2D Covalent Organic Framework Direct Stem Cell Differentiation

Abstract

Two–dimensional (2D) covalent organic frameworks (COFs) are emerging class of crystalline porous organic polymer with wide-range of potential applications. However, poor processability, aqueous instability, and low water dispersibility greatly limit their practical biomedical implementation. Herein, we report a new class of hydrolytically-stable 2D COFs for sustained delivery of drug to direct stem cell fate. Specifically, a boronate–based COF (COF-5) is stabilized using amphiphilic polymer Pluronic F127 (PLU) to produce COF-PLU nanoparticle with thickness of ~25 nm, and diameter ~200 nm. These nanoparticles are internalized via clathrin mediated endocytosis and have high cytocompatibility (half inhibitory concentration ~1 mg/mL). Interestingly, the 2D COFs induces osteogenic differentiation in human mesenchymal stem cells, which is unique. In addition, we are able to load osteogenic agent – dexamethasone within the porous structure of COFs for sustain delivery which further enhance the osteoinductive ability. Our results demonstrate for the first time, fabrication of hydrolytically-stable 2D COFs for sustained delivery of drugs and demonstrate its osteoinductive characteristics.

Graphical Abstract

We report osteoinductive ability of hydrolytically-stable two-dimensions (2D) covalent organic frameworks (COFs) to direct stem cell fate. The high porous network of 2D COFs can be used to load therapeutics (dexamethasone) to promote formation of mineralized matrix.

INTRODUCTION

Two-dimensional (2D) materials have gained significant attention due to their unique layered structure and property combination that have widespread application in electronics, catalysis, sensing, and drug delivery.^{1, 2} Among these, 2D covalent organic frameworks (COFs), an emerging class of crystalline porous organic polymers, have attracted significant attention due to their crystallinity, ordered and tunable porous structure, and high specific surface area.^{3, 4} COFs are synthesized via reversible condensation of metal-free molecular building blocks,⁵ composed of light elements like C, H, B, O, and N, which bestows biodegradability and biocompatibility characteristics indicating their potential in biomedical applications^{6, 7}. However, the difficulty in processing COFs into nanosized⁸ along with their poor physiological stability^{9, 10} limited their application in drug delivery and regenerative medicine. While a great deal of interest is being invested to produce ‘nano’dimensional COFs, significant advancement has been made to enhance the chemical stability of COFs via chemical linkage modification^{11, 12}, pore alkylation¹⁰ and surface engineering¹³. However, many of these approaches results in significant compromise of structural porosity of COFs.^{10, 13} Thus, there is a need to devise new approaches which can impart physiological stability to nanoscale COFs, while maintaining biocompatibility characteristics.

Polymers have widely been integrated with nanoparticles¹⁴ or porous materials^{15, 16} to impart dispersibility or stimuli responsiveness. Integration of polymers with porous metal-organic framework (MOF) via in situ encapsulation of polymers in nanochannels of MOF or post synthetic surface modification^{17, 18}, often produces advanced functional composites exhibiting exciting properties that are superior to those of the individual components.^{19, 20} Integrating polymers with COFs has just begun majorly focusing on polymer assisted exfoliation of bulk COFs.^{21, 22} However, the strategy of enhancing hydrolytic stability of COFs via integrating colloidal COFs with amphiphilic polymer has not been reported so far. Recently, Ditchel and co-worker has developed a strategy of producing colloidal COF-5 with a hexagonal, 2D layered structures via homogenous polymerization²³ while amphiphilic polymers have long been used as phase-transfer agent for dispersing hydrophobic nanoparticles in water.²⁴ To this end, we envisaged that post-synthetic integration of a colloidal COF-5 with an amphiphilic polymer may impart hydrolytic stability and water dispersibility, while maintaining their drug loading efficacy and thereby, enabling biomedical exploration of such COF-based nanocarrier.

Several 2D nanomaterials have been explored for drug delivery and tissue engineering including graphene²⁵, synthetic silicate clays^{26, 27}, transition metal dichalcogenides/oxides (TMDs/TMOs)²⁸, hexagonal boron nitride (h-BN), ultrathin black phosphorous (BP) monolayers, and graphitic carbon nitride (g-C₃N₄) are developed in past decade.^29–31 However, only few of these 2D nanomaterials are bioactive and can induce stem cell differentiation.^{32, 33} In addition, very few of these nanomaterials are able to deliver drugs without any chemical modification.^{34, 35} As 2D COFs are new class of biomaterials, our proposed approach will investigate their application for regenerative medicine. In addition, 2D COFs have highly ordered porosity, biodegradability, and cytocompatibility, which make them ideal candidate for drug delivery.^{6, 36, 37} Overall, this is the first report to investigate 2D COFs for regenerative medicine.

Herein, we report a new 2D COF-polymer nanoparticle for drug delivery by combining colloidal COF-5 and amphiphilic polymer. COF-5 is boronate-based covalent organic framework³⁸ which has not yet been explored in drug delivery due to their high hydrolytic susceptibility. We have first synthesized colloidal COF-5 and functionalized post-synthetically with Pluronic F127 (PLU) to obtain hydrolytically stable 2D nanoparticles of COF-PLU (Figure 1). Interestingly, these 2D nanoparticles COF-PLU induce osteogenic efficacy and sustained release profile of dexamethasone (Dex). The addition of Dex to COF-PLU nanoparticles significantly enhances the osteogenic property of human mesenchymal stem cells (hMSCs). Overall, this study demonstrates a facile strategy for fabricating physiologically stable class of COF based 2D nanoparticles for nanomedicine. To the best of our knowledge, this is the first report demonstrating osteoinductive ability of COFs. In addition, COFs based nanomaterials exhibit high loading and thus can be used for wide range of applications in regenerative medicine and drug delivery.

Figure 1.

Schematic showing fabrication of hydrolytically-stable COF-PLU nanomaterials for sustained delivery of therapeutics (Dex – dexamethasone) for directing stem cell differentiation.

RESULTS

Synthesis and characterization of COF-PLU nanoparticles

COF-PLU are prepared via complexation of colloidal COF-5 and Pluronic F-127 (PLU) followed by their reconstitution in distilled water. First, colloidal COF-5 are synthesized²³ via solvothermal condensation of 2,3,6,7,10,11-hexahydroxytriphenylene (HHTP) and benzene 1,4-diboronic acid (BDBA) in a mixed solvent containing dry acetonitrile, dioxane, and mesitylene in appropriate ratios (Figure 2A). 2D nature of colloidal COF-5 nanoparticles is confirmed by dynamic light scattering (DLS) and atomic force microscopy (AFM), indicating an average particle size ~147 nm with a narrow particle size distribution (PDI = 0.202) and height of ~15 nm (Figure S1). Prior to examine their potential in drug delivery, stability of COF-5 in aqueous solution is examined. Hydrolytic stability study of the colloidal COF-5 evaluated by monitoring turbidity of colloidal suspension over time in presence of added water indicated rapid degradation of COF-5 (Figure 2B). Toward developing a hydrolytically stable water dispersible formulation of COF-5 colloidal COF-5 is complexed with amphiphilic polymer Pluronic F127 (PLU). PLU is an amphiphilic block copolymer containing a hydrophobic block of poly(propylene glycol)) at the center and hydrophilic chains of poly(ethylene glycol) at edges. Addition of Pluronic F127 solution (in acetonitrile) to colloidal suspension of COF-5 (hydrophobic), result in formation of COF-5-PLU composite due to hydrophobic-hydrophilic interactions (Figure 2B, inset). Specifically, poly(propylene glycol) wrap COF-5 to reduce the overall surface energy, which render poly(ethylene glycol) at edges. The presence of poly(ethylene glycol) at edges provide steric stability to these composite COF-PLU nanoparticles.

Figure 2. Synthesis and characterization of 2D COF-PLU nanoparticle.

(A) Schematic showing solvothermal condensation of HHTP and BDBA producing colloidal COF-5 which are unstable in physiological condition. Wrapping of COF-5 with an amphiphilic polymer (pluronic F127 (PLU)) produces a stable COF-PLU nanoparticles. (B) Hydrolytic stability of COF-5 indicating its rapid degradation in water. Inset show complexation of colloidal COF-5 with an amphiphilic polymer pluronic F127 (PLU) forms self-assembled nanoparticles (COF-5 + PLU) which precipitated out. (C) Dynamic light scattering profiles (blue histogram) and transmission electron micrographs (inset) show size of COF-PLU nanoparticles. (D) Atomic force micrographs of COF-PLU reveal the thickness to be ~ 20–25 nm. E) Transmission electron micrographs of COF-PLU nanoparticles at day 1, 3 and 7 indicating their stability in water. (F) ATR-FTIR profile of COF-PLU reveals presence of characteristic peaks for both COF-5 (1347 & 1332) and PLU (1100) in COF-PLU. G) Thermogravimetric analysis profile of COF-PLU revealing presence of two consecutive weight loss events at ~400 °C (characteristics of PLU) and ~650 °C (characteristics of COF-5). The differential temperature is then plotted against time and is shown as differential thermogravimetric (DTG) analysis.

The hydrodynamic diameter of the COF-PLU 2D nanoparticles are found to be ~180 nm as indicated by dynamic light scattering study and transmission electron micrographs (Figure 2C). Atomic force micrographs (AFM) reveals that thickness of nanoparticle is ~25 nm (Figure 2D). Stability of the COF-PLU nanoparticles over time is examined by monitoring the sizes of nanoparticles at different time points using TEM (Figure 2E and S2). No significant changes in sizes are observed till day 7. The nanoparticles also exhibit negative zeta potential values less than −10 mV with time indicating their stability in aqueous suspension (Figure S2). It is worth mentioning that three mass ratios of COF and PLU as COF:PLU (w/w) as 1:1, 1:5, and 1:25 are examined and 1:5 ratio of COF:PLU shows smallest nanoparticle. Although 1:1 (w/w) ratio of COF:PLU forms complex in the mixed solvent, the nanoparticles do not exhibit stable behavior in DLS study whereas the hydrodynamic diameter of COF-PLU nanoparticles with 1:25 (w/w) of COF:PLU are much bigger in size (~383 nm) than that with 1:5 (w/w) of COF:PLU (Figure S2c). It is presumably because that 1:1 w/w ratio of COF:PLU may not be enough to ensure the complete hydrolytic stability while 1:25 ratio of COF:PLU undergoes aggregation due to high content of PLU. Collectively, DLS, TEM and AFM reveals that resuspension of aggregated composite in water lead to formation of disk-shaped stable 2D nanoparticle of COF-PLU with diameter ~200–300 nm and thickness of ~25 nm.

The chemical composition of COF-PLU nanoparticles is examined via attenuated total reflection (ATR)-Fourier transform infrared spectroscopy (ATR-FTIR) and thermogravimetry analysis (TGA). The characteristics bands of functional boronic ester rings (-B-O-)³⁸ at 1347 cm⁻¹ and 1332 cm⁻¹, while for Pluronic F-127 at 1110 cm⁻¹ are present in the FTIR profile of COF-PLU nanoparticles, indicates presence of both COF-5 and PLUF-127 (Figure 2F). Similarly, TGA profiles of COF-PLU nanoparticles (Figure 2G) reveals two consequent weight loss events at ~400 °C and ~650 °C corresponding to the weight loss events of Pluronic F-127 and COF-5, respectively. The differential temperature is then plotted against time and shown as Differential Thermogravimetric Analysis (DTA) curve. The extent of weight loss at ~400 °C and at ~650 °C observed for COF-PLU nanocomposite validates the weight ratio of PLU and COF-5 (5:1) in COF-PLU nanoparticles. The tiny depression at ~100 °C in the TGA curves of COF-PLU and COF is due to the evaporation of adsorbed water. Taken together, the FTIR and TGA studies confirm chemical composition of the new class of COF based 2D nanomaterials. Overall, a water dispersible COF based 2D nanomaterial (COF-PLU) is developed via hydrophobic interaction based post-synthetic integration of colloidal COF-5 with amphiphilic polymer Pluronic F-127.

Cellular interactions of COF-PLU nanoparticles

The cytocompatibility of 2D COF-PLU is evaluated using human mesenchymal stem cells (hMSCs) using conventional MTT assay. More than 85% of hMSCs are alive till 0.1 mg/mL of COF-PLU nanoparticles and half inhibitory concentration (IC₅₀) is found to be ~1 mg/mL (Figure 3A) indicating significantly less cytotoxicity compared with other 2D nanomaterials such as graphene (IC₅₀ ~100 μg/mL).³² To examine whether or not treatment of COF-PLU nanoparticles at lower concentration induces any cell cycle arrest, flow cytometry-based cell cycle analysis is performed. Cell cycle of hMSCs is first synchronized by serum starvation. Then such hMSCs are treated with varying concentrations of COF-PLU nanoparticles ranging from 0.015 mg/mL to 0.3 mg/mL. No significant changes in cell populations are observed when the cells are treated with COF-PLU nanoparticles (Figure 3B) which indicates that COF-PLU nanoparticles do not influence any cell cycle event in hMSCs.

Figure 3.

Cellular interaction of COF-PLU. (A) Cytocompatibility of COF-PLU is evaluated by subjecting hMSc with varying concentration of COF-PLU nanoparticles. Half-maximal inhibitory concentration (IC₅₀) was determined by concentration of COF-PLU nanoparticles to inhibit cell viability by half. (B) Cell cycle analysis of hMSCs treated with COF-PLU nanoparticles (15, 200 and 300 μg/mL) indicating no significant changes in cell cycle upon treatment with COF-PLU nanoparticles. (C) Fluorescence micrographs of hMSC treated with FITC-tagged COF-PLU nanoparticles indicate their cellular internalization. (D) Schematic for proposed cellular internalization pathway of COF-PLU nanoparticles. (e) Flow cytometry profiles of FITC-tagged COF-PLU nanoparticles confirming clathrin-mediated endocytosis as major cellular internalization pathway of COF-PLU (n = 3 in each case).

Next, to examine cellular internalization of COF-PLU nanoparticles, a 4 h cellular uptake study is performed using FITC-tagged COF-PLU nanoparticles. A significant cellular attachment and uptake of FITC tagged COF-PLU nanoparticles (green) is observed (Figure 3C). As COF-PLU can interact with protein and agglomerate, some of these nanoparticles are present on or around cell membrane. We envisaged the 2D nanoparticles of COF-PLU undergoes surface mediated endocytosis majorly via clathrin mediated pathway (Figure 3D) like other 2D nanoparticles.³⁹ To validate this hypothesis, we performed pathway inhibition studies. Specifically, we blocked cellular internalization pathway using pathway specific inhibitors and then monitor nanoparticle uptake using flow cytometry (Figure 3E). We used inhibitors for clathrin-mediated endocytosis, caveolae-mediated endocytosis and micropinocytosis. Interestingly, ~53% of inhibition of nanoparticle uptake was observed when hMSCs are treated with clathrin inhibitor, indicating clathrin-mediated endocytosis was the main pathway for cellular internalization of COF-PLU nanoparticles. Only~3% of uptake of COF-PLU nanoparticles was observed via caveolae mediated endocytosis. Notably, ~19% of COF-PLU nanoparticles were internalized into hMSCs via micropinocytosis which may be due to some extent of non-specific interaction between hMSCs and larger COF-PLU nanoparticles. Collectively, we observed that COF-PLU nanoparticles internalized into hMSCs mainly via clathrin-mediated endocytosis.

Therapeutic (dexamethasone) loaded COF-PLU nanoparticles

After gaining an insight on cytocompatibility of COF-PLU nanoparticles in hMSCs we proceed to explore their potential as delivery vehicle in hMSCs. The hydrophobic cores of COF-5 prompt us to explore loading of a hydrophobic drug, Dexamethasone (Dex). First, Dex loaded COF-PLU nanoparticles are prepared by co-incubation of Dex with colloidal COF-5 in a varying ratio (0.1:1, 0.25:1, 0.5:1, 1:1, w/w) followed by addition with Pluronic F127 and resuspension of the precipitate in sterilized DI water. TEM of the Dex loaded COF-PLU nanoparticles reveals that their hydrodynamic diameter (~200 nm) does not alter significantly upon loading of Dex (Figure 4A). Dex loading in COF-PLU nanoparticles are confirmed by ATR-FTIR spectroscopy (Figure 4B). The co-existence of all the characteristics bands for Dex, COF and PLU in Dex loaded COF-PLU such as bands at 1660 cm⁻¹ for the carbonyl group conjugated to alkene (characteristics band for Dex), bands at 1347 cm⁻¹ and 1332 cm-1 for the functional boronic ester rings (characteristics bands for COF-5) and bands at 1110 cm⁻¹ (that for Pluronic F-127) indicate the presence of Dex in COF-PLU nanoparticle.

Figure 4.

Synthesis and characterization of therapeutic loaded COF-PLU nanoparticles. (A) Transmission electron micrograph of dexamethasone (Dex) loaded COF-PLU (COF-PLU/Dex). (B) FTIR profile of COF-PLU, dexamethasone, dexamethasone loaded COF-PLU nanoparticles (COF-PLU/Dex). (C) Entrapment efficiency as well as loading capacity of dexamethasone on COF-PLU nanoparticles. (D) Release profile dexamethasone from COF-PLU/Dex at different pH to mimic extracellular (pH ~7.4) and intracellular microenvironment (pH ~5 (endosome/lysosome)). (E) COF-PLU nanoparticles induce osteogenic differentiation of hMSCs and sustained delivery of dexamethasone (Dex) from Dex loaded COF-PLU (COF-PLU/Dex) further enhance the osteogenic differentiation. hMSCs are cultured in osteoconductive (OC) media and treated (n = 3) with COF-PLU, Dex (exogenous), and COF-PLU/Dex. hMSCs in OC media is used as control. ALP expression (staining) and activity is measured at Day 7 and 14 using BCIP/NBT staining and ALP kinetic assay. COF-PLU/Dex show strong ALP expression and activity on both day 7 and 14. (F) Treatment with COF-PLU/Dex enhances the expression of RunX2, a key osteogenic protein as indicated by western blot analysis. (G) Significanlty higher production of mineralized matrix is observed in hMSCs treated with COF-PLU/Dex compared to control or dexamethasone treated hMSCs. *P-value < 0.05, **P-value < 0.01, ***P-value < 0.001, and ****P-value < 0.0001.

Next, Dex loading, and release behavior of COF-PLU nanoparticles are evaluated. High Dex loading efficiency (>80%) and drug loading capacity (~14%) are observed up to COF-5:Dex ratio (w/w) of 0.5:1, w/w) (Figure 4C) which falls sharply at COF-5:Dex 1:1 weight ratio presumably due to overcrowding of dexamethasone molecules in porous network of COF-5. The internal regular porosity of COF-5 as well as their low density may contribute to the high drug loading efficiency. Dex release from COF-PLU nanoparticles was evaluated over a period of 7 days using COF-PLU nanoparticles containing 25% of Dex (COF:Dex 1:0.25, w/w). A biphasic release profile with an initial burst release followed by sustained release for an extended period of time is observed (Figure 4D). The initial burst release may be attributed to the loosely attached Dex molecules on the surface of COF-PLU nanoparticles, while sustained release of Dex can be attributed to strong interaction with inner pores of COF-5. As expected, the Dex release rate is higher at lower pH (pH ~5) than that at physiological condition (pH~7.4) presumably due to the acid sensitive boronate linkage at the core of 2D COFs. Similar release profiles were observed for other polymeric drug carrier.^40–42

We then proceed toward evaluating the effect of Dex loaded 2D COF-PLU to direct stem cell differentiation. Dexamethasone is a synthetic glucocorticoid that, in optimum concentration, can stimulate differentiation of hMSCs toward osteogenic lineage in presence of ascorbic acid and β-glycerophosphate. To maintain the optimum concentration at site of bone regeneration many biodegradable sustained drug delivery platforms have been explored.^43–45 Also, stem cell differentiation is a complex developmental phenomenon which is often controlled by several biochemical and biophysical cues including chemicals, stiffness, and topology of the microenvironment. To this end, we envisaged that sustained release of Dex from Dex loaded 2D COF-PLU may enhance the efficacy of osteogenic differentiation in hMSCs compared to exogenous addition of Dex. To investigate this, hMSCs are divided into four treatment groups: COF-PLU nanoparticles, Dex loaded COF-PLU (COF-PLU/Dex), positive control of exogenous addition of 100 nM of Dex (Dex) and untreated hMSCs (negative control). All the treatments for differentiation study are performed in osteoconductive media (in absence of dexamethasone or bone morphogenic protein −2 (BMP-2)).

The osteogenic differentiation of treated hMSCs are investigated by monitoring the alkaline phosphatase (ALP) activity of hMSCs and the production of mineralized matrix as measured by Alizarin red staining (ARS). ALP is one of the major early osteogenic marker and upregulation of ALP activity is a key event occurring during the early time points of osteogenesis. The production and activity of ALP are evaluated at day 7 and 14 post culture by using NBT/BCIP staining as well as ALP activity assay. Importantly, two-fold enhancement of ALP activity is observed at Day 7 in COF-PLU/Dex compared with Dex treated and untreated groups as indicated by BCIP/NBT staining and ALP kinetic assay (Figure 4E). Interestingly, a significant ALP activity is also observed at day 7 and day 14 when the hMSCs are treated with empty nanoparticles of COF-PLU. To elucidate stimulatory effect of COF-PLU 2D nanoparticles the expression level of a key osteogenic gene, Runx2, is measured (Figure 4F). As compared to the control group the expression level of RunX2 is significantly enhanced in hMSCs treated with empty nanoparticles of COF-PLU supporting their osteogenic effect.

To assess production of mineralized extracellular matrix, which is a late osteogenic event, alizarin red staining (ARS) is performed on day 14. Significantly enhanced matrix mineralization is observed in hMSCs treated with COF-PLU/Dex nanoparticle compared to that treated with exogenous Dex (Figure 4G). Interestingly, a significant matrix mineralization is also observed when hMSCs are treated with empty COF-PLU nanoparticle compared to the untreated group indicating distinct osteogenic activity of COF-PLU. Although the bioactivity of COF-PLU is lower compared to Dex only group, which warrant further mechanistic investigating on bioactivity of COF-PLU nanoparticles. Collectively, COF- PLU/Dex shows enhanced osteogenic potential compared to positive control (Dex) presumably via combined osteogenic effect of dexamethasone and 2D COF-PLU.

Mechanistic insight about inherent bioactivity of COF-PLU nanoparticles

To examine whether the osteogenic activity of COF-PLU nanocomposite is mediated via biochemical or biophysical interaction with hMSCs, ALP activity is assessed using COF-PLU nanoparticles as well as its individual ingredients COF-5 and PLU. hMSCs are separately treated with COF-PLU nanoparticles, equivalent amount of COF-5 and PLU for day 7. As expected, much enhanced ALP activity is observed for COF-PLU treatment group compared to positive control (Figure 5A). However, no significant ALP activity is observed in hMSCs treated with only COF-5 or only PLU at this concentration. This indicates that the observed osteogenic activity of COF-PLU nanoparticles in hMSCs is mediated via biophysical interaction. Next, plausible pathways associated with the osteogenic activity of COF-PLU nanoparticles is determined by measuring ALP activity of hMSCs treated with COF-PLU in presence of inhibitors of major osteogenic pathways Wnt, TGF-β, and MAPKs (particularly ERK, JNK & p- 38). No noticeable difference in ALP activity is observed at day 4 (Figure 5B). However, a significant reduction of ALP activity is observed when hMSCs are co-treated with COF-PLU and inhibitors of JNK-pathway for day 8 and day 12 (Figure 5B) compared with hMSCs treated with only COF-PLU. This observation indicates the association of JNK pathway in the biophysical interaction-based osteogenic activity of COF-PLU 2D nanoparticles as depicted in the schematic (Figure 5C).

Figure 5.

Mechanistic insight of osteogenic activity of COF-PLU nanoparticles. (A) Relative ALP activity of hMSC treated with COF-PLU, COF-5, and PLU at day 7. hMSCs treated with and without dexamethasone is used as positive and negative controls respectively (n = 3 in each group, **P-value < 0.01, ***P-value < 0.001). (B) To investigate the mechanism of osteoinductive characteristics of COF-PLU, hMSCs are treated with inhibitors of major osteogenic pathways (Wnt, TGF-β, JNK, p-38, and ERK) in presence and absence of COF-PLU. ALP activity of hMSCs was determined at day 4, 8, and 12 (n = 3 in each group, **P-value < 0.01, ***P-value < 0.001). (C) Schematic showing ability of COF-PLU trigger JNK pathway leading to osteogenic differentiation via upregulation of RunX2.

DISCUSSION

Despite the significant advances in biomaterials there is still need for developing new material that not only can act as drug delivery system but also can actively intervene tissue regeneration. For instance, liposomes,⁴⁶ dendrimers,⁴⁷ silica nanoparticles,⁴⁸ etc. have been used for delivering dexamethasone to induce osteogenic differentiation in stem cells. With the tremendous advances in materials science, recent research focusses on developing bioactive materials that can actively intervene the tissue regeneration. For instances, silica nanoparticles, graphene or synthetic soft polymers e.g. dendrimer are being explored as bioactive materials to induce osteogenic differentiation ^{33, 49},^{50, 51} However, some of them are raising concern in their biocompatibility, biodegradability, size dependent cytotoxicity and genotoxicity.^{33, 49} To this end, covalent organic framework (COF) which is the latest member of porous organic polymer, although in its nascent state, may hold significant promise as drug carrier because of its cytocompatibility and biodegradability⁵² while its inherent bioactivity can further enhance the efficacy. For example, IC50 value of COF-PLU nanoparticles in hMSc is ~10 times higher than that of graphene.³³ Similar to other nanocarriers mentioned above, COF-PLU nanoparticles are water dispersible which may find its clinical use as injectable solution or for localized drug delivery when loaded in hydrogel.

Biomedical exploration of COF-based material has just begun. The major obstacle of biomedical exploration of COF is their poor aqueous stability and difficulty of processability in nano-size. Here, we achieve hydrolytically stable 2D nanoparticles of COF-PLU by post-synthetically functionalized colloidal COF-5 with Pluronic F127 (PLU). PLU is a biocompatible amphiphilic tri block-co-polymer widely used in pharmaceuticals. In water, it forms micelles of hydrodynamic diameter of ~8–12 nm above its critical micelle concentration (CMC) of 0.26 wt%–0.8 wt % which have long been used in drug delivery.^{53, 54} Other than that, PLU has also been widely used to coat or non-covalently functionalize nanoparticles including PLGA, silica, graphene etc. via a strong hydrophobic interaction of the PPO segment and the nanoparticle.^{53, 55, 56} A study by Yang et al. has shown that the hydrophobic interaction-based coating of the nanoparticles with PF127 or other Pluronics with the PPO MW ≥ 3 kDa are as strong as covalent PEG coating.⁵⁵ Also, Li et al. has demonstrated that PF127 at lower concentration (0.05 mg/mL form stable mono-layer in the air-water interface via anchoring of the hydrophobic PPO groups which disappears upon addition of graphene oxide (GO) nanoparticles. The strong interactions of PPO group with the hydrophobic parts of GO surface, resulting in loss of the anchors at the air–water interface and drags the complex PLU/GO into the bulk subphase.⁵⁶ In our study we have used the mixed solvent containing acetonitrile:1,4-dioxane:mesitylene (20:4:1, v/v) for post-synthetically integrating PLU and colloidal COF-5 which immediately (20s) forms an aggregate and then produce nanoparticles of ~200 nm diameter upon resuspension of the aggregate in water which avoids possibility of mono-layer formation unless it dissociates in water. Also, the effective concentration of F127 in aqueous suspension of COF-PLU nanoparticle is 0.125 wt% which is much less than CMC value of PLU. In addition, in DLS and TEM study we did not observe significant population of 8–12 nm hydrodynamic diameter corresponding to micelles of PLU indicating stability of the COF-PLU nanoparticle over time. However, we observed increase in hydrodynamic diameter of the COF-PLU nanoparticle with a higher concentration of PLU (COF:PLU; 1:25 w/w) indicating aggregation at its higher concentration.

Other biocompatible polymer like PLGA, which is approved by FDA for clinical use, is also explored for delivering dexamethasone. Although PLGA nanoparticles exhibit good drug loading behavior, the acidic degradation by-products induce severe local inflammation which negatively regulate bone regeneration.⁵⁷ Moreover, with the breakthrough advances in material science, there is search for bioactive materials that can actively intervene in tissue regeneration process in absence of drug. Some 2D nanomaterials including laponite, graphene can induce osteogenic differentiation in stem cells in absence of any osteoinductive agent. Such 2D nanomaterials induce biophysical stimuli to influence cytoskeletal dynamics of adjacent stem cells which promotes the osteogenic events.^{39, 58} Interestingly, 2D nanoparticle of COF-PLU also exhibit osteogenic activity which is not mediate via biochemical process as its ingredients components fail to induce any such ALP activity in hMSC (Figure 5). Furthermore, when hMSC are treated with 2D nanoparticles of COF-PLU in presence of inhibitors of different plausible osteogenic pathways, significant reduction of ALP activity is observed when hMSCs are treated with inhibitor of JNK pathway indicating it might be plausible pathway of osteogenic activity of the nanoparticles (Figure 5B–C). Clearly, further in-depth studies need to be carried out in future toward obtaining further mechanistic insights. Importantly, studies aimed at developing COF-based biomaterials for bone tissue engineering has just begun and this study hold distinct promise towards developing COF-based bioactive nanomaterial.

CONCLUSIONS

In summary, we have developed a new class of hydrolytically stable 2D COF-based nanomaterial for sustain delivery of therapeutics to direct stem cell fate. Specifically, a boronate–based covalent organic framework (COF-5) is stabilized using amphiphilic polymer (Pluronic F-127). Interestingly, 2D COF is able to induce osteogenic differentiation of human mesenchymal stem cells (hMSCs) in absence of osteoinductive agent, which is unique. In addition, we are able to load osteoinducing agent – dexamethasone within the porous structure of COF to further enhance the osteoinductive ability. Our results demonstrate a new fabrication strategy for developing hydrolytically-stable 2D COF-nanocomposite and for the first time, advocates importance of exploring such materials in regenerative medicine.

EXPERIMENTAL

Materials:

2,3,6,7,10,11-hexahydroxytriphenylene (catalog no TCH0907-005G) and Dexamethasone (catalog no D1961) are purchased from TCI America. Benzene 1,4-diboronic acid (catalog no. 417130-5G) and FITC-BSA are procured from Sigma Aldrich. Acetonitrile (catalog no 42311) and 1,4-Dioxane (catalog no 39118) are procured from Alfa Aesar. PluronicF-127 is purchased from BioVision (Catalog #: 2730). Mesitylene (catalog no 134435) and 4 Å Molecular sieves (MX1583G-1) are obtained from Beantown Chemical and EMD Millipore, respectively. Human mesenchymal stem cells (hMSCs) are purchased from ATCC and maintained in alpha MEM medium supplemented with 16.5% fetal bovine serum (FBS) and 1× penicillin–streptomycin solution in a humidified 5% CO2 incubator at 37 °C.

Synthesis and characterization of nanoparticles:

COF-5, a boronate based covalent organic framework, is synthesized in its stable colloidal form following prior report^{5, 23}. Briefly, solvothermal condensation of 2,3,6,7,10,11-hexahydroxytriphenylene (HHTP) and benzene 1,4-diboronic acid (BDBA) (2 mM HHTP and 3 mM BDBA) in a mixed solvent containing dry acetonitrile: dioxane: mesitylene as 20:4:1 (v/v) at 90 °C under homogeneous conditions for 18 h produces translucent colloidal solution of COF-5. Addition of 200 μL of PluronicF-127 (PLU) dissolved in dry acetonitrile (25 mg/mL) to 1 mL of colloidal COF in room temperature followed by vortex for 20 sec leads to immediate aggregation of colloidal COF to form COF-PLU composite. Centrifugation (5000 rpm, 30 sec) followed by removal of mixed organic solvent and resuspension of COF-PLU composite precipitate in 4mL of water results in formation of 2D nanoparticles of COF-PLU. We examined with three mass ratio of COF:PLU (w/w) as 1:1, 1:5, and 1:25 and we found that 1:5 ratio shows smallest nanoparticle and exhibit stable behavior in DLS with time. It is presumably because that 1:1 may not be enough to ensure the hydrolytic stability while 1:25 ratio being heavy in PLU undergoes aggregation. Hydrolytic stability of colloidal COF-5 is monitored by monitoring turbidity of the colloidal samples added with 5.5 M water (v/v) using a uv-vis spectrophotometer.⁹ Attenuated total reflectance Fourier transform infrared spectroscopy (ATR–FTIR, Bruker Vector 22 FTIR spectrophotometer) and thermogravimetric analysis (TGA) are performed to confirm the chemical composition of COF-PLU nanoparticles. The hydrodynamic diameters and surface potentials of COF-PLU nanoparticles and their drug loaded formulations are measured at 25 °C using Zetasizer Nano ZS (Malvern Instrument, U.K.) equipped with a He–Ne laser. TEM images of those nanoparticles are recorded using JEOL 2010 TEM instrument. Thickness of COF-PLU nanoparticles is measured via AFM tapping mode using Bruker Nanoscope 9.1 instrument and analyzed with Nanoscope Analysis software.

Drug loading and release study:

Drug loading and release studies are performed following a prior reports.⁵⁹ Briefly, varying amount of Dex (10%–100% of Dex with respect to 1 mg of COF) is added to 1 mL of colloidal COF (1 mg/mL) from a stock solution of Dex (22 mg/mL) in 60% aqueous 1,4-Dioxane and incubated for 30 min at 37 °C. PLU taken in acetonitrile is added to it (COF:PLU 1:5, w/w) followed by vortex and centrifugation to obtained Dex loaded COF-PLU composite as precipitate. Removal of solvent supernatant followed by resuspension of Dex loaded COF-PLU composite in 4 ml of DI water and subsequent dialysis against DI water produces Dex loaded COF-PLU nanoparticle (Dex-COF-PLU). The Dex entrapment efficiency of COF-PLU nanoparticles is calculated by measuring absorbance of unentrapped Dex at 241 nm using UV-Vis spectrophotometer (Infinite 2000 Pro). Encapsulation efficiency (%) = (Drug added initially-Unentrapped drug)*100%/Drug added initially. Dex release study is performed at 37°C at two different pH conditions (pH = 7.4 & pH = 5.5) in PBS by dialysis bag method⁶⁰ and Dex release rate at predefined time points are monitored by measuring UV absorbance at 241 nm.

Cell viability and cell cycle analysis:

The cell viability of hMSCs treated with nanoparticles of COF-PLU is determined by conventional MTT [3-(4,5-dimethylthiazol-2-yl)-2,5-diphenyltetrazolium bromide] assay. Briefly, hMSCs are seeded (1000 cells per well) in 96-well plates 12–18 h before treatment and treated with increasing concentration of COF-PLU nanoparticles (across the range of 0.001 to 5 mg/mL) for 24 h. Cells are then incubated with MTT solution (10 μL from 5 mg/mL stock in PBS) for 4 h at 37 °C and cell viability is determined by reading absorbance of formazan at 550 nm using a microplate reader. Results are expressed as percent viability = [A550 (treated cells) − background/A550 (untreated cells) − background] × 100. For cell cycle analysis, hMSCs are seeded in flask (25 cm2), serum starved for 16 h to synchronize cell populations and then treated with nanoparticle of COF-PLU in different concentrations for 72 h. Cells are trypsinized and fixed in ice cold 70% ethanol for 8 h at 4 °C. Cells are then washed with PBS, incubated in a PI staining solution at 37 °C for 20 min and analyzed under flow cytometer.

Cellular uptake study:

FITC-tagged COF-PLU nanoparticles are used to examine cellular internalization in human mesenchymal stem cells. FITC-BSA (8 μL from 20 mg/mL stock) is added in 500 μL of colloidal COF, incubated for 10 min at 37 °C and added with PLU (100 μL from a 25 mg/mL stock in acetonitrile) followed by vortex for 30 sec, centrifugation and resuspension of the precipitate to obtain nanoparticles of FITC-tagged COF-PLU. hMSCs are seeded in 96-well plate (seeding density of ~1000/well) and incubated with FITC-tagged COF-PLU (6 μL) for 4 h. Cells are fixed with 2% formaldehyde and stained for Actin (Rhodamine Phalloidin) and Dapi. Images are recorded under Leica SP8 confocal microscope.

Cellular internalization:

hMSCs are seeded in 6 well plate (25000/well). Cells are divided into five groups. First, second and third groups are separately treated with inhibitors of clathrin-mediated endocytosis (35 μM of Choloropromazine.HCl), calveolar-mediated endocytosis (10 μM of Nystatin), and macropinocytosis (400 nM of Wortmannin), respectively for 1 h followed by incubation with ~50 μL of FITC tagged COF-PLU for additional 4 h. Fourth group is treated with only ~50 μL of FITC tagged COF-PLU for 4 h and 5th group remains untreated. Subsequently, the cells are washed with PBS, trypsinized, suspended in cell culture medium and analyzed under flow cytometer.

In vitro osteogenic differentiation:

Osteogenic differentiation potential of the Dex-loaded COF-PLU nanoparticles are evaluated using 2D culture of hMSCs. All the cells used are within passage 5. hMSCs are cultured in growth media composed of α-modified minimal essential media added with 16.5% FBS and 1% penicillin/streptomycin whereas all the differentiation studies are performed in osteoconductive media which is composed of growth media supplemented with 10 mM β-glycerophosphate (Sigma-Aldrich) and 50 μM ascorbic acid (BDH Chemicals). hMSCs are seeded in a 24- well plate at a density of 4000 cells/cm2 in growth media, replaced with osteogenic media next day and divided into four treatment groups; i) First group is incubated with Dex-loaded COF-PLU nanoparticle (containing 100 nM of Dex and 6 μg/mL of COF-PLU), ii) second group is incubated with equivalent amount (6 μg/mL) of empty COF-PLU nanoparticle, iii) third group is exogenously added with 100 nM of Dexamethasone (positive control), iv) fourth group remained untreated (negative control). hMSCs are treated for 7 days and 14 days. Media is replenished twice a week. Conventional osteogenic assays, specifically, alkaline phosphatase (ALP) staining and kinetic assay, alizarin red staining (ARS) are performed to analyze osteogenic differentiation. Both of the ALP assays (staining as well as kinetic assays) are performed at day 7 and day 14 whereas alizarin red staining (ARS) is performed at day 14. For ALP staining, hMSCs are first fixed with 2.5% glutaraldehyde for 15–20 min and then stained for ALP using NBT/BCIP 1-steps solution (nitroblue tetrazolium/5-bromo-4-chloro-3-indolyl phosphate, Thermo Fisher) with NBT/BCIP 1-steps solution (nitroblue tetrazolium/5-bromo-4-chloro-3-indolyl phosphate, Thermo Fisher) for 30–60 min at room temperature. For ALP kinetic assay, cultures are incubated with ALP yellow (Sensolyte pNPP ALP assay kit, AnaSpec). ALP activity as a function of pNPP metabolism (ΔOD405) is measured using an automated plate reader and activity is normalized to DNA content (PicoGreen, Thermo Fisher). For ARS, hMSCs are also similarly fixed with 2.5% glutaraldehyde and treated with staining solution for 30–60 min on shaker. Bound ARS which is proportional to the matrix mineralization is extracted by cetylpyridinium chloride method and spectrophotometrically quantified by measuring absorbance at 405 nm. Both ALP and ARS staining are visualized with a stereomicroscope (Zeiss).

Inhibition assays:

Inhibitors are added to cell culture media at the following concentrations that are determined empirically to allow complete spreading with no visually apparent changes to morphology. Cells are treated (every 3 days) with COF-PLU in presence or absence of the inhibitors; Wnt inhibitor Cardamonin (8 μM), TGF-β inhibitor SB-431542 (8 μM), ERK inhibitor Selumetinib (9 nM), JNK inhibitor SP600125 (5 μM), p-38 inhibitor SB202190 (5 μM). ALP assays (kinetic assays) are performed at day 4, 8, and 12.

Statistical analysis.

All the results are expressed as the mean ± deviation (n = 3) as well as statistical analysis are performed via a one-way analysis of variance (ANOVA) with post-hoc Turkey analysis using GraphPad Prism software. The statistical significance is presented as *P-value < 0.05, **P-value < 0.01, ***P-value < 0.001, and ****P-value < 0.0001

Data and materials availability:

All data needed to evaluate the conclusions in the paper are present in the paper and/or the Supplementary Materials. Additional data related to this paper may be requested form the authors.