Cartilage‐targeting ultrasmall lipid‐polymer hybrid nanoparticles for the prevention of cartilage degradation

Abstract Current drug delivery approaches for the treatment of cartilage disorders such as osteoarthritis (OA) remain inadequate to achieve sufficient drug penetration and retention in the dense cartilage matrix. Herein, we synthesize sub‐30 nm lipid‐polymer hybrid nanoparticles functionalized with collagen‐targeting peptides for targeted drug delivery to the cartilage. The nanoparticles consist of a polymeric core for drug encapsulation and a lipid shell modified with a collagen‐binding peptide. By combining these design features, the nanoparticles can penetrate deep and accumulate preferentially in the cartilage. Using MK‐8722, an activator of 5′‐adenosine monophosphate‐activated protein kinase (AMPK), as a model drug, the nanoparticles can encapsulate the drug molecules in high capacity and release them in a sustained and controllable manner. When injected into the knee joints of the mice with collagenase‐induced OA, the drug‐loaded nanoparticles can effectively reduce cartilage damage and alleviate the disease severity. Overall, the ultrasmall targeted nanoparticles represent a promising delivery platform to overcome barriers of dense tissues for the treatment of various indications, including cartilage disorders.


INTRODUCTION
Osteoarthritis (OA), characterized by progressive degeneration of articular cartilage, chronic pain, and loss of mobility, is the most common joint disorder and the leading cause of disability worldwide. 1,2 Despite it is an enormous medical need, a disease-modifying treatment of OA remains unavailable. Managing pain and improving joint function are the current standard of care as the disease progresses. 3 Although the exact causes of OA remain unknown, a close association of the disease with the aberrant behavior and abnormal phenotype of the resident chondrocytes has been revealed. 4 As a result, emerging pharmacotherapies are increasingly focused on modulating chondrocyte metabolism, with a hope to halt or reverse OA progression. 5,6 For this purpose, intra-articular injection of therapeutic agents is a preferred route of administration. 7 It offers direct access to the joint space, thereby improving on overall drug bioavailability to the resident chondrocytes while reducing systemic exposure. However, intra-articular injection of drug molecules is challenged by the fact that subsynovial capillaries and lymphatics can rapidly remove the injected drug from the joint space, and thus the therapeutic benefit is short-lived.
The pressing needs of improving intra-articular drug delivery for OA has led to a plethora of nanoparticle designs aimed at enhancing drug retention inside the joint. 8,9 Compared to small molecules, nanoparticles are more likely to be trapped in the joint space and therefore act as reservoirs for prolonged drug release to the synovial tissue. Despite the advantages, this approach does not assure efficient penetration of drugs into the dense extracellular matrix (ECM) of the cartilage to reach chondrocytes in the deep regions of the cartilage.
To address this challenge, adding cartilage-targeting ability to the nanoparticles has been an attractive strategy to facilitate nanoparticle penetration into the cartilage for high intratissue drug concentration. 8 In this regard, positively charged nanoparticles have become popular as they spontaneously bind with negatively charged glycosaminoglycans (GAGs) chains of the ECM through electrostatic interactions. 10,11 Meanwhile, targeting moieties such as anticollagen antibodies and collagen-binding peptides have been conjugated onto nanoparticles for active targeting to the ECM. 12,13 Alternatively, nanoparticles have also been made to bind with the resident chondrocytes for targeting.
In this regard, synthetic ligands such as chondrocyte-affinity peptides and natural ligands from the plasma membranes of the immune cells have been leveraged. 14,15 As researchers continually develop nanoparticles for cartilage targeting, recently, "ultrasmall" lipid-polymer hybrid nanoparticles (denoted "LP-NPs") gained attention for drug delivery. 16 In this development, charged groups were introduced to the polymer backbones to reduce polymer-water interfacial tension during nanoprecipitation, resulting in nanoparticles smaller than 30 nm. 17 Despite their "ultrasmall" sizes, they held a high capacity for encapsulating hydrophobic molecules without burst release. When applied to cancer drug delivery, their ultrasmall size allowed for deep penetration into the dense interstitial space of the tumor microenvironment, and the tumor-binding ligand conjugated onto the nanoparticle surface reduced their outflux. These distinct features allowed ultrasmall LP-NPs to deliver a high concentration of anticancer drugs to the tumor, leading to a better antitumor efficacy compared to nontargeted control nanoparticles.
The unique capabilities of ultrasmall LP-NPs inspire us to apply this nanodelivery platform for drug targeting to the cartilage. Studies have shown that nanoparticles below 40 nm are more likely to penetrate the cartilage than larger nanoparticles. 18,19 In the current study, we engineered ultrasmall LP-NPs with a diameter around 25 nm consisting of a core made from poly(lactic-co-glycolic acid) (PLGA) and a shell from polyethylene glycol (PEG)-modified lipid (Figure 1a). With such a small size, the LP-NPs are expected to penetrate deep into the ECM. We further conjugated the LP-NPs with a short collagenbinding peptide (WYRGRLC) that allowed for nanoparticle entrapment in the ECM (Figure 1b). 19 After confirming their enhanced penetration and retention in the cartilage of mouse femoral head, we loaded them with MK-8722, a potent activator of 5 0 -adenosine monophosphateactivated protein kinase (AMPK) known to regulate chondrocytes energy metabolism in the cartilage and alleviate OA severity. 20,21 In the study, the collagen-targeting LP-NPs (denoted "ctLP-NPs") exhibited sustained drug release over 48 h without drug bursting.
When injected into the knee joints of the mice with collagenaseinduced OA, they significantly alleviated the OA severity. Overall, we demonstrated that ultrasmall ctLP-NPs are a promising nanodelivery platform for cartilage-targeted drug delivery.

RESULTS AND DISCUSSION
In the study, the ctLP-NPs and nontargeted control nanoparticles (LP-NPs) were fabricated through a charge-based nanoprecipitation procedure. 16 Specifically, carboxylic acid-terminated PLGA in acetonitrile was added into a Tris-HCl buffer (pH = 8) containing PEG-lipid molecules, which self-assembled to form nanoparticles. Collagen-binding peptides were conjugated to PEG-lipid to introduce the targeting function. After purification, the dynamic light scattering (DLS) measurements of LP-NPs and ctLP-NPs showed a hydrodynamic diameter of 25 nm (Figure 2a). This value was approximately 5 nm larger than that of the bare PLGA cores, an increase consistent with the previous formulation of similar ultrasmall LP-NPs. 16 A similar size between the two groups suggests a negligible impact of the conjugated peptide on the nanoparticle size. Meanwhile, the zeta potential value of both F I G U R E 1 The design of collagen-targeting ultrasmall lipid-polymer hybrid nanoparticles (denoted "ctLP-NPs") for targeted drug delivery to the joints. (a) Schematic of ctLP-NPs containing a core made from poly(lactic-co-glycolic acid) (PLGA) and a shell made from polyethylene glycol (PEG)-conjugated lipid. For cartilage targeting, the nanoparticles are modified with collagen-binding peptides. Hydrophobic drug molecules are encapsulated inside the PLGA cores for delivery. The nanoparticles have a small size around 25 nm. (b) The ctLP-NPs are able to penetrate deep into the cartilage for effective drug targeting to the chondrocytes nanoparticle groups was less negative than that of the PLGA cores, likely due to the charge shielding by the PEG-lipid shell (Figure 2b). 22 The value of ctLP-NPs was higher, reflecting the presence of the positively charged targeting peptides on the nanoparticle surfaces. When Next, we examined the penetration and retention of ctLP-NPs in the articular cartilage of mice. We hypothesize that the ultrasmall size and collagen-targeting ability together will allow ctLP-NPs to penetrate deep and stay long in the cartilage. To test this hypothesis, we collected femoral heads from mice and incubated them with DiD-  (d) Quantification of protein content on PLGA cores, LP-NPs, and ctLP-NPs, respectively, using a BCA assay (UD, undetectable). (e) The hydrodynamic size of PLGA cores, LP-NPs, and ctLP-NPs in 1× PBS over a week. (f) Fluorescence intensity of DiD-labeled LP-NPs and ctLP-NPs bound onto a type II collagen-coated plate. As an additional control (labeled with "blocked"), the plate was pretreated with free peptide to block the collagen before adding ctLP-NPs. Data presented as mean ± SD (n = 3); n.s.: not significant; ***p < 0.001; statistical analysis by one-way ANOVA contrast, the signal from the femoral heads incubated with ctLP-NPs remained above 10% even when the depth reached 100 μm. We hypothesize that, depending on the nanoparticle size and binding affinity, ctLP-NP binding with the matrix may lead to a steeper intratissue concentration gradient than that of the nontargeted LP-NPs. This effect has been shown to boost both penetration and retention. 15,23 We further examined the nanoparticle retention in vivo by injecting them into the knee joints of mice. As shown in Figure 3d, the fluorescence intensity from both ctLP-NPs (the left knee, marked with "L") and LP-NPs (the right knee, marked with "R") decreased with time.
However, at all timepoints, the joint injected with ctLP-NPs ("L") remained brighter than the control joint injected with LP-NPs ("R").
Further quantification of the fluorescence confirmed the observed differences ( Figure 3e). Throughout the study, the signal of ctLP-NPs remained higher than that of the LP-NPs. At 48 h, 18% of the LP-NPs remained inside the knees. In contrast, 42% of the ctLP-NPs retained.
Overall, these results demonstrate significantly enhanced penetration and retention of ctLP-NPs compared to the nontargeted counterparts in the cartilage of mouse joints.
After having characterized the ctLP-NP formulation, we encapsulated a potent AMPK activator, namely MK-8722, into the nanoparticles and evaluated the drug loading and release properties. 24 As a hydrophobic molecule, MK-8722 can spontaneously incorporate into the PLGA cores during the nanoprecipitation process. To optimize MK-8722 encapsulation, we tested various drug inputs ranging from 0% to 20% of the total nanoparticle weight. Within this range, drug input had an insignificant impact on the nanoparticle size and their zeta potential ( Figure S1). As shown in Figure 4a, the loading capacity of MK-8722 increased as the initial drug input increased. The initial input of 20% resulted in a drug loading capacity of 5.8 wt%, the highest among all groups. Further increase of the initial input led to observable nanoparticle aggregation and precipitation. Therefore, we chose the drug input of 20% for the following studies. We then inves-  Despite the promise, cartilage degradation is a complex and multifactorial process. In our model, partial loss of Safranin-O staining was observed, indicating cartilage ECM degradation. To further assess the beneficial effect of ctLP-NP on preventing cartilage degradation, the efficacy need to be tested in additional mouse models of OA that emphasize different features of OA pathogenesis. 32,33 Mechanistic studies to differentiate roles played by chondrocytes, synoviocytes, and immune cells during the treatment will help to further optimize nanoparticle size, ligand density, and drug release profile. 34 Overall, the targeted ultrasmall nanoparticles hold great promise for effective drug delivery to dense tissues for preventing cartilage degradation. The continuous development of ctLP-NP may lead to a new treatment modality for OA and other cartilage disorders.

Preparation of ultrasmall lipid-polymer hybrid nanoparticles
The LP-NPs were prepared by following a previously published procedure. 16

Nanoparticle characterization
The size and surface zeta potential of the nanoparticles were deter-

Collagen binding study
Collagen (type II, from chicken sternal cartilage, Sigma-Aldrich) was dissolved in 0.25% acetic acid at a concentration of 0.5 mg mL −1 . Then 100 μL of the solution was added into each well of a 96-well assay plate.
The plate was incubated at 4 C overnight. Prior to the binding study, the plate was first blocked with 2% BSA for 1 h at room temperature and

Mouse femoral head degradation assay
Mouse femoral heads were collected from 10-week-old C57BL/6 mice. To study the degradation, 10 ng mL −1 of IL-1β was mixed with 2 mg mL −1 drug-loaded ctLP-NPs or LP-NPs in 100 μL DMEM medium and incubated with femoral heads in 96-well plate to stimulate cartilage degradation for 24 h. The medium was then changed, and the same amount of cytokine-nanoparticle mixture was added to the femoral heads. A total of three treatments, 24 h each, were conducted. Meanwhile, the culture media after the final treatment were collected for measuring the concentrations of IL-6 and TNF-α using an enzyme-linked immunosorbent assay (ELISA, BioLegend).

Mouse models of osteoarthritis
The CIOA mouse model was established by following a published method. 35 Briefly, the male C57BL/6 mice (10-12 weeks old) were randomly housed in filtertop cages with 12 h light-dark cycles and fed with a standard diet. CIOA mice were achieved by two intra-articular injections of 1 U collagenase type VII (Worthington Biochemical) in 10 μL sterile PBS per dose on day 0 and day 2. All the injections were performed in the right knee of the hind legs on the medial side of the knee joint by using a BD ultrafine insulin syringe 29G 1/2 0 (BD) under general anesthesia with ketamine (100 mg kg −1 ) and xylazine (10 mg kg −1 ).
In vivo efficacy study protocol and histological analysis of the knee joints To study therapeutic efficacy with CIOA mice, 20 μL of drugloaded ctLP-NPs or LP-NPs (5 mg mL −1 ) in PBS was injected every other day into the knee joint of mice on days 4-12 (after OA induction). As a negative control, PBS was injected. Following the treatment, mice were sacrificed on day 13 to collect hind legs. The knee joint sections were isolated and fixed in 10% formalin for 24 h.

Quantification of cytokines mRNA expression
The quantitative reverse transcription-polymerase chain reaction