Engineered tissue vascularization and engraftment depends on host model

Introduction:: Developing vascular networks that integrate with the host circulation and support cells engrafted within engineered tissues remains a key challenge in tissue engineering. Major progress towards vascularizing engineered tissues arose from the discovery that endothelial cells have the capacity to self-assemble into microvessel-like networks in vivo. We and others developed a method to pattern endothelial cells and collagen into arrays of parallel channels and then encapsulate the resultant “endothelial cords” within engineered tissues. Upon implantation in athymic mice, these endothelial cords act as “railroad tracks” that guide the formation of chimeric host-graft vessels that became anastomosed with the host circulation.

Despite this progress in leveraging cooperation between the graft and host for tissue vascularization, most previous work in this field has focused on developing new methods to build human vascular networks within engineered tissues prior to their implant in vivo, with substantively less attention paid to the role of the host in tissue vascularization and engraftment. We hypothesized that different host models (here, animal models) and anatomic implant locations play a role in the vascularization of human engineered tissues upon their implantation in vivo.

Here, we test this hypothesis and report divergent responses in vascularization and cardiomyocyte engraftment of engineered tissues after their implantation in athymic mice versus athymic rats. These results suggest that guided vascularization and cardiomyocyte survival are supported by different host factors, which vary across animal models.


Materials and Methods:: To test this hypothesis, we suspended endothelial cells (HUVECs) and stromal cells in collagen within a PDMS mold with parallel channels to form “endothelial cords” and encapsulated these cords within fibrin tissues. Engineered tissues were prepared containing different cell types in the bulk gel for each experiment as described in Table 1. These tissues were then sutured onto either the intraperitoneal (IP) gonadal fat pad or on the epicardial surface (ES) of the heart in athymic nude mice or athymic rats. These engineered tissues were explanted at either 3-, 7- or 10-days post implantation for histological analyses. Hematoxylin and Eosin (H&E) and Sirius Red (SR) staining was used to visualize tissue morphometry and characterize inflammation within explanted tissues. In addition, we stained for human endothelium (CD31), mouse erythrocytes (TER-119), pericyte marker α-smooth muscle actin (α-SMA), human myocardium with β-myosin heavy chain (β-MHC) and Hoescht.

Table 1: Cell types contained in patches for each experiment.

Host	Implant location	Cords	HUVECs (bulk)	hPSC-CMs (bulk)	NHDFs (bulk)
Mouse	IP + ES	Yes	No	No	No
Mouse	ES	Yes	No	6.67 x 106 /mL	1.67 x 106 /mL
Mouse	ES	Yes	3.33 x 106 /mL	6.67 x 106 /mL	1.67 x 106 /mL
Rat	ES	Yes	3.33 x 106 /mL	6.67 x 106 /mL	1.67 x 106 /mL
Rat	IP + ES	Yes	No	No	No

Results, Conclusions, and Discussions:: We first compared engineered tissues that were implanted in either the abdomen or on the epicardial surface of athymic mice. H&E and SR staining of explanted tissues from both locations revealed regularly spaced clusters of cells in a pattern identifying a collagen core surrounded by 3 days that consolidated to form huCD31+ lumens containing red blood cells (TER-119) by 7 days. Quantification of collagen, fibrin, nuclei (Hoescht), huCD31, and TER-119 showed greater huCD31+ area in the epicardial group compared to IP implants at 3 days. By 7 days this had resolved and there were no significant differences in patch architecture, inflammation, or vascularization between implant locations.

We then implanted fibrin patches containing endothelial cords either in the IP space or the epicardial surface of athymic rats. Upon retrieval of the tissues from the IP location, we were able to identify the location of the implant by visualizing the suture but could not identify any of the engineered tissues in any of the animals. Conversely, all epicardial implants were identified grossly at the time of explant and further histologically processed.

We next compared engraftment of engineered cardiac tissues on the epicardial surface of athymic mouse and rat hearts. Tissues containing cords and human cardiomyocytes were implanted on the hearts of athymic rats or mice. After 10 days, H&E and SR staining showed explanted tissues from mice had evidence of guided vascularization, while tissues recovered from rats had no evidence of cords-associated blood consolidations and >3 fold more collagen than tissues from mice. Cords-containing patches in mice had distinct clusters of huCD31+ lumens that contained TER-119+ mouse blood, whereas the tissues explanted from rats had no discernible patterning of graft-derived cells. Furthermore, β-myosin heavy chain (β-MHC) staining showed that explanted tissues grafts explanted from athymic rats were > 3-fold larger grafts explanted from mice.

Our results suggest that the tissue engineering and cellular therapy fields would benefit from paying more careful attention to understanding and tuning host environment to improve reproducibility, accelerate clinical translation of regenerative therapies, and potentially even greatly enhance the efficacy of regenerative engineering strategies.


Acknowledgements (Optional): :

References (Optional): :