Local and targeted drug delivery for bone regeneration

Abstract

While experimental bone regeneration approaches commonly employ cells, technological hurdles prevent translation of these therapies. Alternatively, emulating the spatiotemporal cascade of endogenous factors through controlled drug delivery may provide superior bone regenerative approaches. Surgically placed drug depots have clinical indications. Additionally, noninvasive systemic delivery can be used as needed for poorly healing bone injuries. However, a major hurdle for systemic delivery is poor bone biodistribution of drugs. Thus, peptides, aptamers, and phosphate-rich compounds with specificity toward proteins, cells, and molecules within the regenerative bone microenvironment may enable the design of targeted carriers with bone biodistribution greater than that achieved by drug alone. These carriers, combined with osteoregenerative drugs and/or stimuli-sensitive linkers, may enhance bone regeneration while minimizing off-target tissue effects.

Graphical Abstract

Introduction and motivation

Trauma, osteoporosis, and genetic disease cause approximately 6 million fractures in the United States annually [1]. Between five and ten percent of fractures result in delayed union or non-union, with age, smoking, and diabetes as common comorbidities of poor healing [1,2]. Successful bone regeneration requires a complex and orchestrated cascade of cells and growth factors [3–6]. To emulate this complex cascade, regenerative strategies often employ cell transplantation and/or growth factor delivery. While each of these approaches has yielded different levels of success clinically, there are still significant technological challenges.

To achieve translation of cell-based therapies, there are many hurdles to overcome. For bone regeneration, mesenchymal stem cells (MSCs) are commonly used, as they can differentiate to cells required for bone healing (osteoblasts and chondrocytes). However, other cells (e.g., inflammatory cells, endothelial cells) that play important roles in bone regeneration may also be included. A challenge to the use of any cell type is assurance of cellular uniformity. With MSCs as an example, the International Society for Cellular Therapy (ISCT) issued three criteria for identifying MSCs: plastic adherence, specific cell surface marker expression, and tri-lineage potential [7]. Additional gains in uniformity have been achieved through GMP-compliant isolation procedures [8], culture conditions [9,10], and cryopreservation techniques [11,12]. There are now Food and Drug Administration GMP-approved facilities (the NIH BMSC Transplantation Center and the Upstate New York Stem Cell cGMP Facility) that generate MSCs for clinical investigations [13^•]. However, cell-based therapies require significant lead-time to achieve cell quantities necessary for many regenerative approaches [14]. Further, these approaches still may fall short due to inconsistent and or unpredictable outcomes resulting from donor-to-donor and batch-to-batch variability of cells. For example, it has been shown that patients with many co-morbidities, such as age, osteoporosis, genetic defects, infection, obesity, diabetes, and smoking, exhibit reduced MSC potency and number resulting in poor bone regeneration [15]. While it is possible to transplant regenerative cell types, including MSCs or MSCs pre-conditioned with small molecule drugs [16] or augmented with genetic manipulations [17] to encourage osteogenesis or microenvironmental modulation [18], to date no MSC-based regenerative strategy is approved for clinical use [19^••].

Nevertheless, nearly 20% of the approximately 100 MSC-based clinical trials registered in the United States National Institutes of Health database are for bone and cartilage regeneration [20]. However, there is still a nascent understanding of long-term safety and efficacy of MSC-based therapies as most trials are in Phase I or Phase I/II. Interestingly, for the ~33% of trials completed, there is evidence to suggest MSCs have therapeutic benefits that are not aligned with typical tissue engineering outcomes (e.g., tissue-specific differentiation and tissue production), such as enhancing vascularization in patients with osteonecrosis of the hip [21]. Additional therapeutic effects include attenuating inflammation and stimulating proliferation, suggesting these effects are due to MSC production of cytokines and growth factors that exhibit paracrine and/or autocrine effects [22].

Ultimately, cell phenotype and function are coordinated by a myriad of spatiotemporally regulated growth factors to realize bone regeneration. These factors include transforming growth factors (TGFs), bone morphogenetic proteins (BMPs), stromal cell-derived factor 1 (SDF1), and osteoprotegerin (OPG) (Figure 1). Based on these factors, a variety of drugs, including BMPs, other growth factors, hormones, and monoclonal antibodies, have been explored for bone regenerative effects [20,23^•]. Common strategies focus on BMP therapy, as canonical BMP signaling is integral to bone formation [13^•,24^•]. In fact, drug delivery approaches that increase the availability of factors such as BMP2 have been proven safe and efficacious for bone regeneration (e.g., INFUSE™ [6]); however, protein (e.g., antibodies, growth factors) half-lives are only on the order of an hour [25], necessitating local delivery of supra-physiological doses to achieve desired pharmacodynamics.

Figure 1.

A spatiotemporal cascade of multiple endogenous factors controls normal bone regeneration during fracture repair in four stages. PDGF = platelet derived growth factor; VEGF = vascular endothelial growth factor; FGF = fibroblast growth factor; TNF = tumor necrosis factor; SDF = stromal cell-derived factor; IGF = insulin-like growth factor; BMP = bone morphogenetic protein; OPG = osteoprotegerin; IL = interleukin; TGF = transforming growth factor; Ang = angiopoietin; M-CSF = macrophage colony stimulating factor; RANK = receptor activator of nuclear factor kB; RANKL = RANK-ligand.

As an alternative to BMP, small molecule drugs such as statins that induce BMP signaling [26] and osteoregenerative agents that act on BMP-convergent pathways may also enhance bone remodeling [27^•,28]. These convergent pathways, including parathyroid hormone (PTH) and Wnt, stimulate osteogenesis and preserve osteoblasts when activated (Figure 2). Activated PTH receptor initiates Wnt signaling by complexing with low density lipoprotein receptor-related protein 5 and 6 (LRP5/6) [28], and Wnt signaling regulates targets common to BMP signaling [29]. Teriparatide (PTH 1-34), which is approved to treat osteoporosis, is used off-label in normal bone fractures, delayed bone fractures [30], and non-unions [31], and patients demonstrate accelerated healing. Lithium, a Wnt signaling agonist that negatively regulates glycogen synthase kinase 3 beta (GSK3b) [32^•], and a monoclonal antibody that inhibits the Wnt-inhibitory protein sclerostin [33] accelerate fracture healing since Wnt signaling is critical for fracture repair [34]. Wnt, PTH, and BMP play key roles in specific stages of fracture repair [35,36], but they can also be inhibitory if activated without proper spatiotemporal control [37].

Figure 2.

Mesenchymal stem cell chondrogenesis and osteogenesis are regulated by specific gene expression that is influenced by BMP, Wnt, and PTH signaling. Symbols indicating the various cell types are consistent with Figure 1.

While many growth factor-based bone regenerative therapies have been explored, many require invasive procedures and/or require supra-physiological factor delivery to achieve efficacy, resulting in therapies that are cost prohibitive ($5000 to $6000 per package [38]) and include risks for harmful side effects [39]. As an alternative strategy to augment cell function and enhance bone remodeling, controlled drug delivery approaches amenable to systemic delivery avoid invasive procedures and minimize off-target cell behaviors.

Controlled drug delivery to bone

To augment bone regeneration, there is a need to isolate delivery to the bone [4,40^•,41^•,42^•,43]. This is due, in part, to mitigate off-target effects commonly associated with signaling pathway crosstalk [44] as a result of poorly controlled drug delivery. Drug depots and targeted systemically delivered carriers that present drugs to cells on demand may enhance the efficacy and reduce the potential toxicity of regenerative drugs.

Drug depots

Basic diffusion dependent drug depots consist of a drug loaded within a carrier: for example, BMP2 loaded into the collagen-based scaffold of INFUSE™ [6]. Diffusion dependent depots often demonstrate an initial uncontrolled burst release of drug followed by first order release dictated by the size of drug relative to the pore or mesh size of the carrier. Thus, cells are exposed to drug as soon as the carrier is implanted, which may be undesirable because MSC differentiation is sensitive to the timing of growth factor presentation [45^•].

Greater control over drug release can be achieved through depots that respond to environmental stimuli, including enzymatic activity and changes in pH. A variety of matrix metalloproteinases (MMPs) are expressed during bone formation and remodeling [46,47]. Drugs modified with peptide sequences sensitive to MMPs can be linked within hydrogel depots and released upon MMP activity [48]. Alternatively, drugs can be encapsulated in hydrogels that degrade via MMP cleavage. Designed from α(1)-chain type I collagen, the peptide sequence GPQGIWGQ is cleaved by multiple MMPs [49], and a derivative of this sequence was the crosslinker for poly(ethylene glycol) hydrogels containing BMP2 [50]. Upon implantation, degradation, and increase in hydrogel mesh size, the hydrogels released BMP2. This resulted in regeneration of critical-sized calvarial defects in four weeks, with the amount of bone regeneration directly correlated to crosslinker MMP degradation rate. Radiopacities of the defects were greater than 90% for the faster degrading hydrogels, greater than 75% for the slower degrading hydrogels, and greater than 40% for the non-degrading hydrogels. Similarly, the peptide sequence VPMSMRGG, which is cleaved by MMP1 and MMP2, was incorporated as the crosslinker for maleimide-functionalized hyaluronic hydrogels containing BMP2 [51^•]. When implanted in critical-sized calvarial defects, the fastest degrading hydrogels resulted in significantly greater bone regeneration in six weeks. The ratio of bone volume to total volume in the defect was approximately 0.25 for the fastest degrading hydrogels, whereas the ratio for the slowest degrading hydrogels was only 0.15, which was not significantly different from the ratios for untreated defects or unloaded hydrogels.

Therapeutically relevant concentrations of drug can be released from pH-responsive carriers. For example, osteomyelitis, a condition characterized by an acidic environment, was treated with antibiotic released from pH-responsive poly(d,l-lactic-co-glycolic acid) (PLGA) microspheres impregnated within calcium phosphate scaffolds [52]. These microspheres provide greater control over release versus scaffolds simply impregnated with antibacterial proteins [53]. Additionally, dental biofilms, which also present an acidic environment, were treated with pH-responsive diblock copolymer nanoparticles containing anti-biofilm agents [54,55]. At physiological pH (7.2), 50% of loaded drug was released within 15 h, while at acidic pH (4.5), 50% of loaded drug was released within 7 h. This approach may be useful in combination with or used in lieu of antibiotics to prevent biofilm colonization during bone regeneration [56].

Systemically delivered, targeted carriers

While drug depots provide site-specific drug delivery, similar to cell transplantation, they often require invasive procedures for placement. The primary exception is an injectable formulation that assembles in situ. Although critical sized defects already require surgery to heal, smaller defects or closed fractures may not, and a surgery conducted to place a depot is counterproductive to healing. Alternatively, systemic injection of drug or drug-loaded polymer functionalized with a bone-targeting moiety provides a noninvasive technique for site-specific therapy (Figure 3). It is not necessary to administer this therapy upon insult to bone and instead can be administered at any time, such as in the case of a fracture that is not diagnosed immediately after injury or in the case of regenerating bone in individuals with osteoporosis and other systemic bone diseases.

Figure 3.

Targeted, systemic drug delivery uses targeting moieties (e.g. peptides and aptamers) to deliver drugs to bone. Carriers travel through the bloodstream and exit upon targeting. Drugs are cleaved via MMP and enzymatic activity or due to a change in pH. Symbols indicating the various cell types are consistent with Figure 1.

Reviewed in-depth by Low and Kopeček [57] and Luhmann et al. [58], traditional bone-targeting moieties that bind strongly (dissociation constant, K_D, in the micromolar range) to the mineral phase of bone include bisphosphonates, acidic oligopeptides, and tetracycline. More recently, derivatives of anthraquinone were found to bind hydroxyapatite mineral and may be used to deliver nonsteroidal anti-inflammatory drugs to arthritic bone [59]. It is also possible to target the organic phase of bone by using collagen binding domain (CBD) peptides derived from Clostridium histolyticum (ColH) collagenase, where CBD can be a synthetic extension of a peptide drug [60^••].

Despite gains in bone distribution relative to untargeted drugs, traditional targeted carriers exhibit poor bone accumulation in terms of percent injected dose measures, perhaps due to a lack of specific binding. Alternative high-affinity (K_D in the nanomolar to picomolar range) bone-targeting peptides and aptamers (Figure 3) may be found through libraries of peptide and nucleic acid sequences designed to specifically bind target proteins, molecules, or cells in bone (Table 1). Studies have used phage displays to identify peptides that bind to hydroxyapatite [61,62], collagen [63], tartrate-resistant acid phosphatase (TRAP) [64], bone marrow stromal cells [65], osteoblasts [66–68], and bone marrow endothelial cells [69]. While many of these studies were focused on cell interactions and not targeting drugs to bone, peptides may be incorporated into polymer carriers or synthesized as fusion peptides with therapeutic sequences. Another study used cell-based systematic evolution of ligands by exponential enrichment (cell-SELEX) technology to design an osteoblast-specific aptamer-decorated liposome to deliver siRNA to bone and promote osteoblast function [70^••], and various aptamers have been designed to selectively bind to osteoblasts [71], bone marrow MSCs [72], osteoprogenitor cells [73], and C-telopep-tide of human type I bone collagen (CTX) [74]. While many of these aptamers are used for characterizing cells, they can conceivably be applied to drug delivery approaches.

Table 1.

Specific peptide and aptamer targets and synthetic sequences. Peptides are written N-to-C terminus using standard amino acid abbreviations (γE = γ-carboxylated glutamic acid), and aptamers are written 5′-to-3′ using standard oligonucleotide abbreviations.

Target	Peptide		Aptamer
Collagen	PNNSKETASGPIVPGIPVSGTIENTSDQDYF YFDVITPGEVKIDINKLGYGGATWVVYDEN NNAVSYATDDGQNLSGKFKADKPGRYYI HLYMFNGSYMPYRINIEGSVGR	[60••]	GGTGGTGTTGGCTCC	[74]
	SWWGFWNGSAAPVWSR	[63]
Hydroxyapatite	VTKHLNQISQSY	[61]
	γEPRRγEVCγEL	[62]
TRAP	TPLSYLKGLVTV	[64]
Bone marrow MSCs	DPIYALSWSGMA	[65]	GAATTCAGTCGGACAGCG-N40- GATGGACGATATCGT CTCCC	[72]
Osteoprogenitors			GGGAGACAAGAATAAACGCTCA ACAAATGGGTGGGTGTGGTGGG TGTGAAGGTGCGAGTTGATTCGA CAGGAGGCTCACAACAGGC	[73]
Osteoblasts	MGWSWWPETWPM	[66]	AGTCTGTTGGACCGAATCCCGT GGACGCACCCTTTGGACG	[70••]
	YRAPWPP	[67]
	GKIHRHRGQAVE	[68]	GAATTCAGTCGGACAGCGCACA CGGAACCTCGGAACACAGCTAG CGGGGCTCACTGGATGGACGAA TATCGTCTCCC	[71]
	ESHCLLGISCVL	[68]
Bone marrow endothelial cells	MGGTVESCLAKSHTENSFTNVWKDDK TLDRYANYEGCLWNATGVVVCTGDET QCYGTWVPIGLAIPEN	[69]

Complementing the strategy of primary targeting, a pro-drug approach can ensure drugs are released only upon successful delivery to bone. Linkers placed between a drug and a carrier can be developed to be cleaved only by enzymes or stimuli localized at the site of repair. Just as MMP cleavage is used to release drugs from hydrogel depots, so too are drugs released from carriers following MMP cleavage of labile linkers (Figure 3). Following fracture and during remodeling, osteoclasts produce enzymes, such as cathepsin K, and create an acidic environment [4]. Thus, linkers sensitive to enzymes or low pH could be introduced to enable selective release at resorption surfaces. For example, the peptide sequences GGGMGPSGPWGGK [75], HPGGPQ [76,77], and GGP-Nle [78], which are substrates for cathepsin K, or polyketals [79,80], acetals [80], and hydrazones [81,82^•,83], which are acid-labile bonds, could be incorporated. By using linkers dependent on osteoclast bone catabolism, rate of release may be matched to the required regenerative response because osteoclasts appear early in the repair process, increase in number during soft and hard callus formation, and decrease during resolution [84–86].

Future perspectives

Bone regeneration may benefit from a customizable, targeted approach to treatment. This customization may be at the level of intra-operative decisions, wherein a specific composition of drug and linker may be applied through a patient-specific drug delivery patch, strip, or sleeve [87]. However, it is important to maintain simplicity to facilitate clinical adoption of new technologies. Delivering stimuli-responsive depots or polymers enables customization without a priori knowledge of a patient's condition or on-the-spot modifications since stimuli (e.g. enzymes or low pH) are presented in situ over time. Each stage of bone regeneration is characterized by a distinct microenvironment and a distinct signaling molecule profile. Therefore, if the natural healing cascade is emulated via multi-drug release of osteoregeneratives by enzymatic or other chemical activities that dominate during times at which drugs are most beneficial, a new frontier for bolstering bone regeneration may soon be realized.

Highlights.

• Bone regeneration requires spatiotemporally controlled signals and cells.

• No cell-based therapy is approved to augment bone regeneration.

• Controlled drug delivery minimizes off-target effects of exogenous signals.

• Peptide and aptamer targeting may increase bone biodistribution specificity.

• Environmental stimuli-mediated drug release can match supply to demand.