Shear-Activated Nanotherapeutics for Drug Targeting to Obstructed Blood Vessels

See allHide authors and affiliations

Science  10 Aug 2012:
Vol. 337, Issue 6095, pp. 738-742
DOI: 10.1126/science.1217815

This article has a correction. Please see:

Bio-Inspired Drug Delivery

Noting that platelets naturally migrate to narrowed blood vessels characterized by high fluid shear stress, Korin et al. (p. 738, published online 5 July; see the Perspective by Lavik and Ustin) developed a nanoparticle-based therapeutic that uses a similar targeting mechanism to deliver a drug to vessels obstructed by blood clots. Aggregates of nanoparticles coated with the clot-dissolving drug tPA (tissue plasminogen activator) were designed to fall apart and release the drug only when encountering high fluid shear stress. In preclinical models, the bio-inspired therapeutic dissolved clots and restored normal blood flow at lower doses than free tPA, suggesting that this localized delivery system may help reduce the risk of side effects such as excessive bleeding.


Obstruction of critical blood vessels due to thrombosis or embolism is a leading cause of death worldwide. Here, we describe a biomimetic strategy that uses high shear stress caused by vascular narrowing as a targeting mechanism—in the same way platelets do—to deliver drugs to obstructed blood vessels. Microscale aggregates of nanoparticles were fabricated to break up into nanoscale components when exposed to abnormally high fluid shear stress. When coated with tissue plasminogen activator and administered intravenously in mice, these shear-activated nanotherapeutics induce rapid clot dissolution in a mesenteric injury model, restore normal flow dynamics, and increase survival in an otherwise fatal mouse pulmonary embolism model. This biophysical strategy for drug targeting, which lowers required doses and minimizes side effects while maximizing drug efficacy, offers a potential new approach for treatment of life-threatening diseases that result from acute vascular occlusion.

Disruption of normal blood flow to the heart, lung, and brain is the leading cause of death and long-term adult disability in the Western world (1). Current approaches to acute therapy for ischemic stroke, coronary infarction, and pulmonary embolism require infusion of thrombolytic drugs, which need to be administered systemically or through a catheter placed within the obstructed vessel, usually in an acute care hospital setting (24). To be effective, patients must receive therapy within a few hours after onset of symptoms, and the doses of clot-lysing drugs that can be administered are limited by the potential risk of bleeding because the active drug is free to distribute throughout the body. To overcome these limitations, we designed a thrombolytic delivery system that targets drugs selectively to sites of flow obstruction and concentrates the active drug in these regions.

Stenotic and thrombosed blood vessels exhibit physical characteristics that distinguish them from normal vasculature in that fluid shear stress can increase locally by one to two orders of magnitude, from below ~70 dyne/cm2 in normal vessels to >1000 dyne/cm2 in highly constricted arteries (58). Normal circulating platelets are locally activated by high shear stress in these regions and rapidly adhere to the adjacent surface lining of the narrowed vessels (911), which is a major contributing factor in the development of vulnerable atherosclerotic plaques. Inspired by this natural physical mechanism of platelet targeting, we developed a therapeutic strategy that uses local high shear stress as a generic mechanism to target treatment to regions of blood vessels that are constricted by clots, stenosis, or developmental abnormalities.

Our shear-activated nanotherapeutics (SA-NTs) are similar in size to natural platelets (1 to 5 μm in diameter); however, they are fabricated as aggregates of multiple smaller nanoparticles (NPs). The microscale aggregates remain intact when flowing in blood under physiological flow conditions but break up into individual nanoscale components when exposed to high local shear stress. Because of their smaller size as compared with that of the microscale aggregates, shear-dispersed NPs experience lower drag forces and, hence, adhere more efficiently to the surface of the adjacent blood vessel wall than do the larger microaggregates (fig. S1). The efficiency of this local adhesion can be further enhanced by coating the NPs with molecules that bind to endothelial cells or relevant targets, such as fibrin clots. In this manner, high concentrations of therapeutic agents can be concentrated locally at sites of vascular occlusion or embolism by immobilizing relevant drugs or enzymes on the NPs.

The SA-NTs were produced by spray-drying concentrated solutions of biocompatible, biodegradable, poly(lactic-co-glycolic acid) (PLGA 50:50, MW 17 kD) to form micrometer-sized (3.8 ± 1.6 μm) aggregates composed of small (180 ± 70 nm) NPs (Fig. 1A). Microaggregates of PLGA NPs are stable in aqueous solutions because of their hydrophobicity (12, 13). But when exposed to mechanical forces that overcome the attractive forces holding the NPs together, such as hemodynamic shear stresses, the aggregates break apart (Fig. 1B), much like a wet ball of sand disperses into individual grains when rubbed in one’s hands.

Fig. 1

Microscale SA-NTs disperse into NPs only when exposed to pathological shear stresses. (A) Scanning electron micrographs of the microscale (~2 to 5 μm) SA-NTs (left) and the PLGA NPs (~180 nm) used to produce them (right). Scale bar, 2 μm. (B) Fluorescence micrographs demonstrating intact SA-NTs (top) and NPs dispersed after their exposure to 1000 dyne/cm2 for 10 min by using a rheometer (bottom). Scale bar, 10 μm. (C) Quantification of release of fluorescent NPs from the SA-NTs as a function of shear revealed that exposure to pathological levels of shear (≥100 dyne/cm2 for 1 min) caused large increase in the breakup of the microscale aggregates into NPs compared with physiological levels of shear (1 or 10 dyne/cm2) (*P < 0.005). (D) CFD simulations comparing fluidic shear stress in a normal coronary artery (left) and a stenotic vessel with a 60% lumen obstruction (right). Left inset shows the corresponding angiogram of the stenotic left coronary artery in a 63-year-old male patient.

To determine the shear-sensitivity of this NP deployment mechanism, we used a rheometer to apply controlled shear stresses in vitro to SA-NTs fabricated from NPs labeled with a fluorescent tag. We detected an 8- to 12-fold increase in the concentration of released NPs when the level of shear reached 100 dyne/cm2 or higher (Fig. 1C). This range of fluid shear stress is relevant in many vascular diseases. For example, computational fluid dynamics (CFD) modeling of flow within normal and stenotic human left coronary arteries based on ultrasound imaging (supplementary materials, materials and methods) revealed that the level of shear that induces NP release in vitro is similar to that generated by a 60% lumen obstruction (Fig. 1D), whereas normal coronary vessels experience a 20% lower level of shear stress (~10 to 30 dyne/cm2) that does not cause disruption of the SA-NTs.

To determine whether these SA-NTs can target agents selectively to stenotic regions under relevant hemodynamic flow conditions, we carried out studies in a three-dimensional (3D) microfluidic model of vascular narrowing fabricated from poly-dimethylsiloxane (PDMS) that was designed to mimic regions of living blood vessels with 90% lumen obstruction (Fig. 2, A and B). Based on CFD modeling, such a constriction generates an ~100-fold increase in shear at the stenotic site (Fig. 2C). Perfusion of SA-NTs [100 μg/ml in phosphate-buffered saline (PBS)] through these microfluidic devices resulted in a 16-fold increase in the release of free NPs, as measured in the solution flowing downstream of the obstruction compared with fluid flowing through unobstructed microfluidic channels of similar dimensions (Fig. 2D). Moreover, fluorescence microscopic imaging confirmed that released NPs accumulated in endothelial cells cultured on the inner surface of the artificial microfluidic vessel just distal to the narrowed region, whereas minimal uptake occurred in cells lining the channel before the constriction (Fig. 2E).

Fig. 2

Shear-induced dissociation of SA-NTs and NP targeting under hemodynamic conditions in microfluidic devices. (A) A microfluidic vascular stenosis model showing how SA-NTs (large spheres) should remain intact in the prestenotic region but then break up into NPs (small spheres) when they flow through a constriction (90% lumen occlusion) and can accumulate in endothelial cells lining the bottom of the channel. (B) A photograph of the microdevice that mimics vascular stenosis fabricated in PDMS. (C) CFD simulations of the microfluidic device shown in (B) demonstrating that a physiological inlet shear rate of 1000 s−1 (10 dyne/cm2) upstream from the constriction increases to a pathological level of ~100,000 s−1 (1000 dyne/cm2) in the region displaying 90% lumen occlusion. (D) Graph shows a >10-fold increase in release of fluorescent NPs from SA-NTs when they are perfused through the channel shown in (B) compared with flow through an unconstricted channel (*P < 0.005). Fluorescent micrographs compare the NPs collected in the outflow from the control channel (top) versus the constricted channel (bottom). Scale bar, 2 μm. (E) Graph demonstrates that many more fluorescent NPs accumulate in endothelial cells lining the downstream area (poststenosis) of the constriction relative to an upstream area (P < 0.005). Fluorescence microscopic images show cells from regions before (left) and after (right) the constriction. Scale bar, 20 μm.

To evaluate their functional potential, we fabricated SA-NTs containing fluorescent NPs coated with the U.S. Food and Drug Administration–approved thrombolytic drug, tissue plasminogen activator (tPA), using biotin-streptavidin chemistry (~5 × 105 tPA molecules/microaggregate) and tested their ability to dissolve blood clots. To examine the general utility of this shear-targeted nanotherapeutic approach (Fig. 3A) for removal of natural clots formed endogenously in vivo, we studied the effect of bolus injection of thrombolytic SA-NTs in an established mouse arterial thrombus model, in which clot formation is triggered by injuring the vessel wall by direct exposure to ferric chloride (1416). Real-time, intravital, fluorescence microscopic studies confirmed that this treatment resulted in formation of large blood clots within minutes in injured mesenteric arteries (~100 μm diameter, normal wall shear stress ~30 dyne/cm2) that occluded the diameter by more than 80% (Fig. 3, B and C) and caused the local shear stress to increase by more than 15-fold (~450 dyne/cm2) in these regions, as determined by using an optical Doppler velocity meter (17). Fluorescently labeled tPA-carrying SA-NTs that were injected intravenously 8 min after chemical injury preferentially accumulated in the regions of clot formation, resulting in clear microscopic visualization of these lesions (Fig. 3B). In addition, the locally deployed tPA-coated NPs induced progressive surface erosion of the thrombi, with complete clearance of occlusions occurring within 5 min after SA-NT injection (Fig. 3, B and C, and movie S1). Continuous monitoring of unobstructed vessels for up to 15 min in the mesenteric bed revealed that intact microscale NP aggregates were present throughout the course of the study, indicating that circulation of the SA-NTs through the normal vasculature did not induce microaggregate disruption.

Fig. 3

Shear-targeting of a thrombolytic drug in a mouse arterial thrombosis model using SA-NTs. (A) Schematic of the experimental strategy. Ferric chloride injury initiates formation of a thrombus (top) that grows to partially obstruct blood flow (top middle). Intraveneously injected SA-NTs dissociate into NPs at the thrombus site because of the rise in local shear stress (bottom middle). Accumulation of tPA-coated NPs and binding to the clot at the occlusion site progressively dissolve the obstruction (bottom). (B) Sequential intravital fluorescence microscopic images of a thrombus in a partially occluded mesenteric artery recorded over a 5-min period beginning after bolus injection of fluorescent tPA-coated SA-NTs (1 mg NPs; 50 ng tPA) 8 min after injury initiation. Scale bar, 100 μm. The NPs accumulate at the clot, first visualizing its location and then demonstrating clearance of the clot within 5 min after injection at the bottom (movie S1). (C) Sequence of intravital fluorescence microscopic images recorded over a 5-min period showing fluorescently labeled platelets accumulated within a forming thrombus that partially occludes a mesenteric artery 8 min after injury. Thrombosis is then treated with injection of either tPA-carrying SA-NTs (50 ng tPA) (left) or PBS (right). Scale bar, 100 μm. The clot on the left is greatly reduced in size within 5 min after SA-NTs injection (movie S2), whereas the control vessel on the right fully occludes over the same time period. (D) Graph showing that a bolus injection of SA-NTs carrying 50 ng tPA (tPA-SA-NT) significantly delayed the time to full vascular occlusion in FeCl-injured vessels (***P < 0.0005), whereas administration of the same concentration of soluble tPA (free tPA), uncoated SA-NTs (bare SA-NT), tPA-coated NPs that were artificially dissociated from SA-NTs before injection (dispersed tPA-NPs), and heat-fused NP microaggregates with tPA coating that do not dissociate (fused SA-NT) did not produce any significant delay in thrombosis.

Shear-induced release of tPA-coated NPs from the SA-NTs reopened the obstructed mesenteric arteries and significantly delayed the time to vessel occlusion (29 ± 7 min with tPA-coated SA-NTs versus 12 ± 3 min with PBS), when vessel patency was monitored by using intravenous injection of fluorescently labeled platelets (~2.5% of total platelets) (Fig. 3, C and D, and movie S2). In contrast, administration of the same dose of soluble tPA (free tPA), predissociated tPA-NPs, or heat-fused tPA-NP microaggregates (which do not dissociate in high shear) did not delay thrombosis in this model (Fig. 3D). Careful analysis of these results also revealed that even when a vessel is almost fully occluded, the microscale tPA-coated SA-NTs that bind to the surface of the clot can actively degrade and “recanalize” the clot. Once this happens, flow and shear stress rapidly increase once again, and this feeds back to activate other tPA-carrying SA-NTs, resulting in full clot removal (Fig. 3C and movie S2). Taken together, these results provide proof-of-principle that the SA-NT technology can be used to target clot-lysing agents to vascular occlusions, in addition to providing a way to image these lesions in real time in situ.

To explore the potential value of the SA-NTs for treatment of life-threatening embolic occlusions, we first tested their ability to dissolve experimentally induced fibrin clots in vitro. When preformed fibrin clots [250 ± 150-μm diameter produced by a water-in-oil emulsion technique (18)] were injected into microfluidic channels that contained constricted regions (80 μm high, 500 μm wide), the fibrin emboli lodged in the devices and partially obstructed flow in the channels (Fig. 4A). When SA-NTs (100 μg/ml) carrying tPA (50 ng/ml) were infused at physiological flow rates through the clot-occluded microfluidic channels, the shear-dispersed fluorescent tPA-coated NPs accumulated at the surface of the artificial emboli, progressively dissolving the clots and reducing their size by one half within an hour of treatment (Fig. 4A and movie S3). In contrast, treatment with soluble tPA at the same concentration and flow conditions had negligible effects (<5% reduction in clot size) (Fig. 4B).

Fig. 4

Shear-targeting of a thrombolytic drug to vascular emboli in vitro and therapeutic delivery in a mouse pulmonary embolism model. (A) Time-lapse fluorescence (top) and bright field (bottom) views of artificial microemboli (~250 μm) in a microfluidic channel before (0 min) and 1 or 60 min after injection of SA-NTs coated with tPA (50 ng/ml), showing progressive lysis of the clots over time (movie S3). Scale bar, 100 μm. (B) Graph showing enhanced emboli lysis kinetics induced by tPA-coated SA-NTs (50 ng/ml, blue line) compared with soluble tPA (red line). (C) Fluorescence (top) and phase contrast (bottom) views of histological sections of normal (left) versus obstructed (right) pulmonary arteries showing local accumulation of fluorescent NPs within the obstructing emboli in a mouse ex vivo lung ventilation-perfusion model. Scale bar, 100 μm. (D) Graph showing almost a 20-fold increase (P < 0.005) in accumulation of fluorescent NPs in regions of obstructed versus nonobstructed vessels, as detected with microfluorimetry. (E) Real-time measurements of pulmonary artery pressure in the ex vivo pulmonary embolism model showing that the tPA-coated SA-NTs (blue line) reversed pulmonary artery hypertension within approximately 1 hour, whereas the same concentration (50 ng/ml) of free tPA was ineffective (red line). (F) Graph showing that tPA carrying SA-NTs normalize pulmonary artery pressure within 1 hour, whereas the same concentration of free tPA (50 ng/ml) or a 10-times-higher dose (500 ng/ml) did not reduce pulmonary artery pressure (*P < 0.005); only a 100-fold-higher dose (5000 ng/ml) produced similar effects. (G) Survival curve showing that almost all (86%) of the mice injected with the tPA-coated SA-NTs survived, whereas all control mice died within 45 min after injection of fibrin clots that caused acute emboli formation.

Next, we tested the ability of this shear-activated tPA delivery system to reverse the effects of acute pulmonary embolism in an ex vivo whole-mouse-lung ventilation-perfusion model. A solution containing the preformed fibrin clots similar to those tested in the microfluidic channel were infused (0.1 ml/min for ~5 min; 1 × 103 clots/ml) through the pulmonary artery of the perfused lung. Occlusion of pulmonary blood vessels by multiple microemboli (Fig. 4C) caused the pulmonary artery pressure to increase by approximately threefold compared with its normal value (30 versus 8 mm Hg) (Fig. 4E). We then perfused tPA-coated SA-NTs (100-μg/ml microscale aggregates containing NPs coated with 50 ng/ml tPA) through the pulmonary artery at a physiological flow rate (0.5 ml/min). Fluorescence microscopic analysis of tissue sections again confirmed that the tPA-NPs localized selectively at regions of vascular occlusion, producing a >25-fold increase in accumulation of NPs at these sites, (Figs. 4, C and D). Progressive lysis of the emboli by the tPA-NPs resulted in normalization of pulmonary artery pressure levels within 1 hour in this ex vivo model (Fig. 4E). In contrast, perfusion of soluble tPA at the same concentration as that delivered on the injected tPA-coated NPs (50 ng/ml), or even at a 10-times-higher dose (500 ng/ml), failed to produce any meaningful response (Fig. 4F). In fact, similar clot-lysing effects and hemodynamic changes were only observed when we administered a 100-times-higher concentration of soluble tPA under identical flow conditions (Fig. 4F); this dose in mice (~2 mg/kg) is comparable with the therapeutic dose commonly used in humans (~1 mg/kg).

We then studied pulmonary embolism in living mice by infusing smaller preformed fluorescent fibrin clots (<70 μm; ~1000 clots) into the jugular vein of anesthetized mice, which accumulate in peripheral blood vessels in the lungs, as previously described (19). SA-NTs coated with tPA were then infused either immediately after injection of emboli or 30 min after they formed. Quantitation of the total area of fluorescent emboli visualized in the lungs by use of computerized image analysis confirmed that administration of the tPA-coated SA-NTs resulted in reduction of both total clot area and clot number by >60% when administered immediately after injection of emboli and by >30% when infused one half hour after embolism (fig. S2).

To further examine the potential clinical relevance of our approach for treatment of life-threatening acute massive embolism, we infused a solution containing larger fibrin clots (150 ± 80-μm diameter; ~ 100/injection), which accumulate in the main pulmonary arteries (19) much as they do in humans with pulmonary embolism, and the mice were then immediately infused with tPA-coated SA-NTs or with carrier fluid for 45 min. All control animals died within 1 hour after infusion of the clots (0% survival, n = 7 mice) (Fig. 4G), whereas >80% of the treated mice survived (6 out of 7), and none of these SA-NTs–treated animals displayed any visible symptoms of respiratory distress.

The major potential advantage of the SA-NTs is their ability to enhance the safety of thrombolytic therapies by substantially reducing the drug dose required to be effective, as demonstrated by the ability of SA-NTs to clear pulmonary emboli when coated with a tPA dose ~1/100th of that required for induction of similar clot-lysing effects by free tPA. SA-NTs also could help to minimize unwanted bleeding and neurotoxicity because they are cleared rapidly from the circulation (80% clearance in 5 min) (fig. S3), and because of their larger size, they should not diffuse as easily into injured tissues as free tPA does. Additionally, we have not observed any abnormal bleeding in mice during the extensive surgical manipulations that were involved in our pulmonary embolism and mesenteric artery occlusion models. Nevertheless, before advancing these SA-NTs toward clinical studies in the future, finer control over the size of the microscale aggregates and their pharmacokinetics will be required to ensure that they safely pass through all microvessels and are sustained in the circulation at effective levels (20). Alternative methods to link tPA to NPs [such as direct conjugation by amine-carboxylate coupling or coupling based on biocompatible heterobifunctional polyethylene glycol linkers (21, 22)] also will need to be explored so as to avoid immune responses associated with streptavidin/biotin conjugation and to increase conjugation efficiency as well as optimize tPA activity.

A previously described thrombo-prophylaxis strategy based on coupling plasminogen activators to carrier erythrocytes has shown promising results in preventing thrombosis in various animal models (2325). The SA-NTs described here can potentially be used to prevent formation of thrombi that partially occlude vascular flow, such as occurs, for example, when a stable atherosclerotic plaque is transformed into a life-threatening vulnerable plaque. However, in contrast to the erythrocyte delivery approach, which is limited to prevention of nascent clot formation, the shear-activated drug targeting strategy also offers the ability to treat and dissolve preexisting fibrin clots, such as those found in patients with stroke and myocardial infarction as well as atherosclerosis. In this proof-of-principle study, we focused on SA-NTs coupled to a thrombolytic agent, tPA, which has high affinity for fibrin, and thus, they were efficiently delivered to fibrin clots (fig. S1). In addition, these tPA-coated SA-NTs were loaded with a fluorescent dye that also concentrated at vascular occlusion sites. In the future, it should be possible to design and fabricate SA-NTs containing various drugs or imaging agents for localized treatment and real-time visualization in a variety of pathologies associated with vascular obstruction.

These findings illustrate how nanoengineering approaches inspired by pathophysiological mechanisms can be used to develop safer and more effective therapeutic strategies. In contrast to drug targeting mechanisms that focus on expression of distinct molecular species that can vary between tissues or patients, shear stress increases as a function of narrowing of the lumen diameter in all patients, regardless of the cause or location of obstruction, thus offering a robust and broadly applicable targeting strategy. With further refinement, the shear-activated drug targeting nanotechnology described here might, for example, allow the immediate administration of clot-busting drugs to patients suspected to have life-threatening clots in the brain, lung, or other vital organs by emergency technicians or other caregivers, even before the patient has reached a hospital setting.

Supplementary Materials

Materials and Methods

Figs. S1 to S3

References (2632)

Movies S1 to S3

References and Notes

  1. Acknowledgments: We thank D. Huh, K. Roberts, and R. F. Valentini for helpful comments and K. Johnson and D. Stanton for help with the graphics. This work was supported by a U.S. Department of Defense Breast Cancer Innovator award BC074986 (to D.E.I.), grants from Novartis Pharmaceuticals and Boston Scientific (to C.L.F. and A.U.C.), and the Wyss Institute for Biologically Inspired Engineering at Harvard University. N.K. is a recipient of a Wyss Technology Development Fellowship. Harvard University and the authors (D.E.I., N.K., M.K.) have filed two patents related to this work: (i) Shear-Activated Nanotherapeutics for Drug Targeting (patent application pending) and (ii) Shear Controlled Release for Stenotic Lesions and Thrombolytic Therapies (patent application pending WO 2012/074588).
View Abstract

Navigate This Article