Blood-derived biomaterial for catheter directed arterial embolization

Abstract

Vascular embolization is a life-saving minimally invasive catheter-based procedure performed to treat bleeding vessels. Through these catheters, numerous metallic coils are often pushed into the bleeding artery to stop the blood flow. While there are numerous drawbacks to coil embolization, physician expertise, availability of these coils and their costs further limit their use. Here, we developed a novel blood derived embolic material (BEM) with regenerative properties that can achieve instant and durable intra-arterial hemostasis regardless of coagulopathy. In a large animal model of vascular embolization, we show that BEM can be prepared at point-of-care within 26 minutes using fresh blood, it can be easily delivered using clinical catheters to embolize renal and iliac arteries, and achieve rapid hemostasis in acutely injured vessels. In the swine arteries, BEM increases cellular proliferation, angiogenesis and connective tissue deposition suggesting vessel healing and durable vessel occlusion. BEM has significant advantages over embolic materials used today, making it a promising new tool for embolization.

Graphical Abstract

Blood-derived biomaterial was engineered for catheter directed arterial embolization. This embolic biomaterial shows remarkable modulus, radiopacity, retrievability, and bioactivity capable of achieving durable embolization in porcine iliac and renal arteries. The blood derived biomaterial is a simple and affordable embolic biomaterial that can also be prepared at the point of care for emergency embolization procedures including the treatment of bleeding arteries.

In the early 1970s, Charles Dotter, an interventional radiologist, performed the first catheter directed embolization using clots derived from patient’s own blood to treat an acutely bleeding artery of the stomach.^[1] This creative approach had numerous benefits; for example, blood clots could be quickly made at the point-of-care and blood is widely available lowering the cost and access to these life-saving embolic agents. However, autologous blood clots for embolization were quickly abandoned because natural thrombolysis led to recanalization within hours to days resulting in re-bleeding. To embolize a bleeding artery today, metallic coils are pushed through catheters repeatedly until the coil mass inside the artery slows the blood and clots. These coils, however, have many drawbacks including limited efficacy in the anticoagulated or the coagulopathic patient, they produce significant imaging artifacts limiting evaluation of the adjacent soft tissue, when deployed they are not designed to be retrievable and are not cost-effective.^[2] For example, to embolize an aneurysm, many coils are often used; in fact, in one study it was shown that the cost of embolizing an aneurysm can exceed $150,000, imposing a major cost burden to healthcare.^[3] Liquid embolic agents such as N-butyl-2 cyanoacrylate and ethylene vinyl alcohol polymers are also available, but they are limited to the treatment of cerebral arteriovenous malformations (AVM) and they have their own disadvantages.^[3–4] There is concern for entrapment of the catheter to the liquid embolic agent, polymerization of these liquids can spread beyond the intended site of embolization and some of its components such as DMSO is toxic, which can cause vasospasm and necrosis.^[1c] Embolic materials to-date aim to achieve only occlusion of the blood vessels. However, the next-generation embolic agents should also leverage cellular mechanisms to enhance connective tissue deposition and promote vascular healing so that the occlusion is durable and rebleeding does not occur in the injured arteries or ruptured aneurysms.

Platelet-rich fibrin (PRF), which is derived from the blood, has gained considerable attention due to its regenerative properties, simplicity in preparation, and biocompatibility.^[5] Platelets are responsible for the release and activation of growth factors such as transforming growth factor beta (TGF-β), platelet derived growth factor (PDGF), and vascular endothelial growth factor (VEGF), adhesion molecules, coagulation factors, cytokines and cells such as fibroblasts, neutrophils, macrophages and mesenchymal stem cells that are critical to wound healing.^[6] These components induce the mitogenic activity of fibroblasts and endothelial cells, the migration of fibroblasts and mesenchymal stem cells, and decrease the proteolytic degradation of endogenous fibrogenic factors, thereby providing the necessary elements for wound healing.^[7] In vivo studies, conducted in various animal models such as mice, rats, rabbits and pigs, have shown the beneficial effects of PRF on wound healing.^[6] For example, faster wound healing has been observed when PRF is applied to wounds in dorsal tissues of rats.^[8] Roy et al. has found that PRF can yield improved angiogenesis and collagen deposition in porcine chronic wounds.^[7b] In addition to cutaneous wound healing, two studies have demonstrated favorable effects of PRF on heart-related injuries.^[9] PRF has also shown to be effective and safe in numerous clinical settings including the management of chronic leg ulcers such as diabetic foot ulcers and venous leg ulcers, and the repair of facial soft tissue defects, urethra-cutaneous fistulas, chronic rotator cuff tears and periodontal tissue defects.^{[5a, 10]}

Here, we developed two novel blood-derived embolic materials (BEM) with regenerative properties that can be rapidly prepared and delivered using clinical catheters to achieve instant and durable hemostasis regardless of coagulopathy (Scheme 1). BEM has significant advantages over embolic materials used today, making it a promising new tool for embolization.

Scheme 1.

Fabrication of blood-derived embolic material (BEM) for transarterial embolization. Schematic shows the components of the two types of BEM; for elective procedures, PRF derived from blood is lyophilized for long-term storage at 4 °C. When needed for an endovascular procedure, it is mixed with nanoclay to produce BEM. For emergent procedures, the purified PRF is mixed with nanoclay to produce the point-of-care BEM; in this form, it can be used immediately, i.e., to embolize the renal or the iliac artery.

To make BEM, the platelet rich fibrin fraction (PRF) from a freshly collected aliquot of pig whole blood was isolated. This straw-colored, gel-like material, which includes polymerized fibrin mesh, growth factors, and platelets, offers several favorable features for an embolic agent such as antibacterial and regenerative properties.^{[5b, 11]} PRF was further processed to prepare lyophilized PRF (L-PRF) (Scheme 1 and Figure S1, Supporting Information); in this version, L-PRF could be stored at 4°C for later use to make BEM. SDS-PAGE of L-PRF and PRF were similar with no significant evidence of degradation during the lyophilization process (Figure S2A, Supporting Information). Furthermore, preserved integrity of selected growth factors in L-PRF and freshly prepared PRF was demonstrated by immunoblotting and enzyme-linked immunosorbent assay (ELISA) experiments (Figure S2B to D, Supporting Information). Some variability was noted in the detection level of VEGF-A; this may represent normal variation in the PRF composition of each pig (Figure S2D, Supporting Information). To demonstrate the biological activity of these growth factors in PRF and L-PRF, cell migration and proliferation assays were performed using L-929 mouse fibroblasts. Both demonstrated significantly enhanced cellular proliferation (p < 0.0001) and migration (PRF; p < 0.05, L-PRF; p < 0.01) (Figure S2E to G) suggesting that L-PRF and PRF based BEM will provide adequate bioactivity which is important to promote fibroinflammatory response following arterial embolization to create a durable occlusion avoiding recanalization.

Next, to make L-PRF injectable through clinical catheters, they were combined with laponite nanoclay (NC) to develop a shear-thinning BEM. NC is known to have antibacterial properties and has shear-thinning characteristics that is favorable for injectability.^[12] It is comprised of nanosized silicate disks carrying negative charges on the surfaces and positive charges along the rims, which could help to form ionic interactions with PRF proteins (i.e., fibrin). On scanning electron microscopy (SEM), BEM demonstrated a porous microstructure (Figure 1A and Figure S3, Supporting Information).

Figure 1.

Characterization of BEM. (A) Representative SEM images of L-PRF, NC, and BEM and their gross appearance. (B) Flow curves of NC and BEMs revealing their shear thinning properties. (C) Thixotropy tests showing gels’ recoverability under oscillating low and high strains. (D) Summary of gels’ storage modulus, G’, obtained from amplitude sweeps (n=3). (E) Summary of injection forces generated by NC, BEM and BEM-EO through a 5F catheter (n=3). (F) Graphic summary of average pressure required to displace NC, BEM, and BEM-EO in a vascular occlusion model (n=3); the black dotted line indicates physiologic pressure (120 mmHg); inset shows representative displacement pressure curves. (G) Graph showing relative viability of L-929 cells following 24 hours incubation with PRF, L-PRF, NC, BEM, and BEM-EO extracts respectively (n=12). (H) Summary of sterility testing based on optical density at 600 nm showing no bacterial growth in BEM-EO at 24 hours or 1 week after inoculation; LB broth alone and LB broth inoculated with E. coli were used as negative or, positive controls respectively. (I) Rheological study showing enhanced ΔG’ of BEM-EO in contact with blood compared to blood alone. (J) Images of blood clotting study showing enhanced coagulation when blood is in contact with BEM-EO and clinically used coil fibers. (K) Fluoroscopy images of BEM-EO loaded syringes containing varying concentrations of ethiodized oil. (L) Images of BEM-EO retrieval test in a 3D printed artery model showing complete removal of BEM-EO using Penumbra system. p values determined by two-way ANOVA with Tukey’s multiple comparison, ns, not significant, ****p ≤ 0.0001. Data represented as average ± SEM.

To investigate the interaction between L-PRF and NC, we fabricated three types of BEM formulations (7.2 wt%, 8 wt%, and 9 wt% total solid content) with varying amounts of L-PRF while keeping NC content constant at 6.6 wt% (Table S1, Supporting Information). Decreased viscosity upon increased shear rate suggested the shear-thinning ability of NC and BEM favoring transcatheter injectability in all three formulations (Figure 1B). In addition, the three types of BEM demonstrated excellent recoverability under alternating low and high strain cycles, showing the ability of rapid network disruption and reconstitution regardless of oscillation frequency (Figure 1C). The addition of L-PRF into NC significantly enhanced its storage modulus, Gˈ, an indicator of BEM’s stability. In addition, a three-fold increase in modulus was observed in 9 wt% BEM (12562 ± 475 Pa) when compared to NC gel alone (4097 ± 118 Pa) (Figure 1D). BEM (9 wt%) with the highest possible L-PRF content was selected for further studies due to its excellent mechanical properties for catheter delivery and stability.

To avoid non-target embolization during a procedure, it is critical to see BEM in real-time as it exits the tip of the catheter inside the artery. To allow visualization, BEM was mixed with ethiodized oil, a common contrast agent used in clinical practice with X-ray based imaging modalities, i.e., computed tomography (CT) and fluoroscopy.^[13] A commercially available ethiodized oil was mixed with NC and L-PRF to generate BEM with ethiodized oil (BEM-EO) (Figure S4A and B, Supporting Information). Subsequent rheology testing of BEM-EO demonstrated persistent shear thinning property with excellent Gˈ (39068 ± 575 Pa), which was three times higher compared to BEM alone (Figure S4C and D, Supporting Information).

Next, we focused on the physician experience; the capability to inject by hand is of great importance for ease of use, to reduce procedure time and to lower procedure costs. With compression testing, maximum injection forces for NC, BEM, and BEM-EO were measured to be 23 ± 0.3 N, 32 ± 0.6 N, and 71 ± 0.4 N, respectively, indicating the feasibility for hand injection through clinical catheters (Figure 1E). ^[14] The enhanced modulus of BEM-EO was further confirmed by measuring the maximum pressure required to displace NC, BEM, or BEM-EO (71 ± 7 kPa, 93 ± 8 kPa, and 192 ± 7 kPa, respectively) in an in vitro vascular occlusion model. BEM-EO demonstrated a displacement pressure approximately 12 times higher than normal systolic pressure, suggesting that when injected in the artery, it will remain in place without migration or fragmentation (Figure 1F and Figure S4E and F, Supporting Information).

Next, cytotoxicity of fresh PRF, L-PRF, NC, BEM, and BEM-EO were evaluated according to ISO 10993–5 guidelines using L-929 cells.^[15] No cytotoxicity was observed with any of the tested PRF containing materials, revealing their biocompatibility (Figure 1G). Radiopaque BEM-EO was selected for use in swine experiments and therefore sterility and hemostatic ability of BEM-EO were investigated. Prior to animal studies, preparation of BEM-EO was mixed with LB broth and incubated at 37 °C; these were shown to be sterile at day 1 and day 7 (Figure 1H and Figure S5, Supporting Information). Hemostatic activity was tested using rheometry to observe time-dependent modulus changes as BEM-EO came in contact with blood. Rapid increase in modulus was demonstrated when blood came in contact with BEM-EO compared to blood alone (Figure 1I). Moreover, blood aliquots that were loaded into a 96-well plate were shown to coagulate at 5 minutes, whereas blood with BEM-EO began to coagulate at 3 minutes which is similar to coils used in clinical practice (Figure 1J). The concentration of ethiodized oil in BEM was optimized under x-ray fluoroscopy; 25 wt% of ethiodized oil mixed with BEM produced adequate radiopacity to enable tracking under x-ray based imaging modalities (Figure 1K). Another desired property of an embolic agent is the capability to retrieve them; this would allow rescue of any non-target embolization resulting from accidental delivery. Using the Penumbra Aspiration catheter system (Penumbra, Inc., Alameda, CA), which is an FDA approved device to remove clots that cause stroke, we explored whether BEM-EO could be retrieved (Figure 1L and Video S1, Supporting Information). BEM-EO was shown to be retrievable after delivery, which is a unique property that is not possible using the currently FDA approved embolic agents.

To further evaluate the biocompatibility and biodegradation profiles of NC, BEM and BEM-EO, a rat subcutaneous injection model was used (Figure S6, Supporting Information); the injectate and the surrounding tissue at day 3, 14 and at day 28 post-implantation were explanted. Histological examinations revealed amorphous eosinophilic appearing BEM and BEM-EO in the injection site (Figure 2, Figure S7A, Supporting Information); H&E staining was used for total cell counts and trichrome staining was used to demonstrate collagen deposition. In the BEM-EO group, significantly higher number of cells were observed at D14 (p < 0.01) and D28 (p < 0.05) compared to D3 (Figure 2A and B, arrows). Higher numbers of cells were also present in BEM-EO compared to NC at D14 (Figure 2A and B, arrows). Furthermore, the injectate was surrounded by a layer of cellular infiltration (Figure 2C, black line) and collagen rich fibrous capsule (Figure 2C, dotted yellow line). In D14 group, both BEM (p < 0.05) and BEM-EO (p < 0.001) had significantly thicker layer of cellular infiltration compared to NC group, suggesting that the L-PRF likely had a proliferative effect (Figure 2C and D, Figure S7A and C, Supporting Information, black line). D28 BEM-EO samples also demonstrated a thicker collagen rich fibrotic capsules, representing an increase in collagen deposition around the biomaterial compared to NC (Figure 2C and E, dotted yellow line). This enhanced fibrosis would be beneficial for stabilizing the injected material in the vessel to achieve durable embolization and prevent recanalization in the long term. The change in the injected material volume was measured from reconstructed micro-CT images of explanted tissues, as shown in Figure 2F and G; significant volume reduction was observed in BEM (p < 0.05), and BEM-EO (p < 0.01) between D3 and D28 on micro-CT images (Figure 2F and G, Figure S7E, Supporting Information), revealing their biodegradability. Analysis of the blood samples from all rats further supported the biocompatibility of NC, BEM, and BEM-EO; complete blood count (CBC) studies showed normal levels of leukocytes, red blood cells, and platelets (Table S2, Supporting Information).^[16]

Figure 2.

Assessing the histologic response following subcutaneous implantation of NC or BEM-EO in the rats’ dorsum. (A) Micrographs of H&E stained cutaneous tissue sections of NC or BEM-EO injected sites at 3, 14, or 28 days post implantation (arrow heads point to the injected biomaterial; arrows denotes infiltrating cells). (B) Summary of average cell counts within the biomaterial region showing markedly higher cell infiltration in BEM-EO treated site compared to NC at D14 after injection. (C) Histology images of Mason’s trichrome stained cutaneous tissue sections of NC or BEM-EO injected sites at 3, 14, or 28 days post implantation (dotted yellow line shows fibrous capsule thickness; black line shows region of cell infiltration). (D) Summary of cell infiltration thickness showing thicker cell infiltration layer in the BEM-EO treated site compared to NC at D14 after injection. (E) Morphometric analysis showing a thicker fibrotic capsule layer around the injected BEM-EO at D28. The thickness of fibrotic capsule is marked by dotted yellow line, as shown in (C). (F) Reconstructed micro-CT images of implanted biomaterial. (G) Volumetric analysis of the injected biomaterials showing a decrease in BEM-EO volume at D28 compared to D3. Scale bar, 2 mm in micro-CT images, and 150 μm in histology images. p values determined by ANOVA with Tukey’s multiple comparison, *p ≤ 0.05, **p ≤ 0.01, ***p ≤ 0.001, ****p ≤ 0.0001. Data are represented as average ± SEM (n=4).

Angiogenesis and cell proliferation are essential for soft tissue healing.^[17] Proliferating cell nuclear antigen (PCNA) immunostaining in the rat subcutaneous injection model showed a significantly higher number of proliferating cells in the BEM group compared to NC at D14 (p < 0.001) and D28 (p < 0.05) (Figure S8A and B, Supporting Information). Angiogenesis at the tissue-injectate interface was also evaluated using CD31 immunostaining showing a significantly higher number of vessels formed in the BEM and BEM-EO samples at D28 compared to NC (p < 0.01) (Figure S8C and D, Supporting Information). These findings are also consistent with in vitro cell proliferation and migration assays suggesting that BEM-EO can incite a regenerative response that promotes biodegradation and fibrosis, which are important properties for achieving a durable occlusion in the arterial lumen.

Next, a pre-clinical model of arterial embolization in swine was used to test the capability of BEM-EO to achieve embolization without recanalization in a state of anticoagulation (i.e., Activated Coagulation Time (ACT) > 300 sec). Following intra-arterial access into the carotid artery, a clinical catheter was advanced to the distal aorta and digital subtraction angiography (DSA) was performed revealing the normal arterial anatomy of the pig iliac artery (Figure 3A). Under real-time fluoroscopy, the catheter tip was brought to the mid-portion of the iliac artery and syringes with BEM-EO were connected to the catheter via a Luer lock and injected into the internal iliac artery (IIA) over 15–20 seconds; BEM-EO demonstrated excellent visibility during injection (Figure 3B). Immediately after embolization, DSA from the distal aorta was again performed showing instant occlusion of the iliac artery; no evidence for fragmentation, distal migration or non-target embolization was observed (Figure 3C and Video S2, Supporting Information). In contrast to metallic coils, these anticoagulated pigs demonstrated that thrombosis is not necessary for BEM-EO to achieve occlusion. Following embolization, 4 pigs were sacrificed after 1 hour and another 4 pigs were survived for two weeks and sacrificed following whole-body CT angiogram (CTA) imaging. After 14 days post-embolization, CTAs showed that BEM-EO was still visible in IIA without any artifact, fragmentation, displacement or migration, or BEM-EO recanalization (Figure 3D and E). Fluoroscopic images of all animals and CTA studies of day 14 pigs consistently showed successful embolization (Figure S9, Supporting Information).

Figure 3.

Catheter-directed embolization of iliac and renal arteries using BEM-EO in swine. (A) Pre-embolization angiography showing patency of internal iliac artery (IIA) (white arrow). (B) Single-shot x-ray fluoroscopic image of BEM-EO in the IIA after embolization (white arrow). (C) Post-embolization DSA confirming occlusion of the IIA (white arrows). (D) Axial computed tomography (CT) image of an embolized IIA; white arrow shows BEM-EO. (E) 3D reconstructed CTA image of distal aorta and iliac arteries; embolized iliac artery with the bright BEM-EO casting the IIA (white arrow). (F) Micrographs of stained histologic cross sections of IIA occluded with BEM-EO at 1 hour and 2 weeks following embolization; 2-week survival group, extensive concentric fibroinflammatory reaction with disrupted elastin in the arterial wall (black arrows). (G) Morphometric analysis of arterial wall medial thickness assessed in elastic stained histologic sections. (H) Summary of PCNA positive cell counts shows significant increase at 2 weeks following embolization. (I) Pre-embolization angiography showing patent renal artery segmental branches (arrow). (J) Fluoroscopic image showing BEM-EO in renal artery after embolization (arrow). (K) Post-embolization digital subtraction angiography showing complete occlusion of renal artery with BEM-EO (between arrows). (L and M) Axial CT image and 3D rendered CT image of embolized kidney showing visible BEM-EO inside the artery with no imaging artifact (white arrow). (N and O) Micro-CT, gross view, and histology images of pig kidneys at 1 hour and 2 weeks following embolization showing BEM-EO in renal artery on micro-CT (white arrow); stained histologic images shows an arterial branch filled with BEM-EO at 1 hour and 2 weeks post embolization (black outlined area) and necrotic tubular cells are observed at two weeks after embolization (black arrow). (P) Volumetric analysis of 3D rendered CT scans of pig kidneys at 2 weeks following renal artery embolization compared to control kidney. Scale bars, 150 μm in histology images and 1 cm on gross view images. p values determined by unpaired t-test, *p ≤ 0.05, ns, not significant. Data are represented as average ± SEM (n=4).

Furthermore, on review of the CT scans by a board certified radiologist, all studies demonstrated normal flow to the distal hindlimb without any evidence of non-target embolization; in addition, there was no evidence for lymphadenopathy, or any abnormal findings in the brain, lungs, liver, or the spleen (Figure S10, Supporting Information). Just prior to sacrifice, blood samples were collected, and during necropsy, the embolized arteries were harvested and subjected to high resolution micro-CT imaging and histology. Histologic evaluation of the harvested IIAs was performed by a certified pathologist; in the one-hour non-survival group, flocculent amorphous material was seen casting the arterial lumen with minimal tissue reaction (Figure 3F). In the two weeks survival group, the arterial lumen was completely occluded by BEM-EO with extensive concentric fibroinflammatory reaction rich in macrophages, myofibroblasts and fibrin, resembling granulation tissue (Figure 3F). Scattered multinucleated giant cells and macrophages with fat droplets were also observed (Figure S11, Supporting Information). While the thickness of the tunica media layer between the two groups was similar (p > 0.05), the two-week survival group showed fibrosis and disruption of the elastic fibers of the intima and media layers (Figure 3F (black arrows) and G). Immunostaining for PCNA, however, showed a significant increase in cell proliferation (p < 0.05) in the embolized IIAs at 2 weeks compared to the non-survival group (Figure 3F and H).

In the pelvis, there is often extensive redundancy in the vascular flow that helps bypass an occlusion. IV contrast CTA images in the 2-week survival group showed contrast opacification distal to BEM-EO; while this almost certainly represents collateral blood flow, recanalization could not be entirely excluded. To assess for recanalization, an end organ artery without arterial redundancy was embolized; in this case, the main renal artery of the kidney was selected for BEM-EO embolization. A total of 8 pigs had the main renal artery embolized under fluoroscopy in a state of anticoagulation during the procedure; 4 were sacrificed at 1 hour and another 4 were sacrificed after two weeks following embolization. Figure 3I shows normal renal angiography; subsequently, BEM-EO was injected into the renal artery through a 5F clinical catheter during real-time fluoroscopy (Figure 3J). Following embolization, DSA images showed the absence of flow through the embolized renal artery (Figure 3K and Video S2, Supporting Information). Just prior to sacrifice in the survival group, CTA was performed again showing complete occlusion of the renal artery without any recanalization or imaging artifact (Figure 3L and M). Renal artery embolization with BEM-EO was successfully achieved in all animals (Figure S12, Supporting Information). Concordantly, CTA studies showed that there was no evidence for contrast enhancement in the embolized kidneys at two weeks demonstrating unequivocally that recanalization was not observed (Figure S12, Supporting Information). Both embolized and contralateral normal kidneys were harvested and evaluated by high resolution micro-CT imaging and by histology. Micro-CT images demonstrated complete filling of the renal artery with BEM-EO (Figure 3N and O). On histology, the kidney parenchyma in the non-survival group was still viable; BEM-EO was seen in the hilar and segmental arteries (Figure S13A to D, Supporting Information). In the 2-week survival group, uniform ischemia in the renal parenchyma with a fibroinflammatory reaction in the occluded hilar and segmental arteries were again noted (Figure S13E to H, Supporting Information). These changes were consistent with significant loss of volume (p<0.05) in the embolized kidneys at two weeks compared to the non-treated kidneys, indicating successful embolization of the renal artery (Figure 3P).^[18]

To assess for any systemic side effects of BEM-EO embolization, blood samples were collected before and after embolization in each pig. Complete blood count (CBC), basic metabolic panel (BMP), liver function tests (LFTs) and cytokines levels using a protein array showed values that were unremarkable and within normal range (Table S3, Supporting Information). At two weeks, an increase in creatinine level was observed in the cohort that received renal artery embolization indicating an expected functional outcome of successful embolization with BEM-EO.

On further micro-CT analysis of the embolized iliac arterial segments, the one-hour non-survival group showed complete filling of the arterial lumen with BEM-EO on both coronal and axial images; the corresponding H&E images showed uniform filling of the arterial lumen (Figure 4A). The two-week survival group, however, showed a more heterogeneous appearance on coronal views suggesting that degraded BEM-EO over time had been replaced by fibrotic tissue. Axial CT image and the corresponding H&E image show the characteristic concentric fibroinflammatory response to BEM-EO (Figure 4A). To determine the degradation profile of BEM-EO over two weeks in the survival group, extensive image analysis was performed to segment the BEM-EO inside the artery from the surrounding connective tissue; these data revealed that more than 63% of the BEM-EO had biodegraded at two weeks (Figure 4B). To assess stability of the BEM-EO composition, samples were stored in a tube at 37 °C and serially imaged at high resolution using micro-CT at day 0, 3, 7, 14 and 70 (Figure 4C). Extensive image analysis of these tubes was performed including counting of hypodense foci within the micro-CT images and Hounsfield unit measurements throughout the tube; these data demonstrated no significant differences between the tubes over time suggesting that BEM-EO is stable and that phase separation of its components did not occur (Figure 4D and E).

Figure 4.

Assessing time-dependent structural changes of BEM-EO using micro-CT. (A) 3D rendering of micro-CT scans and a corresponding histologic image of an iliac artery at 1 hour and 2 weeks following embolization with BEM-EO. These images demonstrate time dependent morphologic changes; the non-survival 1 hour group shows uniformly occluded artery that progresses to fragmentation in the 2-week group with intervening non-enhancing regions replaced by fibrotic tissue. Yellow dotted line in specimans P4 and P8 represents the location where the axial micro-CT image and the corresponding H&E image were obtained, scale bar, 2.5 mm. (B) 3D total volume of BEM-EO was computed from micro-CT scans of embolized iliac arteries using segmentation software; these data showed that there was 63% reduction in biomaterial volume over two weeks. (C) Serial scans of BEM-EO inside a tube stored at 37 °C for 0, 3, 7, 14 and 70 days showing consistent dispersion of hypodense foci overtime (arrows indicate similar areas over time), scale bar, 2.5 mm. (D and E) Time-dependent measurement of hypodense foci and radiodensity of BEM-EO loaded in tubes obtained from five levels in each tube as it appear in the schematic. Each tube was separated into 5 compartments and the measurements were performed for each compartment, showing no significant differences over time. This suggests BEM-EO is stable and phase separation into components does not occur over time. Student’s t-test was used to calculate differences in BEM-EO volume and one-way ANOVA with Tukey’s multiple comparisons test was used to assess time-dependent changes on micro-CT in vitro. ****p ≤ 0.0001, ns, not significant. Data are represented as average ± SEM.

Next, we introduce a version of BEM-EO that can be prepared at point-of-care for urgent or emergent embolization procedures. Whole blood from the same pig was centrifuged for 3 minutes at 700 rpm to produce a liquid form of PRF that is also rich in growth factors, cells, and proteins (Figure 5A).^[19] This liquid version of PRF was immediately mixed with NC and ethiodized oil to produce the point-of-care version of BEM, namely pocBEM. The process of preparing a ready-to-use pocBEM syringes from blood collection to loading took 26 ± 0.7 minutes (n=5). pocBEMs that were prepared from different pigs (n=5) were similar in viscosity profiles showing consistent shear thinning properties (Figure 5B). pocBEM also revealed excellent modulus (15685±434 Pa) and recoverability, which are necessary for its mechanical stability as an embolic material to prevent migration or fragmentation inside the artery (Figure 5C and D). The maximum injection force required to inject pocBEM through a 5 French clinical catheter was measured to be 30 ± 1.5 N (n=5), indicating the feasibility for hand injection (Figure 5E). In addition, preparation process of pocBEMs was shown to be sterile, without bacterial growth for 1 week (Figure 5F). Moreover, hemostatic ability of pocBEM was tested; compared to blood alone, blood in contact with pocBEM coagulated faster, which was also supported by rheological studies (Figure 5G and H). Lastly, SEM images and H&E staining of pocBEM revealed the presence of fibrin bundles and platelets in pocBEM (Figure S14, Supporting Information).

Figure 5.

Fabrication and characterization of point of care blood-derived embolic material (pocBEM). (A) Graph showing PDGF-B levels measured in freshly prepared PRF obtained from three different pigs. (B) Shear rate sweeps of pocBEMs from five different pigs showing similar viscosity profiles. (C) Graph showing G’ of different pocBEMs determined for amplitude sweeps performed at 10 rad s⁻¹ (Dashed line indicates the average G’ of 15685 Pa in all pocBEMs). (D) Thixotropy test revealing excellent recoverability of all pocBEM formulations. (E) Compression test showing injectability of pocBEMs with an average force of 30 N (dashed line). (F) OD₆₀₀ measurements obtained at 1 day and 1 week following inoculation with pocBEMs prepared under sterile conditions, showing no bacterial growth. (G) Rheological study showing rapid increase in ΔG’ when blood is in contact with pocBEM compared to blood alone. (H) Representative test of thrombogenic potential of pocBEM showing accelerated clotting time compared to blood alone. p values were determined by two-way ANOVA with Tukey’s multiple comparison, ns, not significant, ****p ≤ 0.0001. Data are represented as average ± SEM.

Following preparation, pocBEM was packed into syringes and injected into the same pig. Iliac arteries of eight pigs and renal arteries of four pigs were successfully embolized using pocBEM. During embolization, pocBEM was visible in real-time under fluoroscopy (Figure S15A, Supporting Information) achieving instant hemostasis, with subsequent DSA showing absence of flow in the embolized artery (Figure 6A to C). Pigs were euthanized at one-hour-post-embolization. All embolized arteries were harvested for micro-CT and histologic evaluation. On micro-CT, pocBEM entirely occluded the lumen of the iliac artery (Figure 6D and Figure S15B and C, Supporting Information). On histology images, pocBEM appeared as an amorphous intravascular material (Figure S15D and E, Supporting Information). Moreover, to compare pocBEM with a clinically used embolic agent, coil embolization of IIA was performed (Figure 6E and F). Following unsuccessful embolization with endovascular coils in an anticoagulated state, delivery of 1–2 cc of pocBEM to the coil mass was able to achieve instant hemostasis rescuing the failed coils (Figure 6G). Furthermore, Figure 6H demonstrates that in the event of an accidental, non-target delivery of pocBEM, pocBEM could be retrieved using the Penumbra Aspiration catheter system to restore blood flow. On high resolution micro-CT, harvested iliac arteries from Figure 6H showed extensive streak artifact caused by the coils and no evidence for residual pocBEM suggesting that the material was successfully aspirated from the LIIA (Figure 6I).

Figure 6.

Transcatheter arterial embolization, retrieval, and rescue of failed embolization with coils using pocBEM in swine. (A) Angiogram of the pig iliac arteries showing patent internal iliac arteries (IIAs). (B) Digital subtraction angiography (DSA) following embolization of right IIA (RIIA) using pocBEM, showing complete occlusion. (C) DSA following embolization of both IIAs using pocBEM showing interruption of blood flow into both IIAs. (D) Micro-CT images of embolized IIAs on coronal and transverse planes (yellow dashed line indicates the location of transverse plane). pocBEM fills the RIIA and LIIA completely without imaging artifacts. (E) Angiographic images showing normal blood flow into iliac arteries. (F) Angiographic image following coil embolization of the LIIA showing failure to stop blood flow in an anticoagulated pig. (G) Angiographic image demonstrates rescue of unsuccessful coil embolization in F following pocBEM injection through the catheter resulting in complete occlusion. (H) Angiographic image of LIIA showing normal blood flow through the coils inside the IIA following pocBEM retrieval using the Penumbra Aspiration system. (I) Coronal and axial micro-CT images of LIIA where coil embolization and retrieval of pocBEM was performed (yellow dashed line shows corresponding axial section below) showing extensive streak artifacts caused by coil. LIIA did not demonstrate opacification suggesting successful aspiration of pocBEM. (J) Angiographic images showing unsuccessful renal artery embolization with coils in an anticoagulated pig as depicted in the schematic image to the right. (K) Failed coil embolization in J was successfully embolized with pocBEM; there is now absence of blood flow to the kidney. (L) Micro-CT image of embolized kidney showing pocBEM filling the renal artery proximal to the coil mass, which causes significant streaking artifact.

Next, renal artery embolization was performed using pocBEM with subsequent DSA images showing complete cessation of blood flow into the kidney (Figure S15F and G, Supporting Information). However, coil embolization of the renal artery failed to achieve hemostasis; 1–2 cc of pocBEM delivery into the coil mass in the renal artery was able to achieve instant hemostasis (Figure 6J and K). Following harvest of the embolized kidneys, high resolution micro-CT imaging was performed. Kidneys embolized with pocBEM alone demonstrated uniform filling of the main renal artery and segmental branches without causing any imaging artifacts (Figure S15H, Supporting Information). In histology, pocBEM was present in hilar and segmental arterial branches in the embolized kidney (Figure S15I, Supporting Information). Micro-CT imaging of the renal artery that received both coils and pocBEM again demonstrated significant streak artifact from the coils and uniform casting of the pocBEM embolized segments of the renal artery (Figure 6L).

To test the value of pocBEM in acute hemorrhage, direct injury to kidneys and pelvic vessels were performed using a 20 cm 18G needle (Figure S16, Supporting Information). In two kidneys, significant injury including pseudoaneurysms as well as active extravasation (Figure S16B and E) was produced using a needle trauma to the kidney under ultrasound and fluoroscopic guidance (Figure S16A to F, Supporting Information). Animals subsequently became hemodynamically unstable including tachycardia and hypotensive; immediately, these animals received catheter directed pocBEM embolization of the renal artery to achieve hemostasis and restore hemodynamic stability (Figure S16C and F, Supporting Information). In another two animals, needle trauma under ultrasound guidance was used to produce bleeding arterial pseudoaneurysms in the pelvic vessels such as branches of the external iliac artery (Figure S16G to L, Supporting Information). These branches were subsequently embolized to achieve instant hemostasis (Figure S16I and L, Supporting Information).

Embolic agents are widely used for hemorrhage control, occlusion of vascular malformations, aneurysms and tumor vasculature. Our patient-inspired embolic agents are easy to prepare using patient’s own blood and can be delivered using standard clinical catheters. Both BEM-EO and pocBEM were tested extensively in pre-clinical anti-coagulated porcine embolization models demonstrating their efficacy. An ideal embolic agent should be widely available, cost-effective, achieve instant hemostasis and should be easy to use enhancing physician experience. While BEM shares these important features, BEM is also biologically active and can be prepared at point-of-care. Current embolic agents lack intrinsic bioactive components such as growth factors, platelets and cytokines relying instead in the bodies response to the metallic coil to occlude the vessel in short term and in long term.^[1c] We showed that BEM can trigger an enhanced local cellular proliferation, promote vascularity, and biodegradation; in pigs, a fibroinflammatory response that is crucial to maintain long term occlusion and prevent recanalization as well as fragmentation.

There are several limitations to our study. Within 14 days, 63% of BEM-EO had biodegraded; we anticipate that within 6–8 weeks all intravascular BEM-EO would be replaced by occlusive connective tissue. However, long term survival experiments beyond 14 days were not performed. For visualization of the embolic material in vivo, we focused on ethiodized oil; however, incorporation of tantalum or intravenous contrast agents such as iohexol were not explored. In addition, given that the composition of the BEM-EO and pocBEM are similar, direct comparison in the large animal models was not performed.

In summary, we present two novel blood-derived embolic materials, which are radiopaque and injectable with shear thinning properties. Designed for elective procedures, BEM-EO can achieve embolization in survival and non-survival porcine models without any evidence for fragmentation, non-target embolization or recanalization. For emergent procedures, the short preparation time, injectability and high modulus make pocBEM an excellent candidate for point-of-care interventions. In addition, visibility of BEM without artifact on CT, its unique efficacy in a state of anticoagulation and its biocompatibility represents a significant advance over coils used today. Altogether, BEM has numerous advantages over current technologies with potential for rapid adoption in clinical practice world-wide.