Evaluation of the Stromal Vascular Fraction of Adipose Tissue as the Basis for a Stem Cell-Based Tissue Engineered Vascular Graft

Abstract

Objective

One of the rate-limiting barriers within the field of vascular tissue engineering is the lengthy fabrication time associated with expanding appropriate cell types in culture. One particularly attractive cell type for this purpose is the adipose-derived mesenchymal stem cell (AD-MSC), which is abundant and easily harvested from liposuction procedures. However, even this cell type has its drawbacks including the required culture period for expansion which could pose risks of cellular transformation or contamination. Eliminating culture entirely would be ideal to avoid these concerns. In this study we utilized the raw population of cells obtained after digestion of human liposuction aspirates – known as the stromal vascular fraction (SVF) – as an abundant, culture-free cell source for tissue engineered vascular grafts (TEVG).

Methods

SVF cells and donor-paired cultured AD-MSCs were first assessed for their abilities to differentiate into vascular smooth muscle cells (SMCs) after angiotensin II stimulation and to secrete factors (e.g. conditioned media) that promote SMC migration. Next, both cell types were incorporated into TEVG scaffolds, implanted as an aortic graft in a Lewis rat model, and assessed for their patency and composition.

Results

In general, cells from human SVF were able to perform the same functions as AD-MSCs isolated from the same donor via culture expansion. Specifically, cells within the SVF performed two important functions, namely, they were able to differentiate into SMCs (SVF calponin expression: 16.4% ± 7.7 vs. AD-MSC: 19.9% ± 1.7) and could secrete pro-migratory factors (SVF migration rate relative to control: 3.1 ± 0.3 vs. AD-MSC: 2.5 ± 0.5). Additionally, SVF was also capable of being seeded within biodegradable, elastomeric, porous scaffolds that, when implanted in vivo for 8 weeks, generated patent TEVGs (SVF: 83% patency vs. AD-MSC: 100% patency) populated with primary vascular components (e.g. SMCs, endothelial cells, collagen, and elastin).

Conclusion

Human adipose tissue can be utilized as a culture-free cell source to create TEVGs, laying the groundwork for the rapid production of cell-seeded grafts.

1. INTRODUCTION

Tissue engineers have developed small diameter (<6 mm) vascular grafts which possess both reduced intimal hyperplasia and thrombosis compared to current clinical standards, representing excellent progress towards clinical application^1-5. However, despite the significant degree of pre-clinical testing that cell-based tissue engineered vascular grafts (TEVGs) have undergone³, few approaches have reached clinical trials^6-8. This is likely attributed to a number of practical rate-limiting barriers still present which hinder the clinical translation of TEVGs, such as appropriate testing of patient-specific cells (e.g. clinically appropriate demographics)⁹, lack of feasible and consistent modes of manufacture to appropriate sizes, the need for culture expansion of cells to a suitable number, and lengthy fabrication time¹⁰. While the first two of these concerns have been validated in recent years^{3, 9}, the necessity of in vitro culture and lengthy fabrication times which are employed by many TEVG designs³ are still major concerns. Specific concerns for clinical application include excess waiting time for the patient, significant costs (such as cell culture reagents and personnel), and potential cellular transformation or contamination.

In order to successfully translate cell-based TEVGs to the clinic, it is important to identify a tissue source which can provide a high number of quality cells in the lowest amount of time. One of the most attractive tissues for this purpose is fat, which can be obtained in abundance from liposuction procedures and provides human adipose-derived mesenchymal stem cells (AD-MSCs). However, even AD-MSCs require time for vitro expansion into necessary numbers¹¹. Alternatively, the digested liposuction population of cells used to obtain AD-MSCs, known as the stromal vascular fraction (SVF), could be utilized as it is progenitor rich, containing AD-MSCs, pericytes, endothelial progenitor cells, and endothelial cells¹². As all of these cell types have been successfully used in vascular tissue engineering³, SVF could hold significant promise as a cell source. Additionally, typical liposuction volumes are sufficient to obtain large enough quantities of cells for construction of a graft of clinically relevant dimensions without additional in vitro expansion^{13, 14}. Utilizing SVF to seed scaffolds for vascular tissue engineering could therefore be ideal for the clinical translation of TEVGs, relieve many regulatory and financial concerns, and make a significant step forward in the fabrication of stem cell-based TEVGs.

In this study, the use of human SVF was validated for vascular tissue engineering using both in vitro and in vivo studies. This initial small-scale study showed that SVF can be used to produce rapidly fabricated TEVGs with robust in vivo performance functionally similar to TEVGs composed of culture expanded AD-MSCs generated from the same human donors.

2. METHODS

2.1 Isolation of Human SVF and AD-MSCs

Adipose-derived cell fractions were isolated from non-diabetic human patients using previously described methods^{12, 15}. Informed consent was not required due to the University of Pittsburgh IRB determination that the study involved solely deidentified waste from the liposuction procedure. Under this type of protocol determination only the presented information (age, gender, BMI, and diabetic status) could be provided (Table I). Liposuction yields were on the order of several kg, from which 200 mL was removed and dissociated via mincing and digestion with collagenase. The resulting solution was filtered to remove debris, centrifuged to eliminate buoyant adipocytes, resuspended in an erythrocyte lysis buffer to eliminate red blood cells, and centrifuged to obtain a final cell pellet. A typical cell yield from this procedure was 30-40 million cells. These cells will be referred to in this manuscript as SVF. For this study, a cultured AD-MSC control was established from each SVF donor by passage expanding a portion of the SVF four times^{9, 15}. Using this procedure, SVF was available for graft fabrication within a matter of hours whereas AD-MSCs took upwards of 4 weeks. Cell culture was performed in 75-cm² tissue culture flasks (Corning) under defined culture media^{9, 15} and was replenished every 2-3 days. To passage expand cells, Trypsin-EDTA (#25200-056; Gibco) was utilized. In addition, for in vitro experimentation SMCs purchased from ATCC were cultured as previously described¹⁵ using their own defined media.

Table 1.

AD-MSC Human Donor Information.

Human Donors Used for Cell Seeding Experiments
ID Number	Gender	Age	BMI
1	F	38	33
2	F	33	28
3	F	43	26

Human Donors Used for Differentiation, Migration, and Implantation Experiments
ID Number	Gender	Age	BMI
4	F	44	23
5	F	26	33
6	F	40	26
7	F	38	35

2.2 Smooth Muscle Cell Differentiation and Migration Assays

Two relevant functions of stem cells in the context of TEVG development are differentiation into vascular SMC^{3, 15-18} and secretion of factors that can promote SMC migration^{2, 15, 19-21}. To induce an SMC-like differentiation of SVF, a previously described method¹⁵ was utilized where cells were stimulated with angiotensin II (1 μM in culture media) for 4 days. As this type of analysis was previously presented in a more robust manner when characterizing multiple populations of AD-MSCs¹⁵ this data is presented in Supplemental Figure 1. In this figure, only staining of calponin expression is presented as markers of terminal differentiation (e.g. myosin heavy chain, smoothelin) do not yet appear with 4 days of angiotensin II treatment, and the early marker smooth muscle alpha actin is ubiquitously expressed in AD-MSCs prior to differentiation. Donor-paired cultured AD-MSCs were utilized as a positive control for differentiation and non-stimulated SVF were used as a negative control. Differentiation was assessed by calponin expression and a morphological shift to a typical SMC spindle-like shape. Quantification of this morphological shift was obtained by measuring each cell's shape factor (Shape factor = 4*pi*area/perimeter², ~0 = ellipsoid, ~1 = circular) wherein differentiation is reflected by conversion from a round (nearer to 1) to spindle-like (nearer to 0) shape.

Stimulation of SMC migration also was performed utilizing previous protocols¹⁵. Briefly, SMCs were plated near confluence in a 24-well plate and a scratch wound was made via a single stroke of a pipette tip per well. During migration SMCs were stimulated with conditioned media from either SVF, donor-matched AD-MSCs, or non-conditioned media. All cells were maintained under culture conditions (37°C, 5% CO₂) for the duration of the experiment. The extent of migration was monitored every 2 hours and was reported as the average rate of scratch wound closure over the first 8 hours. Three donor lines were used in these studies (Table I).

2.3 Scaffold Fabrication

Small (1.3 mm ID, 10 mm length) bilayered, tubular, biodegradable, elastomeric scaffolds consistent with the size of the rat abdominal aorta were constructed from poly(ester urethane)urea (PEUU) as previously described²². Briefly, a porous inner layer was created via thermally induced phase separation, which was then coated with an external mechanical sheath via electrospinning. This scaffold has already shown efficacy in the TEVG context using multiple cell types^{1, 4, 9, 23}.

2.4 Cell Seeding

A rotational vacuum seeding device (RVSD) technology was used to seed PEUU scaffolds as previously described^{1, 4, 23, 24}. Briefly, scaffolds were mounted within the device, infused lumenally with a cell suspension, and exposed to both rotation (15 rpm) and vacuum (−127 mmHg). A total of 3 million cells (either SVF or donor-matched AD-MSCs) were incorporated within each scaffold utilizing this process. Post-seeding, the resulting constructs were incubated in static media for 4 hours to allow for cell adhesion after which they were subjected to a dynamic culture (spinner flask, 15 rpm) for 48 hours which primed them for either seeding analysis or implantation. For analysis of uniform seeding, constructs were fixed (4% paraformaldehyde; 30 min), cut into three axial portions, and frozen. Scaffolds were then sectioned and stained with DAPI to analyze the cell distribution. Images were acquired at 10x using NIS Elements software and were analyzed using custom scripts built within ImageJ (NIH, Bethesda, Maryland)⁹ wherein complete cross sectional images were generated, virtually segmented into radial or circumferential pieces, and cell densities were compared with respect to radial, circumferential, or longitudinal pieces.

2.5 In Vivo Implantation and Explant

Seeded scaffolds were implanted as end-to-end abdominal aortic grafts using our Lewis rat model^{1, 4, 9, 23}. All 4 patients noted in Table I under “Human Donors Used for Differentiation, Migration, and Implantation Experiments” were used as the source of cells to produce, with some replicates, n=6 SVF and n=5 AD-MSC grafts for implantation. Rats were anesthetized and the infrarenal abdominal aorta was exposed. Microclamps were applied to the aorta prior to transection upon which a seeded scaffold was sutured with 10-0 prolene. After securing the graft, the clamps were released and the patency verified by observation of the distal pulse pressure. The animals were then closed with 3-0 polyglactin sutures and maintained on an anti-coagulant (dipyridamole, aspirin) schedule for 4 weeks as previously described⁹. After 8 weeks, the rats were euthanized and an angiogram was performed to assess TEVG patency. All procedures were performed under an approved Institutional Animal Care and Use Committee protocol. TEVGs were then explanted, fixed, and frozen for histologic analysis.

2.6 Immunofluorescence and Histologic Evaluation

Frozen TEVGs were sectioned using a cryo-microtome at an 8 μm thickness prior to being mounted on gelatin coated slides and stained using standard indirect immunofluorescent chemistry. Primary antibodies utilized were as follows: von Willebrand Factor (vWF) [1:250; US Biological #V2700-07], smooth muscle alpha-actin (SMA) [1:1000; Sigma #A5228], calponin [1:250; Abcam #ab46794], and elastin [1:100; EPC #RA75]. The following secondary antibodies were utilized: [Rockland #611-1202 (1:1000)], [Invitrogen #A10521 (1:1000)], [Sigma #C2821 (1:300)]. Explants were also stained with H&E. All images were taken using a 20x objective with an epifluorescent microscope utilizing NIS Elements software. Analysis was restricted to the newly developed luminal tissue within TEVGs (termed as “neotissue”).

2.7 Statistics

All statistical analyses were performed in Minitab (version 16) software. Data was first assessed for assumptions of a normal distribution and homogeneity of variance. Following this, a student's t-test was performed with significance being defined at p < 0.05. Data was expressed as mean ± standard deviation.

3. RESULTS

3.1 Human SVF differentiated into SMC-like cells and secreted factors to promote SMC migration

To initially investigate the relevance of SVF for TEVG applications, the abilities of SVF to differentiate into SMC and to induce SMC migration were analyzed. Both of these stem cell functions are important for achieving maturation of a TEVG into a native-like vessel¹⁵. Both donor-matched AD-MSCs – a cell population previously validated in this manner¹⁵ – and SVF underwent differentiation into SMC like-cells and secreted factors to promote SMC migration. For SMC differentiation this was shown by induced expression of the SMC marker calponin (Supplemental Figure 1A) and a morphologic change to a more spindle-like cell shape (e.g. a decreased shape factor) (Supplemental Figure 1B). It is important to note that even though the broad percentage of differentiated cells was low, discernable donor differences can be detected with this assay¹⁵. The ability of both AD-MSCs and SVF to promote migration of SMCs was evident by the significant increase in migration rate observed in SMCs when stimulated with conditioned media from either cell type (Supplemental Figure 1C). Taken together these results verify that SVF has functions present in AD-MSCs necessary for TEVG maturation and thus can be utilized in the context of vascular tissue engineering.

3.2 Human SVF was able to be seeded within tissue engineering scaffolds and generated robust, patent vessels in vivo

To verify that SVF could be seeded within our porous tubular scaffolds and as a point of quality control, the extent to which SVF was incorporated within our scaffolds using our RVSD was quantified. SVF was evenly seeded throughout the scaffolds in cross section (Figure 1A) and equivalent cell densities were observed across the radial, circumferential, and longitudinal directions (Figure 1B, radial shown). This is in line with previously published work analyzing AD-MSC seeded scaffolds⁹. SVF seeded scaffolds were then implanted in a Lewis rat model where after 8 weeks in vivo they remained patent, as confirmed by angiography, similar to their AD-MSC counterpart grafts (Figure 1C,D). No dilation was observed for any implanted vessel. On gross observation, TEVGs generated from either SVF or AD-MSCs showed an evident tissue-like appearance (Figure 1E). Together these results show that both SVF and AD-MSCs are capable of generating patent, tissue-like TEVGs.

Figure 1. Human SVF can be seeded uniformly within porous tubular scaffolds and generate patent vessels to a similar extent as AD-MSCs.

(A) SVF (n=3) was obtained from human patients and seeded directly into PEUU scaffolds where they seeded uniformly (blue, DAPI stain) throughout the thickness of the scaffold wall. A representative image is shown. (B) This uniformity was quantitatively confirmed by comparing radial distribution densities of cells across three radial regions, which each contained ~33% of the total cells. This is in line with our previous work quantitatively showing uniformity of AD-MSC seeded scaffolds¹³. Data is presented as mean ± SD with a statistically significant difference defined at P < .05. (C) Both SVF and donor-matched AD-MSCs were capable of generating patent TEVGs based on patency rates. (D) Seeded scaffolds with either SVF or donor-matched AD-MSCs were implanted in a Lewis rat model wherein they maintained angiographic patency through an 8 week timepoint. The TEVG location is indicated. (E) All explanted TEVGs displayed a clear tissue-like gross appearance, compared to the unremodeled PEUU scaffold.

3.3 SVF develop vascular-like tissue within TEVGs that is rich in primary vascular components

To analyze the composition of the newly developed tissue within the TEVGs, the primary cellular (SMC, endothelial cells), and extracellular (elastin, collagen) vascular components were detected using multiple imaging modalities. Immunofluorescence (Figure 2) revealed that the cellular components of all explanted TEVGs – either SVF or AD-MSCs – primarily consisted of SMCs (SMA and calponin labeling) with a luminal lining of endothelial cells (vWF labeling). Immunofluorescence also confirmed the presence of elastin in both SVF and AD-MSC-based TEVGs, however, it had not yet formed into fully-developed elastic lamellae. Finally, to detect the presence of the extracellular matrix component collagen within explanted TEVGs multiphoton microscopy was employed. This imaging modality revealed prominent, circumferentially aligned collagen fibers within all TEVGs explants (Figure 2). Taken together, these results show that when SVF is utilized as the cell source for TEVGs a vascular-like composition emerges similar to that observed with donor-matched AD-MSCs.

Figure 2. Human SVF-based TEVGs developed into a vascular-like tissue containing the primary cellular and matrix constituents of a small diameter artery, similar to that observed with AD-MSC-based TEVGs.

TEVGs, using either SVF or donor-matched AD-MSCs displayed significant remodeling. TEVGs stained positively for SMCs (smooth muscle actin and calponin), endothelial cells (von Willebrand Factor), and elastin. Multiphoton imaging demonstrated the presence of collagen. White arrow indicates lumen. The dashed line represents the boundary between newly-developed vascular-like tissue (neotissue) and the PEUU scaffold. For comparison, native artery images are included from¹³.

4. DISCUSSION

The results of this study show that SVF can perform the same functions as paired AD-MSCs derived from the same donor for use in a TEVG. First, SVF was capable of two stem cell-based bioactivities thought to be of possible importance in vascular tissue engineering: differentiation into vascular SMCs^{3, 15-18} and secretion of pro-migratory factors^{2, 15, 19-21} (Supplemental Figure 1). Second, SVF was shown to be incorporated within porous tubular, biodegradable, elastomeric scaffolds and generate patent TEVGs (Figure 1) with vascular-like tissue after implantation (Figure 2). The results of this study indicate the functional efficacy of a freshly-isolated non-cultured adipose-derived cell source in the context of vascular tissue engineering, favorably comparable to published work with cultured AD-MSCs⁹. Ultimately, the use of this cell source reduces rate-limiting barriers that could hinder the clinical translation of TEVGs. Indeed, SVF has previously been shown to perform the same functions as AD-MSCs in other applications^{25, 26} such as Crohn's disease and myocardial infarction which has supported the use of SVF in multiple clinical trials^27-29.

While non-cultured bone marrow cells have already been employed as a non-cultured cell source for vascular tissue engineering³⁰ and can generate an equally sizable amount of cells³⁰ as adipose tissue^{13, 14} (on the order of millions per mL), the use of an adipose-derived cell source confers unique benefits. In addition to the aesthetic attractiveness of harvesting adipose tissue to obtain SVF, performing a liposuction on a patient is associated with a reduction in cardiovascular disease risk factors³¹. This could offer resistance against recurrent vascular disease. On a cellular level, freshly isolated adipose cells (i.e. SVF) contain a higher percentage of cells that can provide anti-thrombogenic effects, compared to bone marrow, which could affect graft patency. These include cells such as MSCs (30% adipose^{12, 27} vs 0.001% bone marrow^{19, 27}), endothelial cells (3% adipose¹² vs 0.1% bone marrow¹⁹), and endothelial precursor cells (14% adipose¹² vs. 2% bone marrow¹⁹). However, this statement remains somewhat conjecture as a side-by-side comparison of SVF vs. fresh bone marrow has yet to be performed in vascular tissue engineering.

One of the major challenges currently present in stem cell-based vascular engineering is fabrication time. To reach clinical trials, tissue engineered products face a number of hurdles³² including extensive pre-clinical testing, complicated and uncertain regulatory pathways, the necessity to improve product practicality, and funding shortages. These are amplified by the landscape of the current market where substantial reimbursement for cell-based technologies does not yet exist¹⁰. Additionally, regulations and recommendations set forth by the Food and Drug Administration (FDA)^{33, 34} encourage the use of fresh cells as well as approaches that minimize time between isolation and use of a patient's cells for clinical trials. Indeed one of the few literature reports of a tissue engineered vascular graft clinical trial utilized a non-cultured cell population (bone marrow mono-nuclear cells)^{6, 8}. The FDA also encourages controllable methods^{33, 35} of manufacture which can be accomplished with cell processing devices. Utilizing our automated bulk cell seeding device in conjunction with SVF makes a significant practical step in the design and fabrication of stem cell-based TEVGs. However, manual methods and xenogenic enzymes are still required to isolate these cells^{11, 12, 15}. Automated technologies³⁶ are currently being tested to alleviate these concerns and technologies which can isolate SVF without enzymes are also being considered³⁷. It is foreseeable that with the combination of these technologies, SVF, and our cell seeding methods, a TEVG could be fabricated at bed-side and within the same surgical procedure.

While this study acts as a small-scale feasibility investigation to the use of culture-free adipose-derived cells (e.g. SVF) to enhance improve upon TEVG fabrication methods, there are still many unaddressed areas that require further investigation. Specifically, multiple questions remain unanswered such as 1) what are the phenotypic differences between SVF and AD-MSCs?, 2) how do these differences contribute to the overall maturation of TEVGs in vivo?, and 3) what are the long term outcomes of these grafts (e.g. at time points greater than 8 weeks)? Each of these areas are of current interest for future study, for example the pro-elastogenic effect of AD-MSC is relevant to another vascular context, abdominal aortic aneurysm³⁸. Additionally, while this proof-of-concept study was performed with immuno-tolerant rats implanted with human cells, large scale studies will need to be performed to fully understand how this graft will function in models with more similar biology and geometry to that of humans.

5. CONCLUSION

A non-cultured human cell source (i.e. SVF) can be isolated from adipose tissue and utilized in vascular tissue engineering similarly to donor-matched culture-expanded AD-MSCs. These cells are capable of essential regenerative functions, can be effectively bulk-seeded using a controllable device, and can generate TEVGs with a vascular-like composition. This study makes a significant practical step in the design and fabrication of stem cell-based TEVGs.

CLINICAL RELEVANCE.

Cells sourced from patient fat after procedures such as liposuction could ultimately be seeded into a tubular scaffold as the basis of a vascular graft for small diameter coronary and peripheral artery bypass procedures. This study serves as a preclinical evaluation of adipose-derived cells, seeded into a tubular scaffold, in the context of a rat interpositional implantation model.

Non-Standard Abbreviations and Acronyms

AD-MSC

Adipose-Derived Mesenchymal Stem Cell

PEUU

Poly(ester urethane)urea

RVSD

Rotational Vacuum Seeding Device

SMA

Smooth Muscle alpha-Actin

SMC

Smooth Muscle Cell

SVF

Stromal Vascular Fraction

TEVG

Tissue Engineered Vascular Graft

vWF

von Willebrand Factor