**4. Discussion**

The results of the present study show that carbogen treatment in an acute short term setting resulted in a profound reduction of intraplaque hypoxia in murine vein grafts lesions in vivo. Long term treatment with carbogen resulted in a beneficial effect on vein graft patency in ApoE3\*Leiden mice, but surprisingly, had no effect on hypoxia, intraplaque angiogenesis and intraplaque hemorrhage. On the other hand, long term carbogen treatment resulted in hyperoxia-induced ROS formation with consequent effects on HIF1a mRNA levels and macrophage apoptosis. A reduction in macrophage content in the vein graft lesions was observed, resulting in less unstable lesions. Moreover, comparable to what was observed in vivo, in vitro induction of ROS using the ROS mimic t-BHP in BMM resulted in a strong increment in DNA damage and apoptosis.

Carbogen inhalation is widely used in the oncological field [20,21]. It has been shown that the time to achieve a maximal increase in tumor oxygenation with carbogen inhalation depends on various factors such as the type of cell involved, the location, and the size of the tumor [22,23]. Moreover, Hou et al. observed an effect of carbogen treatment comparable to what was observed in the present study, both in the short and the long term experiments. Single carbogen inhalation significantly increased tumor oxygenation, while during multiple administrations of carbogen the effect was reduced, indicating that the response to chronic carbogen is not consistent over days [22]. Nevertheless, we showed that prolonged carbogen treatment has a protective role against vein graft occlusions. Vein graft occlusion is a phenomenon often seen after vein grafting in which the vessel lumen is narrowed due to extensive intimal hyperplasia that progress to stenosis and occlusion [24]. This phenomenon is also observed in ApoE3\*Leiden mice that undergo vein graft surgery. Besides the reduction in vein graft occlusions, an increase in vein graft patency due to an increase in lumen perimeter and optimal lumen area of the hyperoxic vein grafts was observed, similar to the study by Fowler et al. [25]. In that study carbogen is used in the treatment of central retinal artery occlusion to increase blood oxygen maintaining oxygenation of the retina [25]. This effect of hyperoxygenation on retinal artery remodeling can be related to the effect of carbogen on patency and increase in lumen perimeter and increase in the optimal lumen found in the present study.

We did not observe a reduction in hypoxia nor an effect on intraplaque angiogenesis in the prolonged carbogen study. Furthermore, no changes in local gene expression of Vegfa were observed in the vein grafts, but interestingly Hif1a was upregulated in the prolonged carbogen exposure study and not downregulated as expected. In fact following our initial hypothesis we would have expected a reduction in intraplaque angiogenesis in parallel with a reduction in Hif1a and Vegfa expression. For this reason, we studied other known processes that regulate Hif1a and observed an accumulation of ROS in the carbogen exposed group when compared to the control group. Repeated exposure to hyperoxia is known to be associated at a cellular level with an accumulation of ROS [26,27]. When the exposure is repeated too often, the oxidant insult is no longer compensated by the host's antioxidant defense mechanisms and therefore cell injury and death ensue [28]. Cell injury induced by ROS comprises lipid peroxidation, protein oxidation and DNA damage [29,30]. We observed an increase in DNA damage measured as an augmented presence of 8OHdG staining in the long term carbogen treated group when compared to the control group, indicating that a daily long term treatment with carbogen gas results in accumulation of ROS that in turns induces DNA damage in the atherosclerotic lesions. Moreover, we also observed an increase in DNA damage in bone marrow macrophages in vitro under ROS stimulation. It is known that DNA damage can be found in the nuclei and in the mitochondria [31,32]. Both in the vein graft lesions in vivo and in the cultured t-BHP treated macrophages in vitro, a strong 8OHdG positive staining in the nuclei of the cells, with an occasional cytoplasmic staining could be seen. The subcellular location of the staining of 8OHdG suggests that the main site of ROS induced DNA damage is nuclear, and not mitochondrial. ROS generated by repeated hyperoxia treatment can alter gene expression by modulating transcription factor activation, like NF-kβ, which then impact downstream targets [33]. It has been shown that hyperoxia also results in nuclear translocation of NF-kβ and NF-kβ activation in several cell types [34]. Our results show that long term carbogen treatment result in Hif1a gene expression upregulation. In addition, in vitro BMM treated with the ROS mimic t-BHP also showed an upregulation of Hif1a gene expression. Interestingly, the transcription of this gene is known to be regulated by NF-kβ transcription factor. In fact, Bonello et al., demonstrated that ROS induced Hif1a transcription via binding of NF-kβ to a specific site in the Hif1a promoter [35]. Those findings could be further investigated in future experiments using antioxidants such as NAC to see whether it can reverse the carbogen treatment.

We showed that the accumulation of ROS in the carbogen treated group caused an increase in apoptosis, accumulated in macrophages rich areas, and resulted in a decrease in the amount of macrophages. Even though we cannot exclude that the association of macrophages with cleaved caspase 3 could be due to efferocytosis of apoptotic cells, macrophage efferocytosis is frequently hampered in atherosclerotic lesions, therefore it is likely that these macrophages are apoptotic. Previously, in contrast with our findings, a strong correlation between macrophage content and hypoxia was shown by Marsch et al. [15]. Moreover, hypoxia potentiates macrophage glycolytic flux in a Hif1a dependent manner [36] in order to fulfill the need of ATP for protein production and migration. Taken together, this points to a high request and high use of O2 by plaque macrophages and a consequent high exposure of these inflammatory cells to ROS accumulated during hyperoxia. We demonstrated that ROS causes accumulation of DNA damage and subsequently an increase in apoptosis and cell death in BMM in vitro. The link between ROS induced DNA damage and apoptosis detected in vitro might explain the observed apoptosis in macrophages in vivo. Moreover, a reduction in the number of macrophages is associated with plaque stability and plaque stability is reflected in an increase in vein graft patency as observed in the present study.

Previously Marsch et al. showed that repeated carbogen treatment in LDLR-/- mice lead to reduction in intraplaque hypoxia, necrotic core size and apoptosis [15]. In the present study we showed that repeated carbogen treatment in accelerated vein graft atherosclerotic lesions in ApoE3\* Leiden mice resulted in increased apoptosis and unaltered intraplaque hypoxia when compared to controls. Accelerated atherosclerotic lesions in ApoE3\*Leiden mice highly resemble human atherosclerotic lesions and, differently from LDLR-/- mice, do present intraplaque angiogenesis. Our results show that although we did not observe reduced intraplaque angiogenesis and IPH daily hyperoxia treatment with carbogen gas in this murine model lead to accumulation of ROS that could not be cleared by anti-oxidant agents and the ROS build-up lead to DNA damage and induced apoptosis. In fact, differently from Marsch et al., who treated mice daily for five days, followed by two days of no carbogen exposure we performed the treatment daily and started our treatment seven days after mice underwent vein graft surgery, when the atherosclerotic lesions already started forming. This starting time point was based on our previous findings [5] in which we found that intraplaque neovascularization in ApoE3\*Leiden mice that underwent vein graft surgery is visible 14 days after surgery. Therefore, we were able to study the e ffect of carbogen treatment on lesion stabilization rather than on lesion formation.

One of the limitations of the current study may be the choice of the model used, the ApoE3\*Leiden mice vein grafts. However, since in most mouse models for spontaneous atherosclerosis intraplaque angiogenesis is absent, and the lesions observed in the ApoE3\*Leiden mice vein grafts show many features that can also be observed in advanced human lesions, including intraplaque hypoxia, angiogenesis and intraplaque hemorrhage, we believe this model is suitable for the current studies. The fact that the most prominent e ffects observed relate to hyperoxygenation induced ROS production, macrophage apoptosis and vein graft patency, whereas the experimental set-up was initially designed to identify e ffects on intraplaque angiogenesis, might indicate another limitation in our study set-up.

Based on the results obtained in the present study we can conclude that although short term carbogen gas treatment leads to a profound reduction in intraplaque hypoxia, the treatment has mixed e ffects. Despite the beneficial e ffects of the hyperoxygenation treatment on vein grafts, i.e., improved vein graft patency and a strong trend towards an increased plaque stability index, chronic hyperoxygenation also induced Hif1a mRNA expression, ROS accumulation and apoptosis. That all will harm the vein grafts in the current model under the current conditions. This indicates that in order to define potential therapeutic benefits of hyperoxygenation treatment further research is needed to define optimal conditions for this treatment in vein graft disease.

**Supplementary Materials:** The following are available online at http://www.mdpi.com/2073-4409/9/2/336/s1, Figure S1: Bodyweight and cholesterol levels acute carbogen exposure, Figure S2: Bodyweight and cholesterol levels chronic carbogen exposure, Figure S3: Total wall gene expression, Figure S4: inflammatory gene expression in cultured BMM, Figure S5: BMM 8OHdG IHC.

**Author Contributions:** Conceptualization, L.P., M.R.d.V. and P.H.A.Q.; formal analysis, L.P. and M.R.d.V.; investigation, L.P., H.A.B.P., F.B. and M.R.d.V.; resources, M.R.d.V. and P.H.A.Q.; supervision, M.R.d.V. and P.H.A.Q.; visualization, L.P., M.R.d.V. and P.H.A.Q.; writing—original draft, L.P.; writing—review and editing, L.P., J.C.S., M.R.d.V. and P.H.A.Q. All authors have read and agreed to the published version of the manuscript.

**Funding:** This research was funded by a gran<sup>t</sup> from the European Union, Horizon 2020 MSCA joint doctoral project, MOGLYNET [Project number 675527].

**Conflicts of Interest:** The authors declare no conflict of interest.
