Rapid fabrication of reinforced and cell-laden vascular grafts structurally inspired by human coronary arteries

Design strategies for small diameter vascular grafts are converging toward native-inspired tissue engineered grafts. A new automated technology is presented that combines a dip-spinning methodology for depositioning concentric cell-laden hydrogel layers, with an adapted solution blow spinning (SBS) device for intercalated placement of aligned reinforcement nanofibres. This additive manufacture approach allows the assembly of bio-inspired structural configurations of concentric cell patterns with fibres at specific angles and wavy arrangements. The middle and outer layers were tuned to structurally mimic the media and adventitia layers of native arteries, enabling the fabrication of small bore grafts that exhibit the J-shape mechanical response and compliance of human coronary arteries. This scalable automated system can fabricate cellularized multilayer grafts within 30 min. Grafts were evaluated by hemocompatibility studies and a preliminary in vivo carotid rabbit model. The dip-spinning-SBS technology generates constructs with native mechanical properties and cell-derived biological activities, critical for clinical bypass applications.

1. Line 123: the statement is made that the GEAL solution is pre-crosslinked, but then the authors state that the dipped layer is exposed to UV light to crosslink the material. please clarify. 2. The phrase "circumferential axis" is used throughout the paper and is unclear to this reviewer. 3. After spraying the PCL layers, do the fibers fuse at the points of contact? What is the data for or against fiber fusion, which will significantly impact the vessel mechanical properties? What is the fiber diameter, is it uniform, how is it controlled? 4. The legend for Figure 4 is very confusing -panels a-c are stated as having data in the grey lines, but these lines are identical across the images, and the dotted lines are not described at all. It is pretty impossible to tell what this figure is saying. Do the authors have data to show on actual stress-strain curves of native human coronary arteries? 5. Similarly the legend for Figure 5 makes no sense -the grey lines and solid squares that are talked about the in the legend simply aren't there in the figure. And, panels b-e contain multiple black lines of unclear significance-if these are replicates, then should not some error bars be included or something? In Figures 5 and 7, where is the "grey region of the range of the model of the mechanical properties" derived from? This is not at all clear to the reader, and serves only to confuse. 6. In Table 1, what is the "n" value for these data? This bears on the statistical significance and differences. 7. Figure 8 needs scale bars. What is the meaning of the blue color -does this signal live or dead cells or proliferating cells? What are the dimensions of the construct? The reagent used does not show cell replication per se -it appears to show mitochondrial activity?? 8. Figure 9 shows cells counts from histology, but there are no histological images shown in the main paper. This is very disappointing. 9. In the Methods section, the Histology paragraph talks about skin transplants and positive controls for immunity, and talks about FACS analysis using different markers for immune cells, but none of this data is shown in the paper. It appears as if this text was taken from another manuscript, since it does not match up with the data in this paper? until failure. 23 In our first draft, we initially performed tensile tests using a fixed upper limit of strain to 24 evaluate mainly the stress-strain J-shape curve, without recording the UTS. Therefore, as well 25 mentioned by the reviewer, a reader could assume that results were plotted to failure. This is 26 now clarified in the text and material and methods. In the main text, Line 313 the following is 27 stated: "The construct exhibited mechanical response similar to human coronary arteries, 28 both in longitudinal and circumferential directions under uniaxial tensile testing up to a 30% 29 strain range ( Fig. 7a-b).", and additionally, in material an methods line 781, the following is 30 mentioned: "Uniaxial tensile testing was performed at a constant strain rate of 1 mm/min and 31 up to 30% strain (see Fig. S2b)." However, since the failure data is relevant for the application 32 of this technology, as stated by the reviewer, we included information about the tensile 33 testing 34 to failure and physiological burst pressure results in the main text and supporting material. 35 We have summarized these results in line 315 for the resistance until failure: "Longitudinally 36 tested SDVG exhibited a maximum failure strength of 520 ± 56 kPa, which is in the range of 37 coronary arteries of individual above 35 years old 18,45 (Fig S4)". In line 363 the following results 38 concerning burst pressure were included: "Due to safety and clinical concerns, vascular grafts 39 must also meet adequate burst pressures and suture retention strength 25,47 .The fabricated 40 full SDVGs exhibited burst pressures of 1630 ± 180 mmHg, similar to values reported for 41 human saphenous vein, as well used as autologous graft for coronary bypass 48 (see Fig S5).". 42 Being the suture retention strength a particularly important aspect of vascular grafts in 43 regards to safety and clinical practice, we have included in line 366 additional results of suture 44 7 30% strain). Beyond this, we have conducted pressurization testing, in which SDVG were 135 previously submitted to 5 cycles of 30% strain in the longitudinal direction after being 136 mounted cylindrically in the pressurization system coupled to the tensile machine (see 137 material and methods line 796) and 5 cycles of 200 mmHG pressure loading prior to test. In an 138 in vivo or clinical implantation setting, the pre-conditioning step may be obviated, although 139 this has not been addressed in the actual study (line 548: "It may even be possible to conduct 140 the preconditioning in situ on implantation by pulsatile blood pressure without need of 141 previous preconditioning and is the subject of future studies."). 142 143 It is not clear why the authors chose to perform uniaxial strip tests on the fabricated vessel 144 material, instead of simply testing biaxial behavior. This should be justified. 145 Response: Indeed, biaxial mechanical testing would be more appropriate for mechanical 146 analysis of vasculature tissues, as mentioned by the reviewer, especially as this type of tissue 147 is submitted to biaxial stretching during pulsatile blood flow. However, and since our study 148 strongly rely on previous experimental and theoretical studies based on uniaxial testing 149 analysis, we decided to use equivalent experimental setting to guide our fabrication design in 150 order to match the native mechanics. 151 152 Table 1 does not appear to be referenced in the text of the manuscript 153 Response: We thanks the reviewer for pointing this out. The rationale for the selection of HUVECs for the proliferation and LIVE/DEAD assay is now 172 clarified in line 381 ("HUVECs were chosen for this study as they have been identified as more 173 oxidative-stress and hypoxia sensitive cells than progenitor stem cells 50-54 , therefore more 174 sensitive to free radical polymerization and hypoxic conditions presented during 175 manufacturing. By performing this assay using HUVECs instead of progenitor cells (e.g. BM-176 MSCs), masking of the level of compatibility by high resistant cells is avoid, and it can be 177 demonstrated with higher certainly that the biofabrication technique is cytocompatible.") 9 BM-MSCs cells were selected for the fabrication of SDVGs and for the immunomodulatory 179 study in mice, the rationale behind this is now given in line 419 ("On the other hand, bone 180 marrow mesenchymal stem cells (BM-MSCs) are known to have immunomodulatory activity; 181 therefore, it is expected that immunoreaction in presence of this cell type in the SDVG be 182 controlled and the effect of endotoxins ameliorated. It is important to remark that in this 183 study, vascular cell functionality, such as contractility, was not under evaluation, instead 184 functionality of BM-MSCs, known as an excellent cell source for vascular remodeling and 185 regeneration 56 and considered within the SDVG design."), and reinforced in line 587 ("In vivo 186 immunogenicity experiments demonstrate that the encapsulation of BM-MSCs lowers the 187 rejection and inflammation response cause by the subcutaneous implantation of an 188 endotoxin-laden SDVG. The immunosupression cell function is of importance in the design of a 189 new vascular graft, considering a cause of graft failure is associated to chronic inflammatory 190 response 5 ") and line 603 ("Additionally, mesenchymal stem cells (MSCs) can be encapsulated 191 as a multipotent source of the biological component with the ability to differentiate to the 192 desired tissues, assist in the recruitment and migration of patient's cells by secreting different 193 chemokines, and help in the formation of a definitive anti-thrombogenic inner layer after 194 implantation 74 ."). 195 In general terms, vascular cells were not considered in this design due to potential 196 immunorejection that these cells could generate in patients in any future clinical application. 197 In order to use autologous vascular cells, these cells need to be harvested from the same 198 patient. This possibility is considered risky in terms of commercial and clinical viability, as cell 199 harvesting and expansion would require invasive procedures, long-term cell culture and 200 accompanying sterility issues, making the SDVG more expensive and less applicable for urgent 201 situations. iPS on the other hand, are potentially an excellent source for this application due to 202 the low invasiveness of their harvesting procedure, however, reprogramming, expansion and 203 differentiation are still long and expensive procedures for this application, without mentioning 204 the high concern of potential teratoma formation. As identified for use in this study, donated 205 BM-MSCs are considered less invasive, storable and immunotolerated. Additionally, these 206 cells could differentiate, immunomodulate and assist tissue regeneration, being considered as 207 key cells in the process of remodeling and tissue integration for tissue engineered grafts. 208 209 210 Please describe the details of the vessel culture from days 1 to 7. 211 Response: This has been added and detailed in the revised version in the materials and 212 methods section, specifically in the cell culture sub-section in line 819 ("Cell culturing of 213 complete or sectioned SDVGs were performed similarly, except that culture media was 214 additionally supplemented with 1X amphotericin B (15290-026, Gibco, USA). This preventive 215 measure was performed to mitigate exposure to microorganisms post fabrication."). 216 217 Cell survival appears to be equated to function, but it should be made clear that vessel 218 function (e.g. contractility) was not tested in this paper. USA), and incubated at 37°C, 5% CO 2 and 96% of humidity."). Lonza provided a certificate of 323 immunophenotypification and tridifferentiation following the ISSCR recommendations for this 324 type of cells. 325 All figure legends have been corrected, as suggested too by the reviewer Nº 2, and material 326 and methods section improved following the reviewer's suggestions, and addressed one-by-327 one below. Regarding statistical analysis, more detailed information has been added in 328 material and methods (line 937: "Data are presented as mean ± SD or mean ± SEM, also 329 indicated in the figure caption. Statistical significance was determined using two-tailed Mann-330 Whitney U test in compliance and cell density of bio-inspired SDVGs (Table 1, and Fig 8). Two-331 tailed Mann-Whitney U test was also applied for the in vivo immunosuppression functional 332 experiments in the mouse model, platelet activation assay and clotting time assay (Fig 9, 10 circulation using a rabbit model, results were presented as preliminary study, therefore only 2 345 animals per group were performed and no statistical analysis presented. 1. Line 123: the statement is made that the GEAL solution is pre-crosslinked, but then the 354 authors state that the dipped layer is exposed to UV light to crosslink the material. please 355 clarify. 356 Response: We thank the reviewer for pointing this out. The GEAL solution was not pre-357 crosslinked, we have now modified the body text to clarify this and revised Figure 2 to include 358 the crosslinking/photo-crosslinking step within the vessel production sequence. We have 359 clarified this in the manuscript in line 143: "Subsequently, the rod containing the orientated 360 PCL fibre sublayers was immersed (dipped) into the GEAL solution and slowly retracted whilst 361 spinning to allow homogenous GEAL layer photo-crosslinking using a lateral UV source (Fig.  362 2b)." 363 364 2. The phrase "circumferential axis" is used throughout the paper and is unclear to this 365 reviewer. 366 Response: This has now been clarified in the revised manuscript. Figure 1 has been revised 367 with a scheme of the multilayer vascular graft, defining the circumferential axis for clarity. 368 369 3. After spraying the PCL layers, do the fibres fuse at the points of contact? What is the data 370 for or against fibre fusion, which will significantly impact the vessel mechanical properties? 371 What is the fibre diameter, is it uniform, how is it controlled? 372 Response: We thank the reviewer for pointing this out. We have included more detail 373 concerning fibre morphologies, diameters and confirmed that the fibres are individualized by 374 SEM and CT investigation ( Fig. 4 and 6), however, some level of fusion between fibres cannot 375 be discarded. We have expanded the text concerning the fibres and the implication of their 376 fusion on mechanical properties. We have included those comments in line 295 ("The fibres 377 are individualized, with minimal fibre fusion evident, according to SEM and CT ( Fig. 4d and 6c, 378 repectively). Fibres fusion is an aspect of concern in the design of the SDVG, because fibres 379 fusion would indeed impact the mechanical response, mainly due to force distribution 380 amongst fibres at fixed contact points. Overall in the actual SDVG, free displacement of fibres 381 would be possible during stretching and recoil. "). Fusion of the fibres can be controlled by 382 fixing the distance between the SBS head (fibre emitting) and collection, if solvent evaporation 383 is insufficient prior to hitting the target then some fusing of the fibres can occur; we specified 384 the collection distance used here to obviate fibre fusion. As specified in the modified caption, human coronary data was obtained from a previous work 424 (Holzapfel et al (2005)), in which 13 donated human coronary arteries were evaluated. From 425 this data a constitutive mathematical model capable to describe the mechanical behavior in 426 uniaxial tensile testing was obtained. In our work, mechanical tensile testing was applied in 427 the same manner as their study, including strain rate as further detail in our response to 428 referee 2.  Table 1, what is the "n" value for these data? This bears on the statistical significance and 476

differences. 477
Response: We thank the reviewer for pointing this out. This has now been corrected in Table 1  478 of the revised manuscript, including details as to how the data was obtained. Specifically in 479 table 1 caption the following was included: " Table 1 wall thickness of the SDVG was 0.59 ± 0.17 mm, relatively similar to the combined thickness of 492 the middle and outer graft layer fabricated separately, and the inner diameter 3.6 ± 0.5 mm. 493 ."). Scale bars were included for the SDVG fluorescent image of SDVG in figure 8. 494 Additionally, the reviewer's concern about proliferation assay and mitochondrial activity has 495 been clarified in the manuscript in line 380 ("In order to evaluate the cell viability after 496 manufacturing, a cell proliferation assay based on mitochondrial activity was performed post-497 fabrication."), and line 390 ("Nevertheless, limited diffusion of nutrients and reagents of the 498 proliferation kit within the graft must be taken into consideration; this could lead to an 499 underestimation of the mitochondrial activity, hence cell survival and proliferation"). 500 501 8. Figure 9 shows cells counts from histology, but there are no histological images shown in 502 the main paper. This is very disappointing. 503 Response: We thank the reviewer for pointing this out. Histological images of rabbit carotid 504 grafted SDVG are presented now in section "Implantability study in arterial circulation using a 505 rabbit model" (line 502) and supporting information (Fig. S8). These images correspond to 506 H&E staining of cellularized SDVG after 14 and 30 days post-implantation (Fig. 10h,i), and H&E 507 staining of acellularized ( Fig S8a) and cellularized SDVG (Fig S8b) after 30 days post-508 implantation. On the other hand, cell counting presented in figure 9, corresponds to cells 509 obtained from graft-draining lymph nodes to evaluate the level of immunoreaction. This has 510 been clarified in the caption and as well in line 425 ("Analysis of immunocompetent mice with 511 24 dorsal subcutaneous implantations demonstrate that SDVGs with encapsulated endotoxins 512 and without BM-MSCs induced graft rejection when implanted subcutaneously, characterized 513 by a lack of graft incision healing (see Fig. S7a), an increased number of total cells isolated 514 from graft-draining lymph nodes (dLNs) (mouse axillary and brachial lymph nodes), and an 515 increased percentage of CD4+ memory T cells and B cells in dLNs compared to BM-MSCs 516 cellularized graft (Fig. 9 b, c, and d respectively) " 517 518 9. In the Methods section, the Histology paragraph talks about skin transplants and positive 519 controls for immunity, and talks about FACS analysis using different markers for immune cells, 520 but none of this data is shown in the paper. It appears as if this text was taken from another 521 manuscript, since it does not match up with the data in this paper? 522 Response: We thank the reviewer for pointing this out. This mistake has been corrected in 523 substantially revised materials and methods section, and as well further clarified in the 524 caption of Figure 9. Concerning the Figure 9, the following underlined modifications has been 525 included: 526 The authors are commended for performing additional SVDG strength characterization and a rabbit carotid artery interpositional implantation study. The reported strength values are reasonable albeit lower than for other TEVG that have attained large animal testing. Unfortunately, acute clotting of the grafts in the rabbit model indicates the graft, despite its noteworthy structure and compliance properties, is not yet a successful TEVG in terms of a large animal implantation; thus, the study, while comprehensive and well presented, is of limited significance given the state of the TEVG field, and I do not think its probable impact merits publication in Nature Communications.

Reviewer #2 (Remarks to the Author):
This is an exciting paper describing an innovative approach for fabrication of vascular grafts that effectively mimic the mechanical response of native vessels and provide a scaffold for vascular cells to function and remodel. The authors have effectively responded to the reviewers' recommendations, and the manuscript is substantially improved, particularly with inclusion of additional results addressing graft strength and performance in a rabbit model. There are still some grammatical issues here and there, but they are not so extensive as to lead to confusion. General Comments: This is a substantially revised manuscript regarding the use of electrospinning techniques to create PCL-gel composite tissues with mechanics similar to native artery. The changes in response to the previous review are extensive, and so this reads almost like a new paper.
Compared to the prior submission, the clarity of the procedures and the characterization of the scaffolds, particularly the mechanical characterization, are vastly improved. This reviewer is again impressed at the extent to which a biological approach to orientation of PCL fibers led to expected and predicted compliance properties. In addition, the mechanical properties of these smalldiameter conduits do now appear to be compatible with arterial implantation. As such, this material characterization and the methodology combine to make this an important advance for research in the area of synthetic arterial grafts.
But while there are many improvements, there are some new weaknesses in the paper, mostly pertaining to the biological characterization of the materials. To remedy some of these deficiencies, it would be suitable (in this reviewer's opinion) to simply remove some of the confusing biological data, since it does not add significantly to the important parts of this story as it currently stands. In addition, the paper is now QUITE LONG, and should be streamlined to about half of its length, in terms of text. Specific comments below.
Specific Comments: 1. In Figure 5, providing more explicit labels on the y-axes would help readers understand which wall stress they were looking at. 2. Figure 6 does not add a lot to the paper, and could be omitted or moved to a supplement. 3. Figure 7 -again, more explicit y-axis labels, with directionality for panels a-c, and with amount of pre-stretch for panels d-f. 4. Table 1 could be omitted, as not adding a lot to the figures already presented. 5. English language correction, line 396, should read: "highly resistant cells is avoided" 6. Regarding results in lines 394-406, it should be noted that 0.02 -0.2% survival is still very poor for HUVEC in the construct. Is poor survival the reason that HUVEC were abandoned for later implantations? Also, what was the survival of BM-MSC when implanted into the constructs? 7. The authors seem to have an incomplete appreciation of the effects of endotoxin. In lines 430-444, there is discussion of endotoxin, MSC, rejection, etc. Some corrective observations: a. HUVEC will be susceptible to endotoxin -it is toxic to endothelium b. Endotoxin is a strong inducer of inflammation, but not of adaptive immunity, per se. Therefore, the B-and T-cells that migrated to the implant were likely part of a non-specific inflammatory response, rather than part of an actual rejection event. If the implants did not have cells, then rejection, in the proper sense, could not have occurred. c. Adding MSC to this cocktail may have resulted in fewer cells on FACS, but the meaning of this observation is really not clear. d. This reviewer would recommend that the cutaneous implants be struck from the paper, since they do not add value in terms of understanding of the construct, and provide some confusing information. 8. For the rabbit carotid implants, it is worthwhile to point out that endotoxin itself can induce endothelial inflammation and hence thrombosis. This may have been why all of the implanted grafts clotted within a short time period. The amount of endotoxin in the constructs should be quantified to gain a better understanding of what is going on here.

Reviewers' comments:
Reviewer #1 (Remarks to the Author): The authors are commended for performing additional SVDG strength characterization and a rabbit carotid artery interpositional implantation study. The reported strength values are reasonable albeit lower than for other TEVG that have attained large animal testing. Unfortunately, acute clotting of the grafts in the rabbit model indicates the graft, despite its noteworthy structure and compliance properties, is not yet a successful TEVG in terms of a large animal implantation; thus, the study, while comprehensive and well presented, is of limited significance given the state of the TEVG field, and I do not think its probable impact merits publication in Nature Communications.
Response: We thank the reviewer for their positive comments regarding our additional investigation to strength characterization and the rabbit carotid study.
Although we understand the author's concern pointing to the fact that the paper is not showing a successful long-term patency in a large animal model, the main focus of the manuscript is the new technology based on an innovative combination of modular and automated technologies capable to produce a tissue engineered product with standardized features, which are broadly considered essential for the efficacy of a vascular graft. We present results that show good potential towards clinical usage, specifically, platelet activation, coagulation, burst pressure, suture strength and hemostasis. Additionally, the main manuscript argues at the end of the discussion section about possible reasons that could cause the thrombus formation, basically indicating that a carotid rabbit model is not optimal for evaluating long-term patency especially for an engineered graft with mechanics and wall sizes corresponding to human coronary arteries. Therefore, it is mentioned too that only in a large animal models, the in vivo relevance of mechanical matching and the graft design will be fully demonstrated, which is part of our ongoing research, moving beyond the scope of the current manuscript. In regard to the possible reasons that caused the thrombus formation after 12 h, we included in the previous revision that luminal unalignment of the SDVG and the native rabbit carotid at the anastomosis zone could cause blood recirculation and stagnation, therefore, coagulation (see lines 554-556 in the new revision). In this new version, we have additionally included a second possible cause, which refers to inflammation-induced thrombogenesis due to the endotoxin levels of the SDVG (see lines 558-566 in the new revision). To complement the discussion, we quantified and reported the endotoxin level in fabricated SDVG and performed in vivo immunogenicity assay to evaluate the immune reaction. These results were included in the new manuscript and new supporting information, respectively, and concluded that inflammation cannot be discarded as a possible cause of thrombus formation in the present study.
The technology presented in this manuscript is a unique rapid manufacturing process that integrate the use of natural biomaterial and cells simultaneously, conforming a disruptive strategy to overcome standardization, manufacturing control, regulatory and commercial hurdles present in the tissue engineered field, which in combination could reduce the gap for this type of product to commercialization and impact in the public health.

Reviewer
#2 (Remarks to the Author): This is an exciting paper describing an innovative approach for fabrication of vascular grafts that effectively mimic the mechanical response of native vessels and provide a scaffold for vascular cells to function and remodel. The authors have effectively responded to the reviewers' recommendations, and the manuscript is substantially improved, particularly with inclusion of additional results addressing graft strength and performance in a rabbit model. There are still some grammatical issues here and there, but they are not so extensive as to lead to confusion.
One minor suggestion: Please define (in the figure caption) the angle alpha shown in Figure 1. Is it the parameter associated with the angles referenced in the caption for Fig. 1?
Response: We thank the reviewer for his comments. We have revised the manuscript for grammatical and English errors throughout.
Additionally, we have included a definition for the angle alpha in the revised caption of figure 1, that effectively correspond to the deposition angle as stated by the reviewer.
" Figure 1: Scheme composition of the middle and outer graft layers. The middle graft layer comprises a series of four PCL/GEAL sublayers, hereafter called middle graft sublayers, with fibres deposited at angles of +/-21° and GEAL sublayer generated after two cycles of dipping and photo-crosslinking. The outer graft layer is composed of a series of five PCL/GEAL bilayers, hereafter termed outer graft sublayers, with fibres deposited at angles of +/-67° and GEAL sublayer generated after three cycles of dipping and photo-crosslinking. The deposition angle is represented by "α". Compared to the prior submission, the clarity of the procedures and the characterization of the scaffolds, particularly the mechanical characterization, are vastly improved. This reviewer is again impressed at the extent to which a biological approach to orientation of PCL fibers led to expected and predicted compliance properties. In addition, the mechanical properties of these small-diameter conduits do now appear to be compatible with arterial implantation. As such, this material characterization and the methodology combine to make this an important advance for research in the area of synthetic arterial grafts.
But while there are many improvements, there are some new weaknesses in the paper, mostly pertaining to the biological characterization of the materials. To remedy some of these deficiencies, it would be suitable (in this reviewer's opinion) to simply remove some of the confusing biological data, since it does not add significantly to the important parts of this story as it currently stands. In addition, the paper is now QUITE LONG, and should be streamlined to about half of its length, in terms of text. Specific comments below.
Specific Comments: 1. In Figure 5, providing more explicit labels on the y-axes would help readers understand which wall stress they were looking at. 2. Figure 6 does not add a lot to the paper, and could be omitted or moved to a supplement. 3. Figure 7 -again, more explicit y-axis labels, with directionality for panels a-c, and with amount of pre-stretch for panels d-f. 4. Table 1 could be omitted, as not adding a lot to the figures already presented. 5. English language correction, line 396, should read: "highly resistant cells is avoided" 6. Regarding results in lines 394-406, it should be noted that 0.02 -0.2% survival is still very poor for HUVEC in the construct. Is poor survival the reason that HUVEC were abandoned for later implantations? Also, what was the survival of BM-MSC when implanted into the constructs? 7. The authors seem to have an incomplete appreciation of the effects of endotoxin. In lines 430-444, there is discussion of endotoxin, MSC, rejection, etc. Some corrective observations: a. HUVEC will be susceptible to endotoxin -it is toxic to endothelium b. Endotoxin is a strong inducer of inflammation, but not of adaptive immunity, per se. Therefore, the B-and T-cells that migrated to the implant were likely part of a non-specific inflammatory response, rather than part of an actual rejection event. If the implants did not have cells, then rejection, in the proper sense, could not have occurred.
c. Adding MSC to this cocktail may have resulted in fewer cells on FACS, but the meaning of this observation is really not clear. d. This reviewer would recommend that the cutaneous implants be struck from the paper, since they do not add value in terms of understanding of the construct, and provide some confusing information. 8. For the rabbit carotid implants, it is worthwhile to point out that endotoxin itself can induce endothelial inflammation and hence thrombosis. This may have been why all of the implanted grafts clotted within a short time period. The amount of endotoxin in the constructs should be quantified to gain a better understanding of what is going on here.
Response: We thank the reviewer for their helpful comments. As suggested by the reviewer, we have moved some biological data from the main manuscript and included in the supporting information (see below for further details). Additionally, we have clarified and simplified the information obtained from the immunogenicity assays and presented in the main manuscript (complete results are included now in the supporting information). Also, we have shortened the text in about 1500 words to adjust to the length of other tissue engineering related papers previously published in Nature Communication and performed new experiment to tackle the endotoxin issue in this work. Following the specific comments, we have undertaken the following modification, addressed here comment by comment (reviewer #3 comments in blue): Reviewer #3 comment 1. In Figure 5, providing more explicit labels on the y-axes would help readers understand which wall stress they were looking at.
-Following the revierwer's suggestions, a schematic representation of the direction of tensile testing was included in our now revised figures 3, 4, 5 and 6 for clarity.
Reviewer #3 comment 2. Figure 6 does not add a lot to the paper, and could be omitted or moved to a supplement.
- Figure 6 of the previous submitted manuscript was removed from the main manuscript as suggested by the reviewer and is now included in supporting information.
Reviewer #3 comment 3. Figure 7 -again, more explicit y-axis labels, with directionality for panels a-c, and with amount of pre-stretch for panels d-f.
-Please see our response to reviewer #3 in comment 2.
Reviewer #3 comment 4. Table 1 could be omitted, as not adding a lot to the figures already presented.
-Following the reviewer´s suggestion, Table 1 has been removed and is now included in the supporting information.
-After shortening the main manuscript, line 396 has been removed, therefore, the specific language correction was not necessary.
Reviewer #3 comment 6. Regarding results in lines 394-406, it should be noted that 0.02 -0.2% survival is still very poor for HUVEC in the construct. Is poor survival the reason that HUVEC were abandoned for later implantations? Also, what was the survival of BM-MSC when implanted into the constructs?
-Although we have removed this part from the newly revised manuscript, the following discussion is necessary to be raised in order to clarify the reviewer´s concern. In the previous submitted manuscript, we have attributed the low signal of metabolic activity to the limited diffusion of reagents of the proliferation kit within the graft; this could lead to an underestimation of the mitochondrial activity of cells within the graft. This is particularly true for encapsulated cells within hydrogels with limited solvent access. However, we cannot discard that the lack or decrease in the capacity of WST-1 reduction of cells is derived either from a metabolic resting induced by a free radical-derived oxidative stress during photo-crosslinking ( 2015 Feb; 0: 281-291), or simply by an interference of the photoinitiatorderived free radical with the reduced form of the 1-Methoxy-5methylphenazinium methyl sulfate during the electron transfer in the reduction process of WST-1 to formazan. Although these phenomena could occur in conjunction and affect the reduction of WST-1 during the assay, viability measurement based on assays that evaluate membrane integrity still indicate that cells are viable (please see Figure 7 of the new manuscript).
Considering the mentioned doubts about the proficiency of this method in measuring cell viability in the actual cell-encapsulation scenario, and the fact that the cell viability and compatibility of cells within GelMA hydrogels has been previously explored by this group (Biofabrication. 2016 Dec 1;9(1):015001) and other research groups (Biomaterials. 2010 Jul;31 (21):5536-44), we have decided to include only live/dead assay and immunosuppression functionality in the in vivo model as proofs of viability of cells in the present manuscript.
Concerning the decision of abandoning HUVECs in the final SDVG design, this is not related to the survival of HUVECs but to the expected and more appropriate role of BM-MSCs in the final design with potential regenerative activity after implantation. Reasonings in this regard are included in the new revision: In lines 363-383: "Having considered the incorporation of a cellular component essential for the design and successful outcome in a transplantation scenario 32,49,50 , important practical implication must be taken into account when choosing an appropriate cell type. Although autologous vascular cells are the preferable choice, invasive harvesting and long-term culturing, specially for elderly patients, make this option risky in terms of commercial and clinical viability. Induced pluripotent stem cells on the other hand, are potentially an excellent source for this application 51 due to the low invasiveness of their harvesting procedure and autologous nature, however, reprogramming, expansion and differentiation are still a long and expensive procedures, and the frequency of point mutations 52 has generated serious concern about the safety of these cells. Allogenic bone marrow mesenchymal stem cells (BM-MSCs) are considered less invasive, storable, economically feasible, immunotolerated, and have been physiologically implicated in vascular repair and remodeling 53 . Additionally, BM-MSCs are known to have immunomodulatory activity; therefore, it is expected that immunoreaction or inflammation in presence of this cell type in the SDVG would be controlled or ameliorated. In this regard, BM-MSCs has been proposed in this study as a source of biological function for the actual SDVG design. Analysis of immunocompetent mice with dorsal subcutaneous implantations has demonstrated that SDVGs with encapsulated BM-MSCs are capable to control an inflammatory response, whereas non-cellularized SDVG not (see Fig. S9, S10 and supporting information for further details). This demonstrates also that cells maintain their viability and functionality after being subjected to the manufacturing process. It is important to remark that in this study, vascular cell functionality, such as contractility, was not under evaluation, instead functionality of BM-MSCs, known as an excellent cell source for immunomodulation, vascular remodeling and regeneration 54 ".
In regard to the BM-MSCs survival, quantification of live cells was included in the new manuscript in lines 340-344: "Using this time bone marrow-derived mesenchymal stem cells (BM-MSCs), and a LIVE/DEAD® cell staining assay, viability analysis of cells located within different sublayers of the fabricated SDVG was performed at different time points of static cell culturing ( Fig. 7fh). Evaluation on day 7, 14, 21 and 28 resulted in 71, 84, 87 and 92% viability respectively (Fig. S8). These results confirm cells survival and even proliferative capacity on day 28 post fabrication (Fig 7i)".
Reviewer #3 comment 7. The authors seem to have an incomplete appreciation of the effects of endotoxin. In lines 430-444, there is discussion of endotoxin, MSC, rejection, etc. Some corrective observations: Considering the reviewer`s suggestion, the subcutaneous implantation was moved to the supporting information and taken only as an additional information to prove the immunosuppressive capacity of BM-MSCs in the present construct. Additionally, the used of a low endotoxin alginate for the SDVG construct was specified in the materials and methods section in lines 593-594 and 600-602, clarifying too that the use of an alginate with higher level of endotoxin was only for the subcutaneous assay in the context of testing the immunosuppressive activity of encapsulated BM-MSCs.
Reviewer #3 comment 7a. HUVEC will be susceptible to endotoxin -it is toxic to endothelium.
Comment 7a: Although the LPS toxicity on endothelial cells has been previously stablished by other authors (Infect Immun. 1993 Aug;61(8): 3149-3156.), in this investigation, low endotoxin alginate has been used in the fabrication of SDVG, as mentioned previously, therefore, the authors of this article did not consider the possible toxicity of endothelial cells after the SDVG implantation. However, discussion and experimental proofs regarding this point are mentioned in the paper and in this document further below.