|Home | About | Journals | Submit | Contact Us | Français|
The therapeutic goal in peripheral arterial disease (PAD) patients is to restore blood flow to ischemic tissue. Stem cell transplantation offers a new avenue to enhance arteriogenesis and angiogenesis. Two major problems with cell therapies are poor cell survival and the lack of visualization of cell delivery and distribution. To address these therapeutic barriers, allogeneic bone marrow-derived mesenchymal stem cells (MSCs) were encapsulated in alginate impregnated with a radiopaque contrast agent (MSC-Xcaps.) In vitro MSC-Xcap viability by a fluorometric assay was high (96.9% ± 2.7% at 30 days postencapsulation) and as few as 10 Xcaps were visible on clinical x-ray fluoroscopic systems. Using an endovascular PAD model, rabbits (n = 21) were randomized to receive MSC-Xcaps (n = 6), empty Xcaps (n = 5), unencapsulated MSCs (n = 5), or sham intramuscular injections (n = 5) in the ischemic thigh 24 hours postocclusion. Immediately after MSC transplantation and 14 days later, digital radiographs acquired on a clinical angiographic system demonstrated persistent visualization of the Xcap injection sites with retained contrast-to-noise. Using a modified TIMI frame count, quantitative angiography demonstrated a 65% improvement in hind limb perfusion or arteriogenesis in MSC-Xcap-treated animals versus empty Xcaps. Post-mortem immunohistopathology of vessel density by anti-CD31 staining demonstrated an 87% enhancement in angiogenesis in Xcap-MSC-treated animals versus empty Xcaps. MSC-Xcaps represent the first x-ray-visible cellular therapeutic with enhanced efficacy for PAD treatment.
Peripheral arterial disease (PAD) results from the progressive arterial lumen narrowing by atherosclerosis with symptoms mostly affecting the lower extremities. The spectrum of symptomatic PAD ranges from intermittent claudication (IC) to chronic limb ischemia. An estimated 5 million people suffer from IC,  and the prevalence is increasing in the U.S. . IC patients are typically managed conservatively with a walking program and cilostazol . Despite these therapies, the vast majority of patients continue to have pain with ambulation, which affects their quality of life. Critical limb ischemia, defined as rest pain and/or tissue loss, develops in 500–1,000 individuals per million per year . Psychological testing demonstrates quality-of-life indices similar to patients with terminal cancer . Twenty percent of patients with limb-threatening ischemia have disease that is so extensive that revascularization by bypass surgery or angioplasty/stent placement is not feasible . The only option for these patients is amputation, a costly and undesirable therapy.
The degree of the arterial occlusion and the amount to which an endogenous arteriogenic response is mounted to compensate for the occlusion dictate a PAD patient’s symptomatology. In an attempt to augment the endogenous arteriogenic response, therapeutic arteriogenesis involves the administration of exogenous agents to induce the development of conductance vessels, such as arterioles, by way of maturation of pre-existing collaterals. Arteriogenesis is important for regional perfusion and results in a 10 to 20-fold increase in blood flow .
Recently, gene, protein, and cellular therapies [7–9] have been explored in PAD patients to enhance the endogenous arteriogenic response. Specifically, bone marrow-derived mesenchymal stem cells (MSCs) have been shown to produce a variety of arteriogenic cytokines  and incorporate into the growing vessel wall [11, 12]. However, due to poor cell survival, a large number of stem or progenitor cells need to be administered. A variety of possible reasons for poor cell survival has been postulated including immune-mediated cell rejection, hypoxia, and the absence of intercellular contact survival cues [13, 14].
To overcome immune-mediated cell rejection in pancreatic islet transplantation, alginate microencapsulation was developed . This alginate microencapsulation method creates a biocompatible, semipermeable membrane that blocks antibody access to encapsulated cells but permits the diffusion of nutrients, cytokines, and waste products. Additionally, we hypothesize that the alginate may provide a mechanism for cellular adhesion and facilitate the requisite survival intercellular contact cues.
The inability to easily visualize the precise location of cell delivery hinders clinical translation of these therapies. We have previously demonstrated the ability to use magnetic resonance imaging (MRI) contrast agents to monitor microencapsulated islet transplantation . However, due to the restrictive nature of performing invasive procedures inside an MRI system, radiopaque alginate microcapsules could offer the ability to use conventional x-ray fluoroscopic imaging for cell tracking . In this study, we sought to optimize a radiopaque barium sulfate, alginate-encapsulated MSC formulation (MSC-Xcaps) formulation for the treatment of PAD using an endovascular rabbit PAD model  to demonstrate bio-compatibility and efficacy of this cellular therapeutic.
The Institutional Animal Care and Use Committee approved all animal studies. Bone marrow was aspirated from the tibia of anesthetized, male New Zealand White rabbits (NZW, < 6-month old, n = 9). Mononuclear cells were isolated using a density gradient (Histopaque-1.077 g/ml, Sigma-Aldrich, St. Louis, MO, http://www.sigmaaldrich.com/), plated, and culture expanded in vitro for two passages prior to freezing .
The proliferation capability of MSCs was demonstrated in vitro by bromodeoxyuridine (BrdU, Accurate Chemical and Scientific Corporation, Westbury, NY, http://www.accurate-chemical.com/) labeling. To determine stem cell multipotency, adipogenic (Cambrex Corporation, East Rutherford, NJ, www.lonza.com) and osteogenic (Stem Cell Technologies, Inc., Vancouver, Canada, http://www.stemcell.com/) differentiation assays were performed according to manufacturer’s protocols, and fixed cells were then stained with oil red O for adipocyte identification or a modified von Kossa’s staining for the presence of phosphate depositions for osteogenesis.
To create radiopaque alginate microcapsules, the classic alginate-poly-l-lysine alginate microencapsulation protocol  was modified to incorporate 5%, 10%, 30%, or 70% (wt/vol) barium sulfate (Sigma, St. Louis, MO, http://www.sigmaaldrich.com/) in 2% sodium alginate (Pronova UP LVG, NovaMatrix, Sandvika, Norway, http://www.novamatrix.biz). Microencapsulation was performed sterilely by suspending 106 MSCs per milliliter of the alginate-barium solution and then extruding the beads from a syringe pump (Harvard Apparatus, Holliston, MA) using an electrostatic generator . Spherical beads were collected in 100 mM calcium chloride solution and incubated for 10 minutes, rinsed with 0.9% saline, resuspended in 0.05% poly-l-lysine) (Sigma, St. Louis, MO, http://www.sigmaaldrich.com, 22–24 kDa) for 10 minutes, rinsed with 0.9% saline, and placed in 0.15% alginate (Pronova UP LVM, NovaMatrix, Sandvika, http://www.novamatrix.biz) for 10 minutes. Microcapsules without barium sulfate were created in a similar manner as were radiopaque microcapsules without MSCs. The viability of unlabeled and barium sulfate-labeled encapsulated MSCs was determined from live/dead staining,  that is, calcein (Trevigen, Inc., Gaithersburg, MD, www.trevigen.com/) and propidium iodide (Invitrogen, Carlsbad, CA, www.invitrogen.com/) staining at time points ranging from immediately after encapsulation up to 30 days postencapsulation. Viability was determined in triplicates at each time point.
Three agarose gel phantoms were designed using six-well culture plates to determine the capsule detection limits relative to barium sulfate concentrations. For the first phantom, each well contained a similar number of alginate microcapsules, but the microcapsules were formed with different concentrations of barium sulfate (i.e., 0%, 5%, 10%, 30%, and 70%). For the second phantom, each well contained different numbers of 10% barium sulfate-alginate microcapsules (i.e., 1, 10, 25, 50, 100, and 200 capsules). For the third phantom, single 10% barium sulfate-alginate microcapsules were embedded with one, two, four, or five capsules total per well with the capsules placed either 1 or 2 mm apart.
Subsequently, x-ray digital radiographs (70 kVp, 21 mA, 22–48 cm field-of-view [FOV], 90 cm source image distance [SID], 3–15 frames per second, Axiom Artis dFA, Siemens Healthcare, Forchheim, Germany, http://www.medical.siemens.com) of the first and second phantom were obtained. The contrast-to-noise ratio (CNR) of each phantom was calculated as: (SIcapsule – SIagarose)/SDnoise, where SIcapsule is the mean signal intensity in a region of interest (ROI) over the barium capsules, SIagarose is the mean signal intensity in the agarose not containing capsules, and SDnoise is the standard deviation of the noise outside the well phantom. For the third phantom, C-arm computed tomography (C-arm CT) was performed on a flat-panel angiographic system (XperCT, AlluraXper, Philips Healthcare, Andover, MA, http://www.healthcare.philips.com/) using standard imaging presets (115–120 kV; 50–110 mA; 47 cm FOV; 256 × 256 image matrix size; 240° rotation; 0.5–0.77° rotation/step; and 120 cm SID). Maximum intensity projections/multiplanar reformats of the reconstructed C-arm CTs and x-ray digital radiographs were reviewed on the vendor workstation, and the minimum number of microcapsules that could be detected or distinguished was determined by a consensus of two observers.
Female NZW rabbits were sedated with ketamine (40 mg/kg) and acepromazine (1 mg/kg) intramuscularly, induced with i.v. sodium thiopental, intubated, and maintained on general anesthesia with intermittent thiopental boluses. Hind limb ischemia was induced by a minimally invasive, endovascular method, where thrombogenic platinum coils (Vortex, Boston Scientific Cork, Ltd., Cork, Ireland, http://www.bostonscientific.com/) were deployed in the left superficial femoral artery (SFA) under x-ray fluoroscopy (Infinix CC-i, Toshiba America Medical Systems, Inc., Tustin, CA, http://medical.toshiba.com/) as previously described [18, 22]. Complete blood counts and chemistries were acquired prior to the induction of hind limb ischemia and at day 14. A 3F pigtail catheter was positioned in the distal abdominal aorta above the aortic bifurcation. After infra-arterial injection of 300 µg of sodium prusside to create maximum vasodilation, digital subtraction angiograms (DSAs) were acquired at a frame rate of 15 frames per second during a 16 ml contrast injection (4 ml/s, Iodixanol, GE Healthcare, Piscataway, NJ, http://www3.gehealthcare.com/en/Products/Categories/Contrast_Media) in the anterior-posterior view prior to SFA occlusion and ~15 minutes after SFA occlusion . Twenty-four hours after ischemia induction, rabbits were reanesthetized and randomized to receive six intramuscular injections (~0.25 ml per injection) into the ischemic medial left thigh of either (a) 13 × 106 MSCs (n = 5), (b) sham 0.25 ml injections (n = 5), (c) 5,000 MSC-Xcaps/injection (n = 5), or (d) 5,000 Xcaps without MSCs/injection (n = 6). X-ray digital radiographs in the anterior-posterior and lateral views were acquired to document the location of the microcapsules. At 14 days after treatment, DSAs using the same preocclusion parameters were obtained to document the location of collateral vessels.
The CNR was determined in anterior-posterior digital radiographs of the medial thigh immediately and at 2 weeks after injection in the animals that received the microcapsules using image analysis software (Amide, http://amide.sourceforge.net/) . CNR was calculated as: (SIcapsule – SImuscle)/SDmuscle, where SIcapsule is the mean signal intensity in a ROI containing the capsule injections, SImuscle is the mean signal intensity in an ROI in the muscle where capsules were not injected in ipsilateral limb, and SDmuscle is the standard deviation of the muscle ROI signal intensity. Because a consistent area outside the animal’s body was not available in the FOV for calculating the SD of the noise, the SDmuscle was used to calculate CNR.
The presence or absence of injection sites in the left medial thigh was determined by a consensus of two observers blinded to the treatment arm. If radiopacities (presumably representing injection sites) were visible, the number and location of injections were recorded for each animal at the same magnification on anterior-posterior and lateral radiographs at the time of injection and at day 14.
DSA images were analyzed in the Toshiba DICOM viewer by two blinded reviewers. A modified Thrombolysis In Myocardial Infarction (TIMI) frame count  was determined as the difference in frames between opacification of the right and left popliteal/saphenous bifurcation on the DSAs (Supporting Information Fig. 1). This served as a method to determine the physiologic perfusion differences between the left and right calves. As collateral dependent flow improved distal to the SFA occlusion, a lower modified TIMI frame count would indicate improved flow.
The angiographic frame in the DSA demonstrating maximal arterial opacification was imported into a personal computer running ImageJ (National Institutes of Health, Bethesda, MD, http://rsbweb.nih.gov/ij/), and the diameter of the distal deep femoral artery (DDFA) in the right and left limbs (Supporting Information Fig. 2) was calculated by two independent observers blinded to the treatment arm . In addition, the number of collateral vessels crossing the left medial thigh (defined as having a stem, middle, and re-entry)  was determined by a consensus of two reviewers blinded to treatment arm.
After humane euthanasia, the perfusion-fixed hind limb was harvested for histological analysis. Inflammatory scoring of the hematoxylin and eosin (H&E) staining was performed by consensus between a board certified veterinary pathologist and a veterinary pathology fellow blinded to the treatment arm at ×20 magnification using a 10 point scale (1, no inflammation to 10, foreign body reaction) based on the area occupied by inflammatory cells (e.g., mononuclear cells and neutrophils) and the degree of hypercellularity as previously described .
Immunohistochemical staining to identify vascular development was performed according to the manufacturer’s protocol on at least five animals in each treatment arm with a mouse anti-human CD31 antibody (Dako, Carpenteria, CA, www.dako.com/) and the Vectastain ABC kit (Vector Laboratories, Burlingame, CA, http://www.vectorlabs.com/) to stain the endothelial cells, and with a mouse anti-human smooth muscle actin (SMA) antibody (Dako, Dako City, Carpenteria, CA, http://www.dako.com) to stain smooth muscles in artery walls. The visualization of CD31 binding was performed with 3,3’-diami-nobenzidine (Sigma, St. Louis, MO, http://www.sigmaaldrich.com/). The visualization of SMA binding was performed with a fluorescent antibody (goat anti-mouse IgG2a Alexa Fluor 488 antibody, Invitrogen, Carlsbad, CA, http://www.invitrogen.com). For CD31 staining, digital photomicrographs (Olympus X51 and IX71 epifiuorescence microscopes equipped with an Olympus DP-70 digital acquisition system, Center Valley, PA, http://www.olympusamerica.com) at ×100 magnification in three high-power fields were acquired in regions adjacent to the injection sites or microcapsules and analyzed using a custom software program (Frida, http://demarzolab.pathology.jhmi.edu/software.html) . A vascularity score was determined as the anti-CD31-positive staining area divided by the total tissue section area. For SMA staining, digital photomicrographs at ×4 magnification in 3–4 high-power fields were acquired in regions adjacent to the injection sites or microcapsules. Green fluorescent positive areas representing vessels larger than 100 µm were counted.
The data are presented as the mean ± SD. All statistical analyses were performed using Stata (Statall, Statacorp LP, College Station, TX). Ordinal data, for example, inflammatory scoring and vessel counting, were evaluated using a Fisher’s exact test. For in vitro and in vivo continuous data sampled at multiple time points, a cross-sectional time-series regression fitted with the random-effects models by using the generalized least square (gls) estimator (xtgls, Stata) with clustering by rabbit to account for repeated measurements in the same rabbit was used. Comparison of TIME frame counts between treatment arms was performed using an analysis of variance (Stata, http://www.stata.com). Pearson’s correlation was performed to compare modified TIME frame counts to the mean CD31 vascularity scoring. A paired Student’s t test was used to evaluate in vivo CNR at day 0 and day 14. A p value of <.05 was considered statistically significant.
Adherent male rabbit bone marrow-derived stem cells (Fig. 1A) retained replicatory capacity in vitro as demonstrated by BrdU staining (Fig. 1B). The mesenchymal lineage and pluripotency of these stem cells were demonstrated by in vitro assays for downstream differentiation into adipogenic and osteogenic cell lines, as confirmed by oil red O staining (Fig. 1C) and von Kossa’s staining (Fig. 1E), respectively. Negative control cells, which were not treated with adipogenic or osteogenic medium, did not differentiate spontaneously during culture expansion (Fig. 1D, 1F).
Microcapsules containing barium sulfate showed uniformity in size with an average diameter of ~500 µm with homogenous barium sulfate incorporation (Fig. 1G). Approximately 30–45 MSCs were contained per capsule using the live-dead staining (Fig. 1H). In vitro viability assessment of encapsulated MSCs showed similar cell viability among unlabeled, 5% barium-impregnated, and 10% barium-impregnated alginate microcapsules immediately after encapsulation (84.1% ± 7%, 85.6% ± 1%, and 84.2% ± 6%, respectively) and at later time points of 1, 2, and 30 days postencapsulation (Fig. 2A). While there was variation in the total number of live-dead cells counted at each time point, there was no statistically significant change in the total number of cells counted over 30 days (Fig. 2B).
The CNR on digital radiographs of the Xcap phantoms increased with increasing concentrations of barium (Fig. 2C). In vitro phantom studies revealed that as few as 10 Xcaps (10% barium) could be seen on x-ray fluoroscopic images (Fig. 3A). Using C-arm CT, individual microcapsules separated by 2 mm could be discerned, whereas capsules separated by 1 mm appeared as a single radiopacity (Fig. 3B).
Qualitatively, MSC-Xcaps and empty Xcaps were clearly visible with digital radiography, whereas injection sites with naked MSCs could not be visualized (Fig. 4A–4D). Furthermore, Xcaps (either with or without MSCs) remained visible with digital radiography 14 days after injection (Fig. 4E, 4F).
Quantitative analysis of Xcap visibility on digital radiographs immediately after implantation demonstrated similar CNR between MSC-Xcaps and empty Xcaps (14.0 ± 2.5 vs. 10.1 ± 4.3, respectively, P = NS) with no significant change in CNR at day 14 in either group (13.4 ± 2 vs. 12.8 ± 3.5, respectively, P = non significant (NS)).
All animals demonstrated angiographic occlusion of the SFA within 15 minutes of coil embolization. The arteriogenic benefit of MSC-Xcaps was assessed by a modification of a traditional TIMI frame count , which reflects the development of conductance collateral vasculature . A lower TIMI frame count, that is, enhanced perfusion in the left calf, was seen in rabbits treated with MSC-Xcaps compared to empty Xcaps (4.2 ± 0.8 vs. 12.2 ± 5, p < .001), unencapsulated MSCs (9.0 ± 1.9, p < .02), or sham injections (8.0 ± 2.7, p < .02, Fig. 5A). No statistically significant difference in TIMI frame count was observed between the sham injections and unencapsulated MSCs, for example, naked MSC groups. Thus, based on TIMI frame counts, MSC-Xcaps enhanced perfusion to the lower calf by 65% versus empty microcapsules.
The number of collateral vessels crossing the medial thigh did not differ between treatment arms (4.8 ± 1 MSC-Xcaps vs. 5.8 ± 2 empty Xcaps vs. 4 ± 1 unencapsulated MSCs, P = NS). However, the radiopaque Xcaps somewhat obscured visualization of small collateral vessels on digital subtraction angiography in the medial thigh (Fig. 5B, 5C). Nonetheless, one could appreciate qualitatively more robust collateral vasculature in the animals that received MSC-Xcaps relative to Xcaps without cells (Fig. 5D–5G). On the other hand, no difference in diameter of the pre-existing DDFA could be shown between treatment groups (Fig. 5H).
Immunohistochemical staining for CD31 expressed by vascular endothelial cells demonstrated a proangiogenic effect, that is, robust functional neovascularization around the MSC-Xcaps with tubular structures that were absent in animals receiving Xcaps without MSCs (Fig. 6A–6C). Quantitative analysis of anti-CD31 staining revealed significantly increased capillary density in animals that received MSC-Xcaps versus empty Xcaps (16.96% ± 0.07% vs. 8.48% ± 0.04%, p < .002, Fig. 6D). Capillary density was significantly lower in animals receiving naked MSCs or sham injections (Fig. 6D). Shorter modified TIMI frame counts, that is, more rapid filling in the ischemic limb, were correlated with higher capillary density around microcapsules (y = −81.4x + 18.6, R2 = 0.64, Fig. 6E). On the other hand, because capillary density in naked MSC-treated and sham injection-treated rabbits was low, there was no correlation between capillary density and the modified TIMI frame counts (y = 41.0x + 7.4, R2 = 0.01, Fig. 6F). Arteriogenesis, based on anti-SMA staining (Fig. 7A–7C), was enhanced in animals treated with MSC-Xcaps (5.1 ± 2.7 vessels) relative to sham injections (2.1 ± 1.8 vessels, p < .001), naked MSCs (2.1 ± 1.9, p < .03), or empty Xcaps (3 ± 2.9 vessels, p < .01).
No adverse events or hematologic abnormalities were seen in the rabbits injected in the medial thigh with Xcaps. Inflammatory infiltrate was minimal in animals that received naked, allogeneic MSCs with little evidence of MSC survival at 2 weeks (Fig. 7D, 7E). Macroscopic gross examination of the medial thigh of animals implanted with MSC-Xcaps or empty Xcaps showed clusters of white microcapsules without obvious fibrosis (Fig. 7F). Histological analysis of H&E-stained sections showed no indication of a foreign body reaction from allogeneic MSCs whether encapsulated or not (Fig. 7D, 7E, 7G, 7H). In addition, H&E staining demonstrated patent capillaries near MSC-Xcaps (Fig. 7G, 7H). Blinded scoring by two observers for the presence of inflammatory infiltrates showed no significant difference in number and type of inflammatory cells between MSC-Xcaps and empty Xcaps (P = NS, Fisher’s exact test). While there was no method to blind the observers to the presence of capsules on the histological sections, the inflammatory infiltrate score was lower in animals that received unencapsulated MSCs (rank 0–1) than animals receiving MSC-Xcaps or empty Xcaps (rank 1–3). However, capsule integrity at 2 weeks after administration was demonstrated in all examined histopathological sections.
These results present the first step toward development of an x-ray-visible cellular therapy for the treatment of PAD. While most cellular clinical PAD trials have used autologous cells and a recent meta-analysis of these intramuscular PAD cell therapy trials has shown potential efficacy,  an increasing body of evidence suggests that patients with cardiovascular disease often have native cells that are impaired [27, 28]. Thus, allogeneic cellular therapy from healthy donors may be preferred. In this study, an x-ray-visible cellular therapy comprises allogeneic bone marrow-derived MSCs encapsulated in alginate impregnated with barium sulfate demonstrated high MSC viability with a high sensitivity for detection.
MSC-Xcaps improved hind limb perfusion in an endovascular rabbit model of PAD with easy visualization of the therapeutic at the time of implantation and up to day 14 using standard x-ray clinical imaging suites. MSC-Xcaps improved hind limb perfusion through both arteriogenic and angiogenic mechanisms. Whereas arteriogenesis is the development of conductance vessels through the remodeling of pre-existing arterioles resulting in an increase in luminal area, this is a different process than angiogenesis. Angiogenesis is the sprouting of new capillaries from pre-existing capillaries. Angiogenesis is important for local tissue oxygen delivery but cannot substitute for a proximal obstruction in a conductance vessel. In this endovascular model of hind limb ischemia, the SFA, the main conductance vessel to the calf, was occluded. The lower modified TIMI frame count in the MSC-Xcap group, that is, the more rapid contrast opacification of the conductance calf vessels in the MSC-Xcap group compared to the control groups, demonstrates the arteriogenic effect of this therapy. The latter was confirmed with almost a twofold increase in the number of small arterioles in the MSC-Xcap group relative to sham or naked cell injections. Interestingly, we were unable to detect a difference in either the size of the DDFA or the number of collateral vessels between groups by angiography to explain the increased conductance blood flow. We hypothesize that the remodeled arterioles in the MSC-Xcap group are larger than in the control group. However, accurate measurement of these small diameter vessels is difficult on digital subtraction angiography as the vessels approach the spatial resolution of the images or on post-mortem sections due to artifacts due to freezing. Furthermore, while Xcaps have the desirable effect of visibility on x-ray imaging, they do obscure the detection of small neovessels. Nonetheless, in addition to the MSC-Xcaps’ proarteriogenic effect, this group also demonstrated a proangiogenic effect of increased capillary density.
In ischemic hind limb experiments, MSCs have been studied extensively. The precise role that MSCs have in promoting arteriogenesis is unclear. There is debate about whether the administered cells have the “leading role” marked by incorporation and/or fusion within the enlarging vessel wall or whether they function in a “supporting role” to produce multiple cytokines in the perivascular tissue in order to enlarge the blood vessel [29–31]. MSCs are known to act in a paracrine fashion through expression of a variety of arteriogenic cytokines: vascular endothelial growth factor; fibroblastic growth factor; placental growth factor; monocyte chemoattractant protein; insulin-like growth factor; interleukins, and so forth [30, 32]. Additionally, MSCs have the potential to differentiate into components of the vascular wall [11, 12, 29, 33–36]. Since there was no microscopic evidence of capsule rupture and capillary density adjacent to the MSC-Xcaps was increased, the MSCs within the XCaps were unable to directly incorporate into new vessels. As a result, the mechanism of enhanced hind limb perfusion in MSC-XCap-treated animals appears to be through paracrine mechanisms, exclusively. Unfortunately, determining the exact profile of minute quantities of cytokines by exogenously delivered allogeneic MSCs from native proteins would be extremely difficult to perform in vivo.
The “cellular protection” afforded by the Xcaps could explain the improved hind limb perfusion in the MSC-Xcaps group compared to the naked MSCs. Cellular encapsulation isolates the MSCs immunologically; the pores within the capsules are large enough to permit the diffusion of nutrients into the capsule, but small enough to prevent antibody or immune cell entry . An alternate or complementary explanation may be that MSC microencapsulation enhances cellular survival in vivo via an ischemic preconditioning mechanism . Ischemic preconditioning refers to short bouts of hypoxia that upregulate endogenous mechanisms to protect against future cell death when exposed to lethal low oxygenation conditions. In this study, cell survival after microencapsulation in culture, where ideal oxygenation and nutrient supply are available, was neither enhanced nor impaired. However, other investigators have shown that the MSC administration in an ischemic environment leads to significant death of naked MSCs shortly after administration in part due to the host immune response and/or the poor oxygenation and nutrient supply [14, 38]. Thus, if encapsulation could prevent early destruction, the therapeutic effect of MSCs, via a paracrine mechanism, could potentially be prolonged or enhanced. Practically, prolonging cell survival will have commercial impact as MSCs are a rare component of the bone marrow, that is, 0.01%–0.001%. Therefore, ex vivo expansion, a time and resource intensive process, is required to produce sufficient quantities for a therapeutic dose. By protecting the cells from the immune system, the administered dose could potentially be decreased and, if combined with the utilization of banked allogeneic stem cells, could result in a more cost-effective acute therapy. In this particular study, a positive angiogenic effect was achieved using an administered dose of MSCs in the MSC-Xcap group that was more than 40-fold reduced compared to naked stem cells. In this study, the MSC-Xcap dose was chosen based on a comparable volume to naked MSC injections. Future studies will be needed to determine whether there will be an additional dose response to larger numbers of MSC-Xcaps or whether microencapsulation could provide an off-the-shelf therapy that can be tracked with minimal cell manipulation, unlike gene therapy  or preincubation with proangiogenic cocktails .
Having an x-ray-visible stem cell therapeutic could have direct bearing on patient PAD trials. Because of the large volume of tissue in the human leg, the locations where the cells are administered need to be more precisely targeted than in rodent models with minimal leg muscle mass. Also, unlike rodent models, there are significant differences between the muscle mass/subcutaneous fat ratios in patients, which can affect delivery location. Prior clinical trials have delivered the therapeutic agent via blind intramuscular injections 1.5–2.5 cm deep into the thigh and calf [41, 42]. Because of the differences in body habitus, some patients received the agent deep within the muscle, while others received the agent in the subcutaneous fat. Potential reasons for the disappointing clinical results [43, 44] are poor injection localization and poor cell survival; two obstacles the Xcap approach could help to overcome. In our study, the MSC-Xcaps were clearly visible on clinical imaging systems, both at the time of implantation and on day 14, which would permit the treating physician to confirm accurate delivery and persistence of the therapeutic payload.
The initial MSC-Xcaps biocompatibility profile appears promising. Once the cells were encapsulated, there was neither significant cell death nor proliferation in the MSC-Xcaps over a 30-day period. Furthermore, naked MSCs and MSCXcaps neither demonstrated a significant difference in the inflammatory scores nor adversely affected hematologic parameters. Indeed, MSC-Xcaps in our study had a lower inflammatory score (mean score, <3) than was seen previously with normal wound healing after surgical ligation of the femoral artery in rabbits (mean score, 5.5) . Whereas encapsulated allogeneic islet cells have often resulted in an inflammatory response that leads to suffocation and cell death [45, 46], the native MSC immunotolerance properties [47, 48] may contribute to further enhanced survival and therapeutic effect in Xcaps. In fact, while the total administered dose of MSCs in Xcaps was several fold smaller than naked MSCs, survival of encapsulated MSCs did not elicit a foreign body reaction. Nonetheless, only a short-term follow-up of 2 weeks was performed in this study. Whereas previous work by our group has shown that barium microcapsules remain intact for at least 1 month [17, 20] and given the positive results of this study, long-term biocompatibility and stability of barium sulfate-impregnated alginate microcapsules beyond 2–4 weeks would be warranted.
The surgical and endovascular animal PAD models, like most animal models, have limitations. Both involve a relatively acute occlusive event in a single vessel segment in an otherwise healthy animal. Our endovascular PAD model in the rabbit hind limb, however, has several features that make it superior to surgical models for the study of therapeutic arteriogenesis. The endovascular model benefits are: (a) like atherosclerosis, vessel occlusion is achieved from within the lumen; (b) there is no hind limb surgical wound-healing response to induce local stem cell recruitment ; (c) pre-existing collateral vessels in the thigh are not disrupted as occurs with surgical dissection; and (d) there is no postoperative pain in the treated limb to impair mobility. As such, endovascular occlusion causes significantly less disturbance to the surrounding tissue and a minimal wound-healing response thereby better resembling the pathophysiology of PAD in patients .
Thus, our biocompatible, alginate-barium MSC encapsulation process improves hind limb perfusion in an endovascular model of PAD using clinical x-ray imaging systems akin to those used in the treatment of PAD patients. Consequently, a potential cell-based PAD therapy has been developed that will allow the physician to precisely monitor delivery of stem cells in a clinically germane manner.
We thank Dr. Nicole Azene and Dr. Veronica Crisostomo for assistance with the animals and angiographic vessel size analysis and Judith Cook and Ellen Tully for assistance with immunohistopathology. This study was funded by NIH R01-HL073223, R21/R33 HL0089029, and K08 EB004348. Platinum coils were provided in part by Boston Scientific Corporation.
Author contributions: D.A.K.: collection and assembly of data, data analysis, manuscript writing, and final approval of manuscript; L.H. and D.L.K.: conception and design, financial support, provision of study material, collection and assembly of data, data analysis and interpretation, manuscript writing, and final approval of manuscript; Y.F. and W.G.: collection and assembly of data, data analysis and interpretation, manuscript writing, final approval of manuscript; K.C.: Collection and assembly of data, data analysis, and final approval of manuscript; B.K.: administrative support, collection and assembly of data, data analysis, and final approval of manuscript; B.B.: conception and design, manuscript writing, and final approval of manuscript; B.S.: data analysis and interpretation and final approval of manuscript; P.W. and K.G.: data analysis and interpretation, manuscript writing, and final approval of manuscript; J.B.: conception and design, provision of study material, manuscript writing, and final approval of manuscript. D.A.K. and L.V.H. contributed equally to this article.
Disclosure of Potential Conflicts of Interest
Dr. Kraitchman receives grant support from Siemens Healthcare, and she is a consultant for Surefire Medical, Inc. Dr. Gilson is a Siemens employee.
See www.StemCells.com for supporting information available online.