**3. Results**

#### *3.1. Acute Carbogen Exposure Reduces Intraplaque Hypoxia*

To evaluate the e ffect of acute carbogen treatment on advanced atherosclerotic vein graft lesions, ApoE3\*Leiden mice that underwent vein graft surgery were exposed for 90 min to carbogen gas or normal breathing air. Mice exposed to carbogen gas (*n* = 8) showed a significant reduction of intraplaque (IP) hypoxia in the vein graft lesion compared to the air breathing group (*n* = 8) as shown by a 84% decrease in the immuno-area positive for pimonidazole in the lesions of carbogen treated mice compared to the control (*p*-value = 0.027) (Figure 1A–C).

**Figure 1.** Short term carbogen exposure drastically reduces intraplaque hypoxia. (**A**) Representative pictures of sections from vein graft lesions in ApoE3\* Leiden mice stained for pimonidazole in the control group (*n* = 8) and (**B**) one-time carbogen treated group (*n* = 8). (**C**) Quantification of pimonidazole positive area. Data are presented as mean ± SEM. \* *p* < 0.05; by two-sided Student's t test.

Regarding the aspect of vein graft patency, single 90-min carbogen exposure directly before sacrifice did not affect vein graft patency (Figure 2A), vessel wall area, lumen perimeter, lumen area or optimal lumen area (Figure 2B,C). Furthermore, weight nor cholesterol levels were changed (Figure S1).

The percentage of collagen present in the lesion was comparable between the two groups (Figure 2D) and at a cellular level, the percentage of macrophages (Figure 2E) and SMCs (Figure 2F) were not altered by the acute exposure to carbogen.

#### *3.2. Chronic Carbogen Exposure Does not Influence Intraplaque Hypoxia*

To evaluate the effect of hyperoxic carbogen treatment on plaque composition and remodeling we performed a chronic carbogen treatment on ApoE3\*Leiden mice with advanced atherosclerotic vein graft lesions. Mice were exposed for 90 min daily to carbogen gas (*n* = 13) or normal breathing air (*n* = 12) for 21 days. Neither weight or cholesterol levels were affected by the treatment (Figure S2).

Surprisingly chronic exposure to carbogen gas did not reduce intraplaque hypoxia in the treated group when compared to the air breathing group (Figure 3). In fact, the degree of pimonidazole staining in the vein graft area was not different between the two groups (Figure 3).

#### *3.3. Chronic Exposure to Carbogen Plays a Protective Role Against Occlusions*

Chronic carbogen treatment resulted in a beneficial effect on vein graft patency, increasing the rate of vein graft patency by 34.5% (Figure 4A). In fact, only 53% of the mice of the control group presented a patent vein graft (Figure 4A), while 87.5% of the mice exposed to carbogen gas had a patent vein graft.

**Figure 2.** Short term exposure to carbogen gas does not influence plaque size nor composition. (**A**) Quantitative measurement of vein graft patency in ApoE3\*Leiden mice from the control and one-time carbogen treated groups. Data are analyzed by Chi-square test (**B**) Representative pictures of MOVAT staining of vein graft sections from control (*n* = 8) and carbogen group (*n* = 8). (**C**) Quantification of vessel wall area, lumen perimeter, lumen area and optimal lumen area. Percentage of positive vessel wall area and representative pictures for (**D**) collagen (*n* = 8 for control and carbogen groups), (**E**) macrophages (*n* = 8 for control and carbogen groups) and (**F**) smooth muscle cells staining (*n* = 8 for control and *n* = 7 for carbogen groups). Data are presented as mean ± SEM.

**Figure 3.** Chronic carbogen treatment does not affect intraplaque hypoxia. Representative pictures of vein graft cross sections stained for pimonidazole in control (*n* = 8) and chronic-treated carbogen groups (*n* = 13) and quantitative measurement of percentage of vessel wall area positive for pimonidazole staining. Data are presented as mean ± SEM.

**Figure 4.** Chronic exposure to carbogen plays a protective role against vein graft occlusions. (**A**) Quantitative measurement of vein graft patency in ApoE3\*Leiden mice from the control and prolonged carbogen treated groups. Data are analyzed by Chi-square test. \* *p* < 0.05. Representative pictures of non-patent and patent vein grafts in ApoE3\* Leiden mice at day 28 after surgery in the right panel. (**B**) Quantitative measurements of vein graft thickening. In the right panel, representative pictures of MOVAT staining in vein grafts from control (*n* = 8) and long term carbogen treated mice (*n* = 13), with vessel wall area as the area between the two dotted lines. (**C**) Quantification of lumen perimeter and (**D**) optimal lumen area. Data are presented as mean ± SEM. \* *p* < 0.05; by two-sided Student's t test.

Vessel wall area thickening was not affected by exposure to carbogen gas since no differences could be detected between the two groups when taken only the patent grafts into account (Figure 4B). More importantly lumen size was affected by carbogen gas. In fact, carbogen treated mice presented a significant increase in the lumen perimeter when compared to control (Figure 4C,E, *p*-value = 0.048), and an increase in the optimal lumen area (Figure 4D, *p*-value = 0.067).

#### *3.4. Chronic Carbogen Treatment Does Not Have an E*ff*ect on Intraplaque Angiogenesis and Intraplaque Hemorrhage*

To see whether exposure to carbogen gas had an effect on the hypoxia triggered IP angiogenesis, the amount of CD31+ vessels in the vein graft lesions (white arrows in Figure 5A zoom in) was evaluated and no difference in the number of neovessels in the carbogen group was observed when compared to the control group (*p*-value > 0.99).

**Figure 5.** Chronic carbogen treatment does not affect intraplaque neovascularization. (**A**) Representative pictures of vein grafts lesions stained for DAPI (white), CD31 (green) and Ter119 (red). (**B**) Quantification of CD31 positive neovessels in the vessel wall area in the control group (*n* = 8) and in the carbogen treated group (*n* = 12). (**C**) Bar graphs representing the quantitative measurements for IPH in the control and long term carbogen treated groups. IPH was scored as not present, low, moderate or high. Total vessel wall gene expression of (**D**) Hif1a, (**E**) Cxcl12, (**F**) Vegfa and (**G**) Hif2a, relative to Hprt, was measured in the control and long term carbogen treated groups. Data are presented as mean ± SEM. \* *p* < 0.05; by two-sided Student's t test.

In addition, when corrected for intimal thickness no di fferences were observed between the groups (*p*-value = 0.91) (Figure 5A,B). As a measure of the quality of the IP angiogenesis the degree of intraplaque hemorrhage was analyzed (yellow stars in Figure 5A zoom-in) as the amount of Ter119<sup>+</sup> cells found outside the neovessels and quantified as not present, low, moderate or high, and no di fferences could be seen when comparing the two groups (Figure 5C).

To determine the e ffects of hyperoxia on angiogenesis related genes the expression of Hif1a was analyzed. We surprisingly found a significant upregulation of Hif1a mRNA expression in the carbogen treated group when compared to the control (Figure 5D, *p*-value = 0.05), while mRNA expression of Cxcl12, Vegfa and Epas1 were not altered (Figure 5E–G). No di fferences between control and one-time carbogen treated group were found when analyzing gene expression in vein grafts from the acute carbogen treatment (Figure S3).

#### *3.5. Chronic Carbogen Treatment Induces Accumulation of Reactive Oxygen Species and Apoptosis*

Although an e ffect on Hif1a upregulation was observed, surprisingly no e ffect on angiogenesis could be seen. Therefore, we looked into other mechanisms that could possibly regulate Hif1a. We hypothesized that the mRNA upregulation of Hif1a in the carbogen treated group (Figure 5D) was caused by an accumulation of reactive oxygen species (ROS) induced by the carbogen treatment. ROS is known to be induced by prolonged hyperoxia [17] and to regulate the transcription of di fferent genes involved in hypoxia and in inflammation such as Hif1a and Il6.

Il6 mRNA expression was studied as a representative for ROS induced factors and quantification of its expression showed a trend towards increased expression in the carbogen group of the chronic exposure study when compared to the control group (Figure 6A, *p*-value = 0.09).

Next the presence of ROS was studied in the vein graft lesions by quantifying the amount of ROS–mediated DNA damage, analyzed by 8-hydroxy-2deoxy-guanosine (8OHdG) immunohistochemical staining. We determined the subcellular location of the staining of 8OHdG. As observed in Figure 6C, a strong 8OHdG positive staining was found in the nuclei of the cells, with an occasional staining outside of the nuclei in the mitochondria, seen as cytoplasmic staining (Figure 6C, right panel). This suggests that the main site of ROS induced DNA damage is nuclear, and not mitochondrial. 8OHdG positive staining could be seen as light blue staining in the nuclei of the cells as a results of co-localized DAPI and 8OHdG staining (Figure 6D zoom-in) and the quantification corrected for the vessel wall area resulted in an increase of DNA damage in the carbogen treated group when compared to the control group (Figure 6B,D), supporting the idea that ROS levels are increased.

ROS is known to induce apoptosis as a consequence of DNA damage. Therefore, the amount of cells positive for cleaved caspase 3 (CC3) in the atherosclerotic vein graft lesions was determined in the carbogen treated and in the control groups (Figure 6E). Due to their high oxygen consumption we hypothesized that macrophages could possibly be the main cell type a ffected by DNA damage induced by ROS and subsequent apoptosis. As shown in the bottom panel of Figure 6D, macrophages rich areas in the lesions of mice treated with carbogen were found to be strongly positive for CC3 when compared to control (Figure 6E bottom panel).

When looking at the total amount of cells positive for cleaved caspase 3 in the intimal area, an increase in apoptotic cells CC3+ in the lesions of mice exposed to carbogen gas was found compared to the air breathing group (Figure 6E, *p*-value = 0.06).

**Figure 6.** Chronic carbogen treatment induces accumulation of ROS. (**A**) Il6 gene expression relative to Hprt in the total vessel wall of control and chronic carbogen treated groups. (**B**) Quantification of the percentage of vessel wall area positive for 8OHdG. (**C**) Representative pictures of DAPI (in blue, left panel), 8OHdG (in green, central panel) and merged (right panel) staining in the vein graft lesions. (**D**) representative pictures of DAPI (blue) and 8OHdG (green) staining in vein graft lesions from control (*n* = 8) and carbogen treated (*n* = 12) mice. Light blue staining represents nuclei positive for 8OHdG and examples are indicated by white arrows. (**E**) In the top panels, representative pictures of DAPI (blue) and cleaved caspase 3 (CC3, in magenta) staining and in the bottom panels representative pictures of DAPI (blue), CC3 (magenta) and Mac3 (green) staining in control and carbogen groups respectively. In the right panel quantification of percentage of intimal area positive for cleaved caspase 3. Data are presented as mean ± SEM. \* *p* < 0.05 by two-sided Student's t test.

#### *3.6. Chronic Carbogen Exposure Reduces Inflammatory Cell Content*

The effects of chronic carbogen gas treatment on intraplaque inflammation were studied on macrophages since they produce high amounts of ROS, consume elevate amounts of O2 and are known to be hypoxic [6]. Interestingly, the group of mice exposed daily to carbogen for 21 days showed a significant reduction in macrophage content when compared to the control group breathing normal air (*p*-value = 0.0126).

When corrected for the differences in vein graft thickening, the relative percentage of macrophages was significantly decreased in the carbogen exposed group by the 15.2% (Figure 7A, *p*-value = 0.0044).

**Figure 7.** Chronic carbogen treatment reduces macrophages infiltration in the plaque. (**A**) Representative pictures of ApoE3\* Leiden mice vein grafts from control (*n* = 8) and chronic carbogen treated (*n* = 13) groups stained for Mac-3. In the right panel quantitative measurements of the percentage of vessel wall area positive for Mac-3. Data are presented as mean ± SEM. \*\* *p* < 0.01; by two-sided Student's t test. Total wall gene expression of (**B**) Ccl2 and (**C**) Tnf relative to Hprt. (**D**) Quantification of the number of cells positive for Mac-3 and Ki67 in the vessel wall area of control (*n* = 8) and carbogen treated group (*n* = 13) and representative pictures of the staining with DAPI presented in blue, Mac-3 in green and Ki67 in red. Data are presented as mean ± SEM. \*\* *p* < 0.01 by 2-sided Student t test.

ϮϬϬђŵ ZK'E

ϮϬђŵ

To study whether the decrease in macrophages was not due to a reduced infiltration of macrophages, nor a reduced proliferation of resident macrophages, local cytokines expression in the vein grafts was studied and the proliferation of macrophages was analyzed.

First, the mRNA expression levels of Ccl2 and Tnf in the vein graft atherosclerotic lesions were examined. The mRNA levels of Ccl2 and Tnf did not differ between the carbogen treated group and the control (Figure 7B,C).

Using a triple IHC staining for Mac-3, Ki-67 and DAPI the amount of proliferating macrophages was determined. As shown in Figure 7D and, there was no difference in the number of proliferative macrophages corrected for the vessel wall thickening (*p*-value = 0.16).

Thus, the data sugges<sup>t</sup> that the reduction of plaque macrophages could be due to enhanced macrophage apoptosis.

#### *3.7. Chronic Carbogen Treatment Does Not A*ff*ect Plaque Size but Increases Plaque Stability*

To evaluate the effect of prolonged carbogen treatment and accumulation of ROS on plaque composition, the amount of collagen (positive collagen area in the total vessel wall) and smooth muscle cells (positive αSMA area in the total vessel wall) was analyzed, two main predictors of plaque stability.

The collagen content in the plaque was not affected by carbogen treatment, and was comparable between the two groups (Figure 8A). Similarly, SMCs content in the carbogen group was not different from the control group (Figure 8B). Interestingly, when calculating the plaque stability index, defined as the amount of collagen and SMCs divided by the vessel wall area, atherosclerotic plaques of the mice daily exposed to carbogen resulted to be more stable than the lesion of the control group (Figure 8C, *p*-value = 0.05).

**Figure 8.** Chronic carbogen treatment does not affect collagen nor smooth muscle cells content in the lesion but increases plaque stability. (**A**) Quantitative measurement vessel wall area positive for collagen in ApoE3\*Leiden mice from the control (*n* = 8) and carbogen treated (*n* = 13) groups. In the right panel representative pictures for collagen staining. (**B**) Quantification of percentage of vessel wall area positive for smooth muscle cell actin and representative pictures from the control (*n* = 8) and carbogen treated (*n* = 13) groups. (**C**) Quantification of plaque stability index. Data are presented as mean ± SEM.

#### *3.8. ROS Increases DNA Damage and Apoptosis in Bone Marrow Derived Macrophages In Vitro*

To unravel the molecular and cellular mechanism underlying the observed changes in macrophage content, in particular whether this could be due to hyperoxia induced ROS accumulation, we treated macrophages derived from bone marrow of APOE3\*Leiden mice with t-BHP, a known ROS mimic [19]. t-BHP treatment increased the occurrence of DNA damage in BMM as measured by 8-OHdG immunocytochemical staining (Figure 9A), confirming its activity as a ROS mimic and the induction of DNA damage by ROS.

**Figure 9.** ROS induces DNA damage and apoptosis on in vitro bone marrow macrophages (BMM) (**A**) Representative pictures of CTRL BMM and BMM treated with 200 or 400 μm t-BHP respectively are shown. Examples images stained for DAPI (blue), Mac3 (green) and 8OHdG (magenta) as well as a merged image are shown per each condition tested. (**B**) Quantification of 8OHdG expression as mean intensity is shown. (**C**) Total mRNA expression of Hif1a relative to Hprt. (**D**) Quantification of CC3 expression as mean intensity is shown. (**E**) Quantification of total amount of cells per condition tested expressed as total amount of positive DAPI nuclei. (**F**) Representative pictures of CTRL BMM and BMM treated with 200 or 400 μm t-BHP respectively. Examples images stained for DAPI (blue), Mac3 (green) and cleaved caspase 3 (CC3) (orange) as well as a merged image are shown per each condition tested. Data are presented as mean ± SEM. \* *p* < 0.05, \*\* *p* < 0.01; by two-sided Student's t test.

Quantification revealed a 2.2-fold increase in DNA damage in macrophages treated with 200 μm t-BHP (*p*-value = 0.006) and a two-fold increase in DNA damage in macrophages treated with 400 μm t-BHP (*p*-value = 0.006) when compared to control (Figure 9B).

We then evaluated the effect of the ROS mimic t-BHP on the expression of several genes. Similar to changes in expression in vivo, we found that t-BHP-induced ROS caused a significant increase of Hif1a mRNA expression (*p*-value = 0.007 and 0.02 respectively) when compared to control (Figure 9C). Interestingly, we also found that ROS caused an increase in the expression of pro-inflammatory genes Ccl2 and Tnf, but decreased Epas1 expression compared to control (Figure S4)

To assess if ROS ultimately causes apoptosis in cultured BMM, we examined the expression of CC3 and found a significant and dose dependent increase in CC3 expression, thus apoptosis, in t-BHP treated BMM when compared to control (Figure 9F). The group treated with 200 μm t-BHP showed a 10% increase (*p*-value = 0.03) and the group treated with 400 μm t-BHP a 27% increase (*p*-value = 0.01) in CC3 expression when compared to control (Figure 9D). Moreover, we observed a drastic reduction in the total number of cells by 72% and 70% in the groups treated with 200 and 400 μm t-BHP, respectively, when compared to control (Figure 9E, *p*-value = 0.01 for both groups). Combined these data demonstrate that ROS directly affects gene expression in macrophages and causes DNA damage and apoptosis.
