Assessment of cell yield among different devices for endovascular biopsy to harvest vascular endothelial cells

Routine surgical collection of vessels limits pathophysiological understanding of vascular disorders, particularly those of the CNS. This consequently slows development and improvement of diagnostic techniques and therapies. Recently developed and validated endovascular biopsy techniques can address this issue [1–3]. To further refine these techniques, this investigation compares yields of harvested endothelial cells (ECs) among several devices used for endovascular biopsy.

A total of 70 sampling procedures were performed – 38 were performed in rabbits, and 32 were performed in humans. ECs from normal vessels were sampled in 25 cases. ECs were sampled in 24 aneurysms, 13 arteriovenous malformations, four extracranial atherosclerotic lesions, and five intracranial plaques. Mean ECs collected by device type are summarized in Table 1. Table 2 summarizes yields by pathology. No differences were found for species, location, or pathology. Phenox clot retriever devices were more navigable to smaller distal arterial branches (1–3 mm) compared with balloons and retrievable stents, which could be fully deployed in larger intracranial branches (2–6 mm). Phenox clot retriever devices yielded more cells compared with other types (LR 33.3; p < 0.001.) This difference was not found for analysis of ECs/mm² (LR 24.1; p = 0.344). Similarly, retrievable stents yielded more ECs compared with other devices (LR 63.0; p < 0.001), but ECs/mm² analysis showed no significance (LR 31.5; p = 0.344). No other devices demonstrated significant differences in ECs yielded or ECs/mm². Combining analysis of Phenox clot retrievers and retrievable stents against other device types also demonstrated higher EC yield (LR 64.2; p < 0.001).

To better realize the promise of precision medicine, innovative techniques are needed to isolate viable cells for genomic and epigenetic analysis. This need is particularly acute in the case of the intracranial circulation, as pathologic study is seldom achievable for living tissue or cells, given the preclusive morbidity of surgical biopsy. To circumvent these limitations, our group has developed and validated endovascular biopsy techniques to sample ECs from vessel walls that can then be analyzed with single-cell RNA sequencing [4–9].

To further refine these techniques, we sought to compare devices employed for endovascular biopsy. As can be expected, those devices that make the most contact with vessel walls yield the highest number of cells. The exception to this finding was angioplasty balloons. We posit that balloon materials are less adherent than the metal of retrievable stents or filaments of Phenox clot retrievers. Additionally; the radial force exerted by these devices, particularly retrievable stents, likely contributes to their superior performance. While our initial investigation in humans focused on sampling ECs from intracranial aneurysms using detachable coils, these devices prove less efficacious in this study, although they can still yield meaningful results in the appropriate setting, that is, sampling within a saccular aneurysm. As might be expected with their small profile, minimal vessel wall interaction, and hydrophilic coating, wires and catheters yielded fewer ECs.

While this investigation offers insight into maximizing yields for endovascular biopsy, it bears several limitations. The exploratory nature of the investigation precluded standardization of devices and techniques, although, once harvested, all cells underwent the same standardized sorting protocol. Additionally, variability between animal and human tissue, normal and diseased arteries, and the various disease states investigated can also introduce variability and potential bias into the data. Nonetheless, these preliminary results can inform future evaluation of endovascular sampling techniques and should be accounted for in future investigations.

Cell harvesting procedures performed initially in rabbits and subsequently in humans were analyzed according to a previously reported protocol that was approved by the Institutional Animal Care and Use Committee and Institutional Review Board [4,10,11]. Human subjects provided informed consent prior to the procedure. Briefly, a sampling device is placed into the vessel segment of interest using standard endovascular techniques. The device is deployed from its delivery catheter in a way that it only contacts vessel wall at the targeted segment. Withdrawing the device into the delivery catheter, the catheter is then removed from the subject's body so that the device contacts no other vessel segments.

Maintaining sterility, the device is cut with scissors to disconnect the portion of the device that contacted the vessel wall, so that the device drops directly into cell dissociation buffer (Gibco, NY, USA). The sample is vortexed for 10 s and the device removed from the sample. The cells, now suspended in dissociation buffer, are centrifuged at 1500 RPM for 10 min at 4°C. The resulting pellet is re-suspended in 1 ml ACK Lysing Buffer (Gibco) for 5 min to lyse erythrocytes. Nucleated cells remaining in the specimen are then stained with seven fluorescently conjugated monoclonal antibodies for FACS sorting [4–6,11]. CD31, CD34, CD105 and CD146 serve as EC markers, while CD45, CD11b and CD42b identify contaminant leukocytes that can be removed to ensure only ECs are isolated [11]. EC candidates successfully screened are then sorted into 96-well plates using single-cell sort mode [11]. Sample FACS output images are provided in Figure 1.

For the present investigation, EC counts were recorded, as well as species, device utilized for sampling, vessel sampled and disease state interrogated, if any. Diseases present included intracranial aneurysm, intracranial arteriovenous malformation, pulmonary arteriovenous malformation, extracranial atherosclerosis, intracranial atherosclerosis and brain tumor. Locations sampled were classified as cerebral, pulmonary, renal or aorta. Devices from which ECs were sampled included microwires (Transend EX Platinum, CA, USA), angioplasty balloons (Transform Compliant, CA, USA; Aviator, Cardinal Health, OH, USA), stent delivery catheters (Acculink Carotid Stent, Abbott Vascular, IL, USA), detachable coils (Target Coils, Stryker Neurovascular, CA, USA; Axium Coils, Medtronic, CA, USA; Orbit Galaxy Coils, Cerenovus Johnson & Johnson, NJ, USA), Phenox clot retriever (Phenox GmbH, Bochum, Germany), retrievable stents (Solitaire, Medtronic, CA, USA; Trevo, Stryker Neurovascular, CA, USA), and vascular plug devices (MVP Micro Vascular Plug, Medtronic, CA, USA). For devices with roughly cylindrical shape that contacts the vessel wall (balloons, Phenox clot retrievers, retrievable stents), the number of cells harvested per square mm was calculated by the equation, cells/(π*width*length). Chi-square analysis was performed to compare means of ECs harvested between the above-listed variables. Statistical analysis was performed with SPSS version 22 (Armonk, NY, USA), with statistical significance defined as p < 0.05.

Among devices investigated, Phenox clot retrievers and retrievable stents yielded greater quantities of ECs compared with other devices employed for endovascular biopsy. Further study of these devices for EC sampling is warranted to increase and refine its use for this purpose.