Tissue Engineered Vascular Grafts

Corresponding Author

The repair of congenital heart defects often requires the use of synthetic vascular grafts. These artificial conduits present with substantial limitations such as the inherent risk of thrombosis, progressive degeneration, calcification, and importantly, the inability to accommodate growth in children. This often leads to size-mismatch and somatic overgrowth over time, resulting in the need for repeated surgeries. To overcome these limitations, the concept of tissue-engineered vascular grafts (TEVGs) with their unique capacity to transform into neovessels with somatic growth potential has generated substantial hope. Over the past decades, a wide array of TEVG strategies has emerged, ranging from autologous, cell-based concepts to cell-free approaches.¹

One concept to creating “living” implants involves a 1-step seeding approach of autologous cells onto starter matrices immediately before implantation.¹ Building on the pioneering work of cardiac surgery legend DeBakey in the mid-1960s and the seminal contributions of Weinberg and Bell to create the first cell-seeded TEVG in vitro, the concept of autologous cell seeding has evolved rapidly. Numerous iterations have been explored, testing various cell-sources on a wide range of prosthetic and scaffold materials.¹

In 2001, pediatric cardiac surgeon Dr Toshiharu Shinoka pioneered the first successful implantation of a TEVG in a 4-year-old girl undergoing pulmonary artery repair.² This landmark event was followed by a clinical trial in Japan involving 25 children with single-ventricle anomalies undergoing modified Fontan surgery requiring a vascular conduit for cavopulmonary anastomosis.³ The TEVGs were created by perioperative seeding of autologous bone marrow–derived mononuclear cells (BM-MNCs) onto a biodegradable tubular scaffold fabricated from a polyglycolic acid fiber–based mesh coated with a 50:50 copolymer of polycaprolactone and polylactic acid.³ The grafts demonstrated a good safety profile, and showed remodeling into compliant neovessels capable of growth, potentially reducing the need for reoperations. However, TEVG stenosis occurred in 30% of patients during late follow-up, which could be successfully treated with angioplasty.³

After Shinoka’s return to the United States, and ever since, he joined forces with his colleague Dr Christopher Breuer to further develop TEVGs for congenital heart defects. Building on the data from Japan, they advanced their TEVG program into a first U.S. Food and Drug Administration–approved clinical trial, again targeting patients undergoing Fontan surgery (NCT01034007). Advanced imaging revealed an initial period of dynamic TEVG remodeling. However, this remodeling resulted in graft-stenosis 6-months post-surgery in all 4 implanted patients, of whom 3 required angioplasty. Due to this unexpected high rate of early graft stenosis, the trial was prematurely closed.^4,5

The pooled results from both trials identified TEVG stenosis as the main graft-related complication limiting broader clinical adoption of this otherwise promising approach. Shinoka and Breuer were forced from the clinic back to the bench, to elucidate mechanisms underlying the high incidence of TEVG stenosis as a prerequisite for continuation of their clinical program. Driven by these sobering facts and their unwavering commitment to their patients, they initiated an unprecedented research program to investigate the natural TEVG remodeling behavior and to develop new strategies to prevent TEVG stenosis.

In this issue of JACC: Basic to Translational Science, the Breuer-Shinoka team investigated potential factors influencing TEVG remodeling and that may predict graft stenosis.⁶ Using a lamb model with inferior vena cava (IVC) interposition grafts, they analyzed neovessel development in 50 animals and conducted angiography at 1 and 6 weeks postsurgery. Regression analysis identified significant correlations between graft oversizing and inflow narrowing at the anastomosis, and subsequent stenosis at 6 weeks. Computational fluid dynamics demonstrated that these factors altered wall shear stress and flow patterns, contributing to neovessel narrowing. They concluded that hemodynamics play a critical role in neovessel formation after TEVG implantation, and that graft oversizing and inflow narrowing may exacerbate stenosis.

These findings are clinically relevant, especially given that graft oversizing is a common strategy in Fontan surgeries to anticipate somatic growth. In contrast, the present findings suggest that such oversizing may be detrimental and instead support a size-matched approach. This showcases that the safe implementation of novel procedures does not only demand sufficient patient selection and surveillance strategies, but also emphasizes the need for dedicated training of the performing surgeons—possibly supported by new technology-focused, highly specialized “Centers of Excellence”.

However, some methodological limitations warrant discussion as they may have biased results: The retrospective study-design allows only for correlation between surgical factors and the occurrence of TEVG stenosis. A randomized, head-to-head study comparing size-matched vs size-mismatched TEVGs is essential. Moreover, assessment of native IVC size to determine related TEVG size match or mismatch represents a key factor to render meaningful conclusions. Notably, large veins such as the IVC are notoriously difficult to measure caused by their dynamic nature and volume dependency. The authors state that TEVG size was selected based on animal size and weight, which is a rather imprecise measure, and no details on how intraoperative sizing and surgical alignment was performed were provided. Furthermore, angiography was only performed at 1 and 6 weeks, respectively. The absence of preoperative imaging limits interpretation, as the 1-week angiography cannot serve as a true baseline. Moreover, the use of retrograde bolus injection to determine presence and degree of TEVG stenosis inherently bears a risk for measuring errors, especially in the light of the complex interplay between the following: 1) the pliable nature of the native IVC; 2) potential size-mismatch and inflow narrowing; and 3) compliance-mismatch between native IVC and polymer-containing TEVG. Thus, complimentary, noninvasive imaging such as magnetic resonance imaging or echocardiography should be considered for validation. Lastly and most importantly, an in-depth, more precise morphologic characterization of the observed TEVG stenosis, together with histology, would have added significant value.

All in all, Shinoka and Breuer have put tremendous efforts into the characterization of the underlying mechanisms of neovessel formation and the development of TEVG stenosis over the past 2 decades. They have utilized various animal models, a multitude of analytical tools such as computational modeling^4,7 and computational fluid dynamics,⁶ paired with a wide range of mechanistic and molecular investigations. Based on this extraordinary research-machinery, a plethora of mechanistic and translational insights was generated with the main aim to systematically tackle the remaining clinical challenge of TEVG stenosis.

For instance, their computational modeling predicted that TEVG stenosis may resolve spontaneously. A striking finding that, in retrospect, might have influenced the decision to close the U.S. trial prematurely.⁴ Additionally, angiotensin-II receptor blockers were introduced,⁸ emulating the previously described chronic treatment with angiotensin-converting enzyme inhibitors targeting intimal-hyperplasia induced stenosis.⁹ This suggest that, beyond size and/or compliance mismatch (relative stenosis), true lumen narrowing caused by similar phenomena plays a role in TEVG stenosis.

The authors characterize their frequently observed TEVG stenosis by an inflammatory process, driven by a foreign-body response to the polymer, followed by a mechano-mediated process. Although their data support this notion to a large extent,^4,5,7 it however also raises the critical question of whether the seeded cells themselves might contribute to this process. Although BM-MNCs are a well-established cell source for tissue engineering, their ambiguous ability to act as a “friend or foe” has also been shown, particularly when it comes to the delicate interplay with the particular polymer used.¹ A possible role of the BM-MNCs in the pathogenesis of TEVG stenosis should therefore be carefully (re)considered. This appears particularly relevant in light of the recent realization by the Breuer group that “cell seeding is not essential for neovessel-formation, though it modulates outcomes.”⁷ A direct comparison of BM-MNC seeded vs unseeded TEVGs would be essential to determine the true need and benefit of cell seeding.

Despite the magnitude of available mechanistic and preclinical TEVG data, the direct transferability of all these findings into the complex pathophysiology and hemodynamic environment of Fontan patients remains limited. Besides the highly variable single-ventricle pathology, these patients ultimately experience multiorgan disease involving lungs and liver. In contrast, the IVC lamb model represents a healthy animal with a normal heart and physiology, in which the TEVG simply serves as a 2-cm interposition graft. From that perspective, the “TEVG Fontan case” exemplifies the translational gap that often exists between controlled preclinical models and real-world clinical complexity.¹⁰

To this end, also the temporal discrepancy in the onset of TEVG stenosis occurring after many years in the Japanese trial, 6 months in the U.S. trial, and 6 weeks in the preclinical IVC model remains unexplained. It triggers the critical question of whether the underlying mechanisms causing TEVG stenosis in the animals are indeed the same as the ones in Fontan patients. Notably, echocardiography-based assessment demonstrated initial TEVG dilation after 2 months followed by stenosis after 7 months in patients from the prematurely closed U.S. trial.⁵ This biphasic remodeling behavior in humans substantially contrasts with the early TEVG narrowing observed in animals.^4,5,7

Although the key issue of TEVG stenosis does remain, the team seems to be prepared to take this challenge. The significant insights from their translational research machinery and the identification of potential pharmacological mitigation strategies have paved the way for a second-generation TEVG trial (NCT04467671), which is currently underway. The results are eagerly awaited with the hope that the critical remaining issue of TEVG stenosis in an otherwise promising approach will be resolved.

Although only time can tell, in any case, the team deserves recognition for their remarkable path—from clinic to bench, and back again. As Shinoka and Breuer continue writing the next chapter in the TEVG story, they leave behind an extraordinary journey of courage, challenge, and importantly, persistence over one-quarter of a century.

Funding support and author disclosures

Dr Zilla is a co-founder of the University of Cape Town Startup Company Strait Access Technologies. Dr Hoerstrup is a shareholder at Xeltis BV and LifeMatrix Technologies AG. Dr Emmert is a shareholder at LifeMatrix Technologies AG.