wt%). Studies demonstrated the advantages of utilizing MOF nanocarriers to protect siRNAs from nuclease degradation, increase siRNA cellular uptake, and promote siRNA escape from endosomes to silence multidrug resistance genes. Therefore, an order-of-magnitude enhancement of chemotherapeutic efficacy of cisplatin was achieved. Similarly, the Liu group reported the synthesis of MIL-101(Fe) as the nanocarrier to co-deliver pooled siRNAs and selenium(Se)/ruthenium(Ru) nanoparticles to reverse multidrug resistance in Taxol-resistant breast cancer cells (Fig. 11a) [130]. The endosomal escape of siRNA was investigated by confocal laser scanning microscopy. After incubation for 3 h, most of the green fluorescence (siRNAFAM) and red fluorescence (lysosome tracker) in the cytoplasm were separated, suggesting the escape of siRNA from the entrapment of endo-/lysosome to accumulate in the cytoplasm (Fig. 11b). The gene transfection efficiency of Se@MIL-101 and Ru@MIL-101 was measured by EGFP transfection assay in MCF-7/T cells (Fig. 11c). The therapy efficacy was enhanced by the silencing of MDR genes and interference of microtubule (MT) dynamics in MCF-7/T cells. Moreover, high targeting specificity to tumor cells, increased antitumor efficacy, and reduced systemic toxicity in vivo were observed. These studies demonstrated the potential of MOF nanoparticles as a novel nanocarrier platform for co-delivery of chemotherapeutic agents and siRNAs to drug-resistant cancer cells.

Fig. 11
figure 11

Reproduced with permission from Ref. [130]. Copyright 2017, American Chemical Society

a Mechanism of Se/Ru nanoparticles and siRNA co-delivery by MIL-101 for the reversal of drug resistance and induced apoptosis by the disruption of microtubule in MCF-7/T (Taxol-resistant) cancer cells. b Time-dependent confocal microscopy of siRNA escaped from endosomes in MCF-7/T cells. Scale bar: 5 μm. c Fluorescence microscope images of MCF-7/T cells transfected by Se@MIL-101 and Ru@MIL-101 for 24 h.

Full size image

Nucleic acid aptamers usually consist of short strands of oligonucleotides. These oligonucleotide molecules can be engineered to recognize and bind to specific molecular targets such as small molecules, proteins, and nucleic acids [131,132,133]. So far, various aptamers have been selected and widely used as effective molecular probes for cancer study based on their high binding specificity and sensitivity, ease of synthesis, improved storage, as well as lack of immunogenicity [134,135,136,137]. Particularly, the Tan group pioneered the whole-cell systematic evolution of ligands by exponential enrichment (cell-SELEX) approach for high-affinity aptamer selection [138,139,140,141,142]. This method allows for the selection of aptamers against specific cell lines to accelerate the discovery of biomarkers (Fig. 12). So far, the group has successfully selected a series of aptamers through the cell-SELEX method. For example, aptamers have been selected against leukemia [143], lung cancer [144], and cells infected with the Vacinia virus [145], as well as aptamers specific for phosphorylation epitopes of tau protein [146].

Fig. 12
figure 12

Reproduced with permission from Ref. [139]. Copyright 2009, American Chemical Society

Schematic representation of the cell-SELEX approach for aptamer selection.

Full size image

With the development of aptamer selection for molecular medicine, MOF nanocarriers for aptamer delivery have been investigated during the last few years, taking advantage of the unique properties of aptamers. For instance, Fang and coworkers demonstrated that AS1411 aptamer-functionalized UiO-66@AgNCs@Apt can be internalized effectively by target cancer cells (MCF-7 cells) with high selectivity through AS1411-mediated endocytosis [147]. Upon one-pot incorporation of the anticancer drug DOX, this drug delivery system exhibited high capability for targeted delivery and intracellular controlled release, resulting in enhanced antitumor effect in vitro.

Several efforts have been made toward controlled release of drugs utilizing aptamer-functionalized MOF nanoparticles. This was achieved by designing MOFs responsive to different triggers, e.g., ATP and glucose. The Willner group modified the external surface of MOF nanoparticles (UiO-68) with ATP-AS1411 hybrid aptamer in caged configurations [148]. ATP is upregulated in cancer cells, while AS1411 aptamer identifies the nucleolin receptor sites on the cancer cell membrane. In the presence of ATP, the MOFs were unlocked by ATP–aptamer complex formation, releasing the loaded drug molecules (DOX). Experiments revealed high cytotoxic efficacy and highly selective permeation of these dual aptamer-modified MOF nanocarriers into MDA-MB-231 breast cancer cells as compared to MCF-10A normal epithelial breast cells. The group subsequently designed glucose-responsive MOF nanocarriers for controlled release of drugs [149]. ZIF-8 nanoparticles were loaded with glucose oxidase (GOx) and antivascular endothelial growth factor aptamer (VEGF aptamer). Upon GOx-mediated aerobic oxidation of glucose, the products gluconic acid and H2O2 acidified the microenvironment and caused pH-induced degradation of MOFs to release drugs (Fig. 13a). The VEGF aptamer could potentially inhibit angiogenic regeneration of blood vessels. The loadings of VEGF aptamer and GOx were confirmed by confocal microscopy imaging (Fig. 13b). Panels I and II suggested that the Cy3-modified VEGF aptamer (red) and the coumarin-functionalized GOx (blue) were successfully incorporated into ZIF-8. The Fan group reported immunostimulatory DNA–MOFs (isMOFs) containing cytosine–phosphate–guanosine (CpG) oligonucleotides, which exhibited high cellular uptake, organelle specificity, and spatiotemporal control of Toll-like receptors (TLR)-triggered immune responses [150].

Fig. 13
figure 13

Reproduced with permission from Ref. [149]. Copyright 2018, American Chemical Society

a Mechanism of glucose-driven release of VEGF aptamer from ZIF-8 caused by degradation of MOFs under local acidified conditions created by GOx-catalyzed aerobic oxidation of glucose to gluconic acid. b Confocal microscopy images of ZIF-8 loaded with Cy3-modified VEGF aptamer (I) and coumarin-functionalized GOx@ZIF-8 (II), and the bright field and merged image of the loaded MOF (III and IV, respectively).

Full size image

3.3 Proteins

Proteins are macromolecules consisting of one or more long chains of amino acid residues. They serve a large number of functions, such as DNA replication, metabolic reaction catalysis, and molecular transport. Since proteins have large size, charged surface, and environmental sensitivity, it is difficult for proteins to naturally cross cell membranes without losing structural integrity. In order to utilize proteins for therapeutic purposes, MOF nanoparticles for intracellular delivery of proteins have attracted increasing attention in recent years.

For example, Farha et al. selected NU-1000 and PCN-222/MOF-545 nanoparticles as the host for insulin encapsulation [151]. The surface of MOFs was modified with phosphate-terminated nucleic acids for increased colloidal stability and cellular uptake. Compared to the native protein, a tenfold enhancement of cellular uptake was achieved. The Zheng group synthesized a pH-sensitive nanocomposite with a core–shell structure as the drug delivery system [152]. Biocompatible bovine serum albumin (BSA) and DOX (BSA/DOX) core was protected by the ZIF-8 shell. The BSA/DOX@ZIF-8 showed greater antitumor efficacy than that of free DOX against breast cancer cell line MCF-7.

Recently, Mao and coworkers have developed zeolitic imidazole framework-90 (ZIF-90) as a general platform to deliver different proteins into the cytosol, independent of their size and molecular weight [153]. Protein encapsulation was performed by self-assembly of imidazole-2-carboxaldehyde, Zn2+, and the protein (Fig. 14a). Degradation of nanoparticles to release protein was observed in the presence of ATP. HeLa cells were treated with ZIF-90/GFP nanoparticles for cellular uptake study. According to the flow cytometry analysis, the cellular uptake of ZIF-90/GFP increased proportionally with the concentration of GFP increasing from 40 to 100 μg mL−1 (Fig. 14b). Next, different endocytosis inhibitors were selected for pretreatment. Among them, only sucrose reduced the cellular uptake efficiency significantly (down to 17%) (Fig. 14c), indicating that ZIF-90/GFP is mainly internalized via clathrin-mediated endocytosis. After incubation of HeLa cells with 50 μg mL−1 ZIF90/GFP nanoparticles, a significant accumulation of GFP in the cytosol was observed by CLSM imaging (Fig. 14d). Furthermore, ZIF-90/protein nanoparticles were used to successfully deliver cytotoxic RNase A for tumor cell growth inhibition, as well as genome-editing protein Cas9 to knock out the green fluorescent protein (GFP) expression of HeLa cells.

Fig. 14
figure 14

Reproduced with permission from Ref. [153]. Copyright 2019, American Chemical Society

a Schematic illustration of the synthesis of ZIF-90/protein nanoparticles and ATP-triggered protein release in the cell. b Cellular uptake efficiency of ZIF-90/GFP. c Cellular uptake efficiency of ZIF-90/GFP in the presence of different endocytosis inhibitors. d CLSM images of HeLa cells treated with ZIF-90/GFP. LysoTracker Red was used for endosome/lysosome staining. Scale bar: 10 μm.

Full size image

Enzymes are a class or proteins that can catalyze many complex reactions in organisms with high selectivity. So far, enzyme–MOF composites have been widely studied for catalysis, sensing, and detection [154,155,156,157]. Recently, cellular delivery of enzymes by MOF nanocarriers for cancer therapy has been reported by the Zhou group. These MOFs showed better selectivity and less systemic toxicity than conventional chemotherapy [158]. Tyrosinase was encapsulated into PCN-333(Al) (TYR@PCN-333) to form an enzyme–MOF nanoreactor to activate the cancer prodrug paracetamol (APAP). The reaction generated reactive oxygen species (ROS) and depleted glutathione (GSH), inducing cytotoxicity in drug-resistant cancer cells. Compared to free enzyme, the MOF nanocarrier provided protection against enzyme deactivation and thus extended the antitumor efficacy of TYR@PCN-333.

3.4 Challenges

Although remarkable achievements have been made to apply MOFs for drug delivery, several challenges still exist in this field. First, only limited studies on the kinetics of drug loading and release have been reported so far. Recently, the Horcajada group has demonstrated that the drug loading process is governed by the accessibility of cages in MOFs, while the loading capacity is influenced by the hydrophobicity/hydrophilicity of MOFs and the drug molecules [159]. For instance, the loading rates of hydrophilic acetylsalicylic acid (AAS) and hydrophobic isobutylphenylpropanoic acid (IBU) into UiO-66 are 0.0301 and 0.0295 M h–1, respectively. However, higher total drug loading capacity of IBU (35.5%) was observed compared to that of AAS (25.5%). It is worth noting that the solvent may also affect the drug loading rate. According to their studies, both the structure of MOFs and the hydrophobic/hydrophilic nature of the drug molecules could affect the rate of drug release. For example, a faster release of AAS from the open-structured MIL-100 (1 day) was observed, while a slower AAS release from the narrow 1D pore system of MIL-127 (6 days) was detected. Furthermore, a mismatch in hydrophobicity and hydrophilicity could result in a fast drug release. For example, hydrophilic AAS underwent a quick release from hydrophobic UiO-66 (1 day).

Another major challenge for clinical applications of MOF-based DDS is the potential toxicity. However, the existing literature is very limited and insufficient to draw a conclusion about the toxicity of MOF nanoparticles. So far, many in vitro toxicity studies have been conducted on different cell lines, making it very difficult to compare the obtained results. For instance, nanoZIF-8 (200 nm) was evaluated against three human cell lines, namely NCI-H292, HT-29, and HL-60. Results suggested that nanoZIF-8 is nontoxic to these cells [101]. However, in another report, nanoZIF-8 (90 nm) showed cytotoxicity toward HeLa and J774 cell lines [160]. Recently, the in vivo toxicity of nanoscale MOFs has been assessed against zebrafish embryos [161]. The study revealed that the toxicity of MOFs was mainly attributed to the leached metal ions. In contrary, three different Fe(III)-based MOF nanoparticles (MIL-88A, MIL-100, and MIL-88B_4CH3) were injected in rats at high doses. The results suggested that these MOF nanoparticles exhibited low acute toxicity and were rapidly sequestered by liver and spleen. According to the studies by Baati et al., the MOF nanoparticles could undergo further biodegradation and elimination in urine or feces without metabolization and causing significant toxicity [162]. In order to reach the clinical development stage of MOF nanoparticles, the performance of MOF-based DDS should be optimized for preclinical evaluation by conducting systematic in vivo studies on their stability, degradation mechanics, and side effects on normal organs.

4 Conclusions and Perspectives

During the past few decades, MOFs have been extensively studied for a variety of applications by their well-defined structure, high surface area, high porosity, tunable pore size, and easy functionalization. In particular, exploring MOFs as a nanocarrier for drug delivery in biomedical applications has attracted great interest in recent years. Currently, various molecules have been investigated as the therapeutic agents for disease treatment, such as anticancer drugs, nucleic acids, and proteins. In this review, we summarized four strategies commonly used to functionalize MOFs with therapeutic agents for drug delivery. They include surface adsorption, pore encapsulation, covalent binding, and functional molecules as the building block. The van der Waals interaction, π–π interaction, and hydrogen bonding are the main forces involved in surface adsorption and pore encapsulation approaches. Functional molecules are covalently bound to the framework through inorganic metal clusters or organic linkers by the covalent binding method. Moreover, functional molecules can be incorporated into the framework as organic ligands. Then, we thoroughly discussed recent progress of biological applications of MOF nanocarriers for drug delivery. Benefiting from unique advantages of MOFs, many drug molecules have been efficiently delivered by MOF nanoparticles. Among them, drugs, nucleic acids, and proteins were selected for discussion in this section.

Despite remarkable achievements made in this field, several challenges remain to be solved. First, although many functionalization methods have been reported, they all possess some limitations. For instance, molecules incorporated by surface adsorption and pore encapsulation tend to leak gradually owing to weak interaction forces. Covalent binding provides stronger interactions, but it requires complex synthetic procedures and may influence the activity of functional molecules. On the other hand, the organic ligands suitable for MOF synthesis are usually rigid and highly symmetrical, which makes it difficult to directly utilize biomolecules as the building block. Such limitations call for the development of advanced functionalization strategies to incorporate a wide variety of potential therapeutic agents into MOFs to explore their clinical applications. Second, the kinetics of drug loading and release, in vivo toxicity, degradation mechanism, and pharmacokinetics of MOF nanoparticles are still under study. Further investigations are required to rationally design MOF–drug conjugates with enhanced biostability, biocompatibility, and therapeutic efficacy. In conclusion, MOFs possess unique properties and show great promise for intracellular drug delivery to treat diseases. In the future, efforts should be focused on overcoming the noted challenges to fully realize the potential of MOFs as drug delivery systems in clinical applications.