The free radical NO mediates a number of physiological functions. The Moncada group has described how NO production is one of the most important function of the endothelium (Moncada and Higgs, 1991). It regulates blood flow and cell–cell communication, and it is also a key component in the anti-atherosclerotic properties of the endothelium (Wever et al., 1998). NO is generated in the body when nitric oxide synthase (NOS) enzymes degrade L-arginine (Wever et al., 1998). Endothelial NOS is NADPH and Ca²⁺/calmodulin dependent. The biological impacts of NO have been extensively investigated, and are only briefly summarized below. A variety of stimuli such as bradykinin have been shown to initiate NO generation (Palmer et al., 1987; Böger et al., 1996). NO is a very reactive gas with a short half-life in the body. At healthy levels, NO often reacts with heme-containing molecules, either through the protein moity to form S-nitrosothiols or directly with the metal ion. For example, NO activates guanylate cyclase (GC), converts guanosine triphosphate (GTP) to cyclic guanosine monophosphate (cGMP), and activates protein kinase G (PKG) that can lead to several results important for vascular hemostasis (Figure 2A) (Ignarro et al., 1987). Healthy NO signaling has several impacts including mediating vasodilation, inhibit platelet aggregation, decreasing SMC proliferation, and reducing leukocyte adhesion and migration (Garg and Hassid, 1989; Jeremy et al., 1998; Sarkar and Webb, 1998; Gewaltig and Kojda, 2002). However, NO can also react with oxygen or oxygen radicals in the body (Figure 2B). The compounds that result from this reaction have been associated with atherosclerosis and other inflammatory diseases. For example, high concentrations of the reaction product peroxynitrite can result in cytotoxic peroxynitrous acid (Pryor and Squadrito, 1995), hydroxyl radical toxicity (Beckman et al., 1990) and protein fragmentation by nitration of amino acids (Ischiropoulos et al., 1992). Thus, the pro-inflammatory potential of NO generally occurs indirectly and is most prevalent at high concentrations of NO. Peroxynitrite can react directly with some targets (e.g., to activate MMPs) but most of its impacts are through oxidation via a free radical reaction (Szabó et al., 2007). Most of these effects of peroxynitrite promote an inflammatory response. The dual effects at different doses may contribute to the fact that the benefits of NO therapy in adult patients with pulmonary hypertension are still uncertain (Griffiths and Evans, 2005).

The formation of peroxynitrite is expected to be very low in normal, healthy physiological conditions because of the reaction kinetics (Wever et al., 1998). It has been estimated to be in the pico- to nanomolar range. In addition, there are molecules such as glutathione that can protect against oxidation by peroxynitrite (Sies et al., 1997). However, in diseased cardiovascular tissues (e.g., arteries with atherosclerosis), both the oxidative stress and inducible NOS (iNOS) will increase significantly (Wever et al., 1998; Mudau et al., 2012). This will result in higher rates of peroxynitrite production that can potentially cause cell damage. In addition, patients with hypercholesterolemia can have impaired function of glutathione, reducing the protection against reactive species derived from NO (Ma et al., 1997). One question is what impact specific doses of exogenous NO could have if delivered as part of a therapy. Inhaled NO delivered to infants with pulmonary hypertension are typically provided at ≤40 ppm (Weinberger et al., 2001). A clinical trial of inhaled NO (20 and 80 ppm in some patients) showed that the treatment was considered safe as monitored by signs of seizures requiring therapy or intracranial bleeding (Neonatal Inhaled Nitric Oxide Study Group, 1997). Clear toxicity has been observed in the clinic with exogenous doses over 80 ppm. However, modified tyrosine residues, which is a potential concern, were observed in the lungs of infants treated with 20 ppm doses (Hallman et al., 1998; Weinberger et al., 2001). Evidence suggest that there is a dose dependence where inhaled NO at a high 100 ppm dose increased ROSs, but 50 ppm dose exhibited anti-inflammatory properties (Weinberger et al., 2001). Overall, it is clear that the dose must be considered when incorporating NO and NO releasing molecules as part of a vascular tissue engineering strategy.

Carbon monoxide is generated naturally in the body when the heme oxygenase (HO) enzyme degrades heme molecules (Tenhunen et al., 1969). CO is most commonly known because high doses, such as those found with smoking, can be pro-inflammatory and higher doses are fatal. This is probably one reason why exploration of biomedical uses of CO started more recently than NO. However, CO in appropriate doses has anti-inflammatory properties (Motterlini, 2007). In addition, animal models with an abnormality in endogenous CO metabolism within cardiovascular tissues contribute to various pathological situations including hypertension, cardiac failure, and inflammation (Wu and Wang, 2005). Further, a patient died because of a unique clinical scenario where HO enzyme generation was reduced, and as a result less CO was present (Yachie et al., 1999). HO-1 has been shown to provide oxidative stress protection by itself (Yachie et al., 1999). However, CO has also independently been shown to provide anti-inflammatory properties. For example, it has been shown that CO acts by inhibiting the expression of the pro-inflammatory cytokines tumor necrosis factor alpha (TNF-α), interleukin-1beta (IL-1β), and macrophage inflammatory protein-1beta (MIP-1β), as well as by increasing the production of the anti-inflammatory cytokine IL-10 (Otterbein et al., 2000). Similar to NO, there are multiple mechanisms of action for CO. The Center for Disease Control (CDC) defines two modes of action which are (a) the hypoxic one that make CO dangerous in high dosage and (b) the non-hypoxic ones from endogenous CO (Wilbur et al., 2012). It is these non-hypoxic mechanisms that are of interest for CO delivery for tissue engineering applications.

Defining specific levels for high and low CO doses, as well as the associated signaling pathways, have been major areas of research. There has been progress, but there are still also many remaining questions. The specific impacts of CO and the associated mechanisms have been reviewed in detail in other review articles and book chapters (Ryter and Otterbein, 2004; Motterlini and Otterbein, 2010; Untereiner et al., 2012; Wilbur et al., 2012; Motterlini and Foresti, 2017). The non-hypoxic impacts range from vasodilatory, anti-apoptotic, and anti-inflammatory impacts important for vascular tissue engineering to neural signaling and prevention of ischemia-reperfusion injury during organ transplant (e.g., kidney). Below, we provide a summary of the some of the important pathways with a focus on those relevant to tissue engineered vascular grafts.

Carbon monoxide binds to fewer targets than NO, but there are still several heme-containing compounds that it has been shown to bind to. These include hemoglobin, myoglobin, soluble guanylate cyclase (sGC), cytochrome c, cytochrome p-450, and NOS. In addition, there are a variety of pathways that have been identified. The toxic effects at high doses have been linked to CO binding to hemoglobin and myoglobin that replaces oxygen (O₂). CO binds hemoglobin about 220 times stronger than (O₂), replaces O₂ from oxyhemoglobin, and forms carboxyhemoglobin (COHb) (Nguyen and Boyer, 2015). As a result, reduced oxygen transport capacity of the red blood cells can inhibit oxygen delivery to tissues. In addition, the binding of CO to myoglobin, which typically provides an oxygen reserve to skeletal muscle cells, has been linked to muscle fatigue (Wilbur et al., 2012). There is the possibility that the negative effects of CO binding to hemoglobin and myoglobin also include eliciting pro-inflammatory effects, such as neutrophil recruitment in the brain and lipid peroxidation (Peng et al., 2013). However, a CDC publication concludes that it is uncertain if plasma inflammation and coagulation markers increase at low to medium levels of CO exposure because some of the literature results appear to be seemingly contradictory. The negative impacts and mechanisms of high CO doses are generally considered well established. However, a Motterlini and Foresti (2017) review article mentions that CO binding to hemoglobin and myoglobin is actually a protective mechanism and that it is the CO that does not bind to these heme compounds and instead binds extensively to cytochrome c oxidase within cell mitochondria that cause the toxic effects. They cite relative CO affinities to the three heme-containing compounds as well as a dog study that suggested that free CO gas is more dangerous than CO bound to hemoglobin. Although the Motterlini hypothesis about hemoglobin toxicity will require further investigation, it suggests the importance of considering biodistribution in tissue and within cells in understanding the difference between toxic and protective CO doses.

The beneficial effects of CO have been shown to occur through binding to other heme-containing compounds within the cell. Early work focused on CO binding to sGC that leads to increased cGMP production. This is the best characterized pathway, and it has been shown to lead to impacts such as vasodilation, increased thrombolysis, and decreased SMC proliferation (Suematsu et al., 1994; Fujita et al., 2001; Ryter and Otterbein, 2004). However, other heme-containing targets for CO and the corresponding pathways have continued to be identified. Some of these are shown in Figure 3. The variety of pathways help to explain the different effects that CO has been shown to have on different cell types. For example, CO appears to promote endothelial cell proliferation but reduce SMC proliferation (Figures 3A,B) (Ryter and Otterbein, 2004; Untereiner et al., 2012). Both of these results are beneficial in preventing vessel stenosis and intimal hyperplasia. The suppression of SMC proliferation has been linked to CO binding to either sGC or NOS, and the signaling proceeds through a cGMP-dependent pathway (Morita et al., 1995; Song et al., 2002; Botros and Navar, 2006). This has also been shown to involve transcription factors (e.g., E2F) (Ryter and Otterbein, 2004) and proteins that regulate the cell cycle, including increased levels of the cyclin-dependent kinase inhibitor p21 (Zhou et al., 2005). However, the promotion of endothelial cell proliferation by CO has been linked to CO binding with NOS, and involves Ras homolog gene family member A (RhoA) and Akt (Wegiel et al., 2010). This is one of the pathways that demonstrates an interaction between both the gasotransmitters CO and NO. In addition, CO has been shown to provide anti-apoptotic effects on endothelial cells through a p38 mitogen activated kinase pathway (MAPK) that includes upregulation of anti-apoptotic genes [e.g., inhibitors of apoptosis 2 (IAP2)] (Brouard et al., 2002). Interestingly, CO may also provide dose-specific modulation of cellular apoptosis through mitochondrial swelling. Queiroga et al. (2011) have reported that 10–100 μM CO prevented mitochondrial permeabilization and prevented release of pro-apoptotic factors from the mitochondria but 250–500 μM CO caused mitochondrial swelling.

The anti-inflammatory properties of low CO doses also appears to act through MAPK-dependent pathways. Activation of p38 has been discussed most extensively, but more recently ERK downregulation has also been implicated in these effects (Untereiner et al., 2012). A continuing question about this pathway is how CO initiates the response since it does not bind directly to any of the upstream components. It may be that mitochondria-derived ROS is what initiates the p38 signaling pathway and leads to the anti-inflammatory effects (Otterbein et al., 2000). This is supported by work by Bilban et al. (2006) that showed that macrophages exposed to CO produced temporary bursts of ROS that led to peroxisome proliferator-activated receptor gamma (PPARγ) expression (Bilban et al., 2006). They further showed that blocking PPARγ prevented the anti-inflammatory effects from CO. An anti-inflammatory effect through ROS would be expected to be CO dose dependent since levels of ROS that are too high can lead to a damaging oxidative stress environment.

Carbon monoxide does have a couple of potential advantages for controlling the inflammatory response to vascular grafts despite the increased importance of dose control compared to NO. CO is a more stable molecule than NO (Zhuo et al., 1993). In addition, CO targets only transition metals (e.g., iron) (Motterlini and Otterbein, 2010), unlike NO that interacts with a large number of cell targets. The stability of CO may be a benefit when in the high oxidative stress environment that would typically be found in diseased tissue and after initial graft implantation (Desmard et al., 2012). In this situation, the half-life of released nitric oxide would be very limited and some of the reaction products could cause negative side-effects. Work by Motterlini and Foresti (2017) has suggested that NO is more common in healthy tissues, but HO1-derived CO is more important in tissues with higher levels of oxidative stress where NO is less stable. It is known that NO can increase endogenous CO through an increase in HO-1, and alternatively CO can increase NO generation through binding to NOS (Untereiner et al., 2012). For example, it has been shown that inhaled CO reduced pulmonary hypertension in a mouse model by generating eNOS and NO (Zuckerbraun et al., 2006). Their eNOS deficient control mouse did not exhibit the treatment benefits observed from wildtype mice when CO treatment was provided (Zuckerbraun et al., 2006). These results are supportive of a comparative benefit of CO over NO in an oxidative environment. However, it has also been mentioned that the negative impacts of higher doses of CO (e.g., blood COHb levels ≥2.4%) are more pronounced in patients with previous cardiovascular disease than those who were healthy (Wilbur et al., 2012). The benefits of NO and CO therapy in an inflammatory environment will require more investigation of the dose dependent differences and differences in distribution between particular tissues and cells.

One of the biggest challenges with interpreting the results of previous CO studies has been determining what constitutes a high and low CO doses. The CO dose has typically been defined as either ppm in the inhaled gas or percent CO bound to a heme compound – COHb in vivo in the blood or COMb for in vitro tests. Figures 4, 5 show different studies and reported results at different CO doses. Common endogenous levels are also shown. The endogenous levels range from 0 to 0.5%, but are hard to define. HO-2 is expressed at varying levels in different tissues, with the highest expression in tissues where there are increases in estrogen and glucocorticoids (Wilbur et al., 2012). In addition, HO-1 is induced upon injury, so the specific levels will vary. The higher the HO levels, the higher the amount of CO will be.

While endogenous and high lethal levels can be roughly defined, there are questions about the impacts of CO at levels in between. The chart in Figure 4 shows that ppm CO in the inhaled gas provides conflicting information about positive and negative levels when different clinical and pre-clinical studies are compared. CO has been reported to demonstrate significant medical benefits at concentrations of the inhaled gas ranging from 10 to 500 ppm (Mann and Motterlini, 2007; Motterlini and Otterbein, 2010), but the CDC report discussed concerns to particular tissues at ppm levels as low as 12 ppm (Wilbur et al., 2012). The results could be influenced by differences in duration and frequency of exposure, so COHb is often used as an alternative measure. Figure 5 shows COHb levels which are reported less often than ppm CO in the inhaled gas, and are often theoretically predicted instead of experimentally determined (Wilbur et al., 2012). This figure demonstrates that consideration of the COHb levels also exhibits apparent discrepancies for positive and negative levels between studies. For example, the CDC defines 2.4% COHb as the level for the lowest observed adverse effect but other studies have demonstrated beneficial impacts for higher levels of CO [e.g., less than a 15% COHb toxic threshold for a study conducted on healthy, non-smoking, young-men (Benignus et al., 1987)]. It has also been noted that there is significant patient to patient variability in the response to COHb levels, with patients with the same COHb level exhibiting very different responses (Gorman et al., 2003; Grieb et al., 2008). Some patients have been shown to be without symptoms at comparably high COHb levels in their blood as the Benignus et al. (1987) study demonstrated (Motterlini and Foresti, 2017). The difference in levels is important because the projected CDC levels are similar to what a body would be exposed to in normal ambient air, but the higher values in other studies suggests a benefit of additional supplemental CO. Finally, it is even more uncertain how COMb levels found with in vitro studies relate to the results in the clinic and in animal models. These studies can provide comparative release rates and biological impacts for different compounds and doses, but an extrapolation to the in vivo environment would require mathematical modeling and consideration of mass and fluid transport.

Carbon monoxide biodistribution is likely important to interpret these results and determine optimal levels for supplementary CO doses, but this information is not as well known. This information is also important because different tissues have varying sensitivity to CO. For example, the respiratory track is one of the least sensitive. However, this would stress the importance of local delivery to avoid any systemic effects as well as prevent the binding of most of the CO to hemoglobin in the blood stream so that it can target the heme compounds within the cell (Motterlini and Foresti, 2017).

Since the dose of NO or CO are very important, the drug delivery strategy is also critical. This is the reason why molecules that degrade to release NO or CO in a controlled manner or in response to a stimuli are being investigated.

Several different NO releasing molecules have been used (Frost et al., 2005). These include diazeniumdiolates and S-nitrosothiols (SNO) that release NO when in contact with aqueous solutions such as body fluids. The use of SNOs was likely inspired by the native SNOs found in the body that provide a reservoir of NO (Stamler, 2004). One advantage of SNO is its ability to be regenerated endogenously in the body. Some of these SNOs are compounds that have been synthesized to extend the release time for NO (e.g., by making them more hydrophobic to shield them from metal ion interactions) (Kumari et al., 2014). SNOs have been shown to be active pharmaceutical agents, including inducing vasodilation, and have been incorporated within a wide variety of drug delivery strategies (Liang et al., 2015). These SNOs have also been tested on vascular cells (e.g., SMCs and adventitial fibroblast) including in culture with the antioxidant AA (Gregory et al., 2011). Gregory et al. (2011) found that AA and their particular SNO are complimentary, with AA increasing NO release (Gregory et al., 2011).

A variety of carbon monoxide releasing materials (CORMs) have been developed, as have been described in several review articles (Foresti et al., 2008; Motterlini and Otterbein, 2010). The classes of these CO-releasing molecules can be categorized as aldehydes, boroncarboxylates, metal carbonyl complexes, organometallic compounds, oxalates, and silacarboxylates (Romaõ et al., 2012). The majority of these compounds are metal carbonyls that quickly release CO in the presence of water, but many of which contain heavy metals. There is also a boron carbonate compound that does not contain heavy metals and releases CO at a slower rate. However, CORMs that release CO in response to light are being investigated to allow for a more controlled release. These light responsive CORMs include a manganese complex and organic molecules. CORMs have been shown to elicit anti-inflammatory properties at the right dose for a variety of tissues (Motterlini and Otterbein, 2010). For vascular applications, there are currently limited results but one study showed that a CORM induced extended vasodilation and reduced acute hypertension in explanted rat aortic rings (Motterlini et al., 2002).

For tissue engineering, incorporation within a scaffold is important and can provide controlled, local delivery of the gasotransmitter (Motterlini et al., 2002; Foresti et al., 2008). NO releasing materials have been incorporated within tissue engineered scaffolds as well as modified PTFE traditional vascular grafts (Gregory et al., 2011). For example, NO has been added to the surface of synthetic scaffolds both covalently for extended delivery (i.e., diethylamine/N₂O₂) and non-covalently (spermine/N₂O₂) (Frost et al., 2005). The main application of these NO releasing materials incorporated within scaffolds has been as a coating of extracorporeal circulation devices such as those used for hemodialysis (Yang et al., 2015). The primary goal for this application has been to provide a NO-releasing substitute for endothelial cells that can prevent platelet deposition and thrombosis, and the studies have demonstrated success with this goal. For example, Skrzypchak et al. (2007) coated poly(vinyl chloride) tubing used as an arteriovenous shunt in a rabbit model with for diamine-based diazeniumdiolates material that provided clotting protection with DBHD/N₂O₂ loadings of 10 – 50 wt%, but not for a lower 2% dose. Toxic nitrosamines can form when many of the NO releasing materials leach out of the coating, but techniques such as lipophilic modification of molecules to reduce leaching and development of new NO releasing materials with lower toxicity concerns are current strategies to overcome this challenge (Yang et al., 2015). Another possible limitation has been that the lack of extended NO release.

Nitric oxide can have other effects relevant for vascular tissue engineering, but these have been less of a focus with the current studies. An NO-releasing poly(diol-co-citrate) elastomer has been shown to reduce SMC proliferation in cell culture and reduce intimal hyperplasia when used a perivascular wrap in a rat model (Serrano et al., 2011). This was performed on a vessel injured by balloon inflation. The incorporation of NO releasing materials within tissue engineering scaffolds that are intended to integrate with the surrounding artery has also been performed. Andukuri et al. (2011) developed a fibrous tubular structure with both endothelial cell binding peptides and polylysine NO donors that reduced platelet adhesion and SMC proliferation but increased endothelial cell proliferation in vitro. In a novel technique by Wang et al. (2015) the surface of their TEVG was modified with β-galactosidase to provide the ability to catalyze removal of the galactose group of an NO releasing prodrug (Wang et al., 2015). This technique can allow for greater stability of NO releasing materials injected into circulation. The study grafted the enzyme modified TEVG into a rat aorta and then provided a tail vein injection of the NO prodrug, resulting in increased von Willebrand factor expression from the endothelial cells but also a thicker layer of intimal hyperplasia than the control group without NO. There are also studies that have demonstrated NOS activity from cells seeded within a tissue engineered scaffold. For example, Lim et al. (2008) demonstrate that bone marrow-derived cells pre-seeded on a TEVG differentiated after grafting into the abdominal aorta of dogs and generated eNOS. Another study by McIlhenny et al. (2015) took the extra step of transfecting pre-seeded adipose-derived stem cells with the eNOS gene to decrease the thrombogenicity of their TEVG in a rabbit aorta model. However, most of the TEVG work with NO releasing materials involves in vitro studies, and the current TEVG approaches that have shown promise in large animal and clinical trials do not include NO releasing materials (Shojaee and Bashur, 2017). The lack of an established safe therapeutic dose in the body has been described as a current limitation for commercialization of these technologies (Jen et al., 2012). For tissue engineering applications, it will be important to determine how well the TEVGs with NO releasing molecules remodel and integrate in animal models of disease.

The incorporation of CORMs within vascular tissue engineering scaffolds has also been limited, and has only involved fibrous scaffolds generated through electrospinning (Bohlender et al., 2014; Michael et al., 2016). CORMs have also been delivered without a scaffold and incorporated within micelles (Peng et al., 2013), but micelles can be washed away. CORMs in scaffolds are the focus of the section below. The local release from the scaffold has the potential to provide controlled delivery to the vascular graft and the surrounding artery and reduce systemic effects. The effective dose can also be reduced because it bypasses the blood stream where the CO would bind to hemoglobin (Romaõ et al., 2012; Heinemann et al., 2014).

It is important that the CORMs can provide controlled release when incorporated within a tissue engineered scaffold. Fibrous scaffolds have several advantages as described previously (Bashur et al., 2013). In addition, these scaffolds can provide a hydrophobic carrier for the CORM materials with proper control of the scaffold composition. This is important for some CORM materials, such as the organic photoactivatable ones that our lab works with that need to avoid hydration to maintain their activity (Michael et al., 2016). The first incorporation of CORMs within a scaffold was performed by Bohlender et al. (2014). Interestingly, their goal with incorporating the CORM was to provide anti-bacterial properties through CO release (Bohlender et al., 2014). The photoactivatable Mn-based material provided the capacity for significant release of CO. In fact, the doses that they provided released CO so quickly that it generated bubbles that were toxic to the fibroblastic cells seeded on them. This is not a surprise since even bubbles composed of air that are generated by ultrasound technology can be toxic to cells (Wu et al., 2008). This study emphasizes the importance of dose control with use of CORMs for a tissue engineering application.

The first incorporation of CORMs within a scaffold for tissue engineering was performed by Michael et al. (2016). The CORM was blended with a hydrophobic polymer and electrospun to generate a fibrous mesh. Importantly, this mesh allowed a time frame in which vascular cells can be seeded prior to activation and then the construct was used for a tissue engineered vascular graft application. The time frame with the specific CORM tested in this study was only 1 h, and the strategy may need to be modified to provide a longer time frame. However, it is important that the cytotoxic effects noted in the Bohlender study were not observed with the lower CO doses and slower CO release that was used in the study by Michael et al. (2016). They did not find a significant decrease in SMC viability with 2% (w/w) CORM incorporation and activation. The SMCs were also shown to attach and express phenotypic markers. Currently, there is limited research on the impact of CORMs and CO dose that would explain the results of these studies. However, Nobre et al. (2016) evaluated the ability of CORMs to deliver CO to bacterial and mammalian cells and investigated the toxicity to mammalian cell for several CORM compounds. They found that most CORMs and the released CO were non-toxic to eukaryotic cells even at high concentrations that resulted in bactericidal activity. Specifically, the antimicrobial effects of the metal-based CORM (CORM-3) were noted at 50-fold lower concentrations than those that were toxic for eukaryotic cells (Desmard et al., 2009). This interesting finding suggests that CO may be able to simultaneously provide antibacterial properties as well as support cell function in vascular grafts at appropriate doses, although additional research is required. Finally, the impact of the inflammatory environment in the body on the tissue engineering use of CORMs has not yet been determined.

Gasotransmitters have shown promise for improving TEVGs and represent the most common approach to deliver pharmaceutical compounds with tissue engineered vascular grafts. However, these studies are in the early stages. One of the critical areas of research will be verifying the appropriate doses and potential side effects. Careful control of the dose and investigation of the pharmacodynamics are especially important for NO and CO delivering compounds because of their concentration dependent effects. This will include an assessment of the biocompatibility of the NO donor or CORM itself in case it leaches out of the scaffold, the released gas, and the degradation product after activation and release of the gasotransmitter. This is demonstrated by Nobre et al. (2016) where they showed that there was no correlation between the bactericidal activity and the ability of CORMs they tested to release CO into the medium. New CORMs have been developed that provide lower toxicity potential. As an example, the CORM used by Michael et al. (2016) is an α-diketones that has been shown to have limited cytotoxicity for KG1 myeloblast cells and rat SMCs within doses up to 40 μM tested in cell culture (Peng et al., 2013). The desired rate of CO release still requires further investigation by the CORM field. Finally, the degradation product of the Michael et al. (2016) CORM is anthracene, which has a low toxicity for aromatic compounds. However, detailed analysis of the toxicity profile will be need for any new anthracene derivatives or other CORM molecules that will be employed. Overall, more work is needed, but these delivery strategies may allow for generation of a viable tissue engineered vascular graft that would provide needed options for patients with coronary or peripheral artery disease.