In the present study, we have evaluated the presence of oxidants and anti-oxidants in experimental pulmonary hypertension, their enzymatic source, their impact on pulmonary cellular recruitment and vascular structures. We used a combined inflammatory-/increased flow-model of pulmonary hypertension, first described by Okada and co-workers [
21]. The main finding of our study, is that increased flow in MCT-induced pulmonary hypertension potentiates three pathological parameters: production of reactive oxygen species (ROS), remodeling of pulmonary arteries and hemodynamic deterioration. These findings provide new evidence for a vicious circle comprising increased, mast-cell induced OS in the range of high-flow challenged small pulmonary arteries, homing of immune-modulating cells to this "Achille's heel", and their contribution to vascular remodeling through inflammation.
Mast cell steered OS is increased in PH with additional blood-flow, after pneumonectomy
In our model of combined inflammatory/increased flow-induced PH, we observed significantly elevated expression of pro-oxidative enzyme NOX-4 in lungs of rats from the MCT/PE-group, as compared to controls and to the MCT-group. Pro-oxidative enzyme NOX-2 was significantly increased in both, MCT/PE and MCT. Concomitantly, mRNAs of anti-oxidative enzymes Mn-SOD and glutathion-peroxidase-1 were significantly elevated in whole lung tissue. Interestingly, we found a 1.5-fold decrease of anti-oxidant enzyme hemoxygenase-1 (HO-1) in animals with MCT-treatment and increased blood-flow, versus MCT-treatment only. In addition, mRNA-levels of catalase, another major anti-oxidative player were significantly reduced in both, lungs from the MCT- and from the MCT/PE-group. We tried to localize the increase of the two main pro- and anti-oxidative players NOX-4 and Mn-SOD, respectively, within the diseased lungs. Immunohistochemistry experiments allowed us to literally highlight a cell population, which we had not noticed in the first place:diseased lungs of the MCT/PE, and to a lesser extent of the MCT group displayed an overwhelming influx of large, cytoplasm-rich interstitial NOX-4+ and Mn-SOD+ cells, corresponding to tryptase+ mast cells. In addition to the wide-spread interstitial distribution, NOX-4+ and MN-SOD+ mast cells were observed in the range of pulmonary arteries and bronchioles. In an attempt to reveal a possible mechanism for the massive pulmonary homing of mast cells the recruitment pathway through the SDF-1/CXCR4 couple was positively tested: SDF-1 was expressed by endothelial cells of pulmonary vessels and by perivascular inflammatory cells, while immunofluorescent double-labeling revealed CXCR4-expression by NOX-4+ mast cells. In line with our observation, CXCR4 is known to be expressed by blood-derived mast cells and mast cell chemotaxis has been reported to be mediated by SDF-1 [
22,
23]. Our group has previously described the implication of CD117+ cells, comprising tryptase-positive mast cells and tryptase-negative bone-marrow derived progenitor cells in human PAH [
24], and increase of pulmonary mast cells has been described in human PAH [
25], as well as in experimental left-heart independent and left-heart-dependent PH-models [
26,
27]. Dahal and co-workers have elegantly shown in the monocrotaline rat-model, that mast-cell numbers are significantly increased and that preventive treatment with the specific c-kit inhibitor PLX and the mast-cell stabilizer cromolyn sodium salt (CSS) reduce mast-cell numbers and ameliorate pulmonary vascular remodeling, right-heart hypertrophy and hemodynamics [
26]. Interestingly, they were not able to attenuate these three indicators of PH in a therapeutic (non-preventive) fashion. When considering the vascular and peri-vascular localization of oxidative stress markers malondialdehyde and nitrotyrosine (see beneath), one seducing hypothesis appears to be a 'bow and arrow' setting, in which the mast cells (bow) will have produced enough ROS (arrows) to injure their distant target (pulmonary arteries) and trigger a self-supporting pro-oxidative and pro-inflammatory process, which will not be affected by any mast-cell dependant therapy.
Supporting our findings of increased oxidative imbalance in the MCT/PE model, hemodynamic stress induced by increased pulmonary blood flow has been previously proposed as a cause of superoxide anion production by oxidant enzymes, such as NOX-4, in cultured human endothelial cells [
18]. Of note, increased ROS production was observed only after cell exposure to pulsed, unidirectional flow, as compared to non-pulsed flow. Differentiation between pulsed and non-pulsed flow in the study of endothelial stress originates from an earlier work by De Keulenaer and co-workers [
28], where induction of NOX-4 activity is only observed after exposure of endothelial cell cultures to pulsed oscillatory shear stress, but not to non-pulsed unidirectional shear-stress. Originally, these data support the hypothesis of OS generation in plaque-prone regions of systemic atherosclerotic arteries. It is quite well established, that biomechanical forces exerted on the vessel wall by the flowing blood tend vary at these predilection sites due to oscillations in blood flow [
29]. As a consequence, laminar stress is reduced, whereas volume-dependent deformation of the vessel wall is enhanced. In fact, laminar shear stress is believed to protect arteries from developing intimal lesions, such as atherosclerosis, by maintaining endothelial cell nitric oxide synthesis and by enhancing the activity of anti-oxidant enzymes, e.g. HO-1. On the other hand, site-dependent cyclic vessel deformation is known to promote atherosclerosis through an increased formation of ROS and impairment of HO-1 [
30,
31]. Further more, it has been recently shown, that stimulation of HO-1 induces an improvement of endothelial dysfunction in spontaneously hypertensive Wistar-Kyoto rats by reducing OS and increasing NO availability [
32]. In a latest study, Yu and co-workers have shown that ROS overproduction and NOX-4 increase in a trauma/hemorrhage rat model can be prevented and trauma-impaired endothelium-dependent vaso-relaxation restored through stimulation of HO-1 expression [
33]. However, our findings also demonstrate deterioration of the oxidative balance through decrease of anti-oxidative enzyme catalase mRNA in both models as compared to controls, and of HO-1 mRNA in the MCT/PE group as compared to MCT alone. This raises the question of a possible mechanism for the downregulation of such protective enzyme-encoding genes. A plausible pathway for this deleterious phenomenon might be triggered by HIF-1 signaling: Loboda and co-workers have recently reported that induction of HIF-1 in human endothelial cells through hypoxia and dimethyloxaloylglycine attenuated the expression of IL-8 and of HO-1 through down-regulation of transcription-factor Nrf2 [
34]. The increased expression and important role of HIF-1 in human PAH and experimental PH has been addressed by several authors and is well recognized [
35]. It has been reported that experimental pneumonectomy elicits HIF-1 signaling [
36]; hence, a decrease of HO-1 and also of catalase would be plausible in our model, especially in the MCT/PE model. Further evidence of a connection between pulmonary arterial remodeling and oxidative/anti-oxidative balance alterations has been recently provided by Podlutsky and co-workers [
37]. The investigators tested several major enzymes of the oxidative/anti-oxidative system in F344-rats of different age (3 to 28 month old) in order to obtain an age-related profile on the presence of OS. As aging in the systemic circulation of the elderly is associated with generalized endothelial dysfunction and increased OS probably contributing to increased morbidity and mortality from cardiovascular diseases [
38], the rationale of the study was based on observations that pulmonary artery pressure and vascular resistance increase with normal aging in humans. The authors found, that aging in rat pulmonary arteries is associated with impaired acetylcholine-induced relaxation and vascular OS. Amongst others, expression of NOX-4 (mRNA) significantly increased in aged vessels, whereas expression of catalase significantly decreased. In contrast, expression of Mn-SOD, and glutathione peroxidase remained unaltered. Interestingly, this rodent enzymatic aging-profile of pulmonary arteries is congruent to our findings in permanent high flow-challenged monocrotaline treated rats, as compared to controls.
The increasing appearance of pulmonary artery adventitial fibroblasts (PAFB) in the setting of experimental hypoxic PH and their contribution to arterial remodeling has been reported in the past [
39]. In the context of human PAH, Li and co-workers have recently studied the expression of different NOX subunits in PAFB of human donors under normoxic and hypoxic conditions, as well as in IPAH-patients [
40]. Under hypoxic conditions NOX-4 was significantly upregulated at mRNA and protein levels. Silencing of NOX-4 by siRNA caused reduction of ROS levels under both normoxic and hypoxic conditions and suppressed the significant hypoxic-induced ROS increase. In addition, PAFB proliferation was significantly decreased in cells transfected with NOX-4 siRNA, whereas apoptosis was enhanced. A significant increase of NOX-4 mRNA expression was observed under hypoxic conditions in PAFB from the lungs with IPAH compared to healthy donors.
Further more, the role of NOX-4 in human pulmonary artery smooth muscle cells (HPASMC) has been the subject of a study by Ismail and co-workers [
41]. The investigators show that hypoxia induced HPASMC proliferation
in vitro is accompanied by increased reactive oxygen species generation and NOX-4 gene expression, and is inhibited by antioxidants, the flavoenzyme inhibitor diphenyleneiodonium (DPI), and NOX-4 gene silencing.
Our findings suggest that there may be an accentuated attenuation of the biological anti-oxidant defense system in MCT/PE-rats as compared to controls, but also as compared to monocrotaline challenge only. An imbalance between oxidative and anti-oxidative molecules could contribute to exuberant pulmonary arterial remodeling, since histomorphological and hemodynamic differences between our two PH-models were marked.
Markers of reactive oxygen species-activity and inflammation are overexpressed in pulmonary arteries of hypertensive rats
To test the hypothesis of direct ROS-involvement in accentuated flow-associated vasculopathy, we studied the expression of malonedialdehyde and nitrotyrosine in all groups by immunohistochemical experiments. These two molecules are established markers of ROS-activity on lipids, proteins and DNA, respectively [
42‐
44]. We found increased production of malonedialdehyde and nitrotyrosine in lung-tissue from MCT- and MCT/PE-animals, as compared to controls. The main sources of both markers were endothelial cells, an observation that has been previously made in lungs from patients with PAH [
20]. In addition but to a lesser extent, medial smooth-muscle cells stained for both markers. Of interest, malonedialdehyde was markedly produced by perivascular leucocytes of both pulmonary arteries and veins, suggesting a self-supporting stress circuit through recruited inflammatory cells. This possibility is supported by a recent experimental study from De Miguel and colleagues [
45]. The authors report an increased infiltration of T-lymphocytes into the kidney of rats with experimental salt-sensitive arterial hypertension and an increase of NOX-subunits within the attracted lymphocytes associated with an increase of OS-markers in the kidney. These pro-oxidative effects were attenuated by preventive immune-suppression through tacrolimus, resulting in a decrease of lymphocytic infiltrations and eventually decrease of arterial hypertension. This appears of particular interest regarding the close association of inflammatory elements and pro-oxidative activity in our experimental PH-model: as compared to MCT-animals, hypertensive rats with increased blood flow and maximal inflammatory infiltrates within the pulmonary arterial wall exhibited a higher number of remodeled vessels with production of OS markers. Cytokines IL-6, RANTES and fractalkine were overexpressed in MCT/PE and to a lesser extent in MCT, indicating a proinflammatory state within lung-tissue. IL-6 expression was localized to pulmonary arterial endothelium, while the two latter proteins were mostly expressed by perivascular leucocytes in our model.
Our previous articles have highlighted an overproduction of IL-1, IL-6, CCL2 (MCP-1) CXCL1 (fractalkine) and CCL5 (RANTES) [
46]. Recently, Furaya et al [
47] proposed a hypothetical mechanism leading to pulmonary vascular remodeling via overexpression of IL-6. IL-6 induces proliferation and anti-apoptosis in vascular smooth muscle cells through upregulation of VEGF, and downregulation of BMPR2 and TGFβR2. Upon IL-6 exposure, endothelial cells undergo apoptosis through repressed Tie2 signaling via downregulated Ang-1 expression in smooth muscle cells. Production of CX3CL1 results in recruitment of inflammatory cells, such as lymphocytes and monocytes, which produce enormous amounts of IL-6, while vascular smooth muscle and endothelial cells also produce IL-6 upon stimulation with IL-6. This hypothesis is confirmed experimentally as interleukin-6 overexpression induces pulmonary hypertension in mice [
48]. All the inflammatory mediators we have shown to be overproduced in PAH are linked to this key cytokine. Indeed, in mouse serum, mouse CCR2 protein (the receptor for CCL2) is necessary for expression of mouse IL6 protein that is increased by experimentally induced sepsis in mouse [
49], and IL1 protein increases expression of CCL2 protein [
50]. IL-1 also increased directly IL-6 mRNA levels by a protein kinase C-independent mechanism [
51] and IL-1 receptor antagonist treatment reduces pulmonary hypertension generated in rats by monocrotaline [
52]. Moreover, IL-1 may impair the therapeutic effects of prostacyclin analogues such as iloprost and carbaprostacyclin by attenuating cyclic AMP production by human pulmonary artery smooth muscle cells in response to these drugs [
53]. At last, in dendritic cells, mouse CCL5 protein increases expression of mouse IL6 mRNA [
54].
Further more and in accordance with previous reports [
55], we found decreased expression of PH-associated genes BMPR2 and of BMPR1A in both, MCT and MCT/PE, underlining the molecular similarities of those two animal-models with human familial or idiopathic PAH when it comes to alterations of BMP/TGF-β signaling.
Increased blood flow leads to modified arterial wall proportions in MCT-induced PH
The combination of MCT and pneumonectomy in young rats as a model for pulmonary hypertension has been studied for the first time by Tanaka and colleagues [
56]. The authors describe excessive intimal and medial remodeling of distal pulmonary arteries in those animals, as compared to MCT alone, which is nearly absent in animals undergoing pneumonectomy alone. In a more recent study, Homma and co-workers utilized a MCT-penumonectomy model in order to test preventive and curative effects of dehydroepiandrosterone in experimental pulmonary hypertension [
57]. One week after MCT injection, left pneumonectomy was performed and sacrifice was done 4.5 weeks later (day 40). PH was accompanied by severe pulmonary vascular remodeling, consisting of small pulmonary arterial medial wall thickening, increased adventitial cellularity and to a lesser extent arteriolar neointimal lesions.
Indeed, microscopic morphometrical evaluation revealed significant differences between our two pulmonary hypertensive groups. First, excessive muscularization of arterioles of less than 50 μm in external diameter was observed in both, MCT- and MCT/PE-animals. However, this phenomenon was quantitatively more pronounced in the latter group. In addition to arteriolar muscularization, the MCT/PE group displayed discrete intimal thickening with increased cellularity in small arterioles, as it has been described by Tanaka and by Homma. This observation might be of importance in the context of an experimental PH-model, since most PH-models generate medial thickening and adventitial remodeling, but lack intimal lesions, a hallmark of human PH.
Our results are Further more consistent with recent data showing the consequences of right-sided lobectomies in a rat model indicating that arteriolar muscularization occurs in rats with triple-lobectomy (quasi equal to pneumonectomy), but not in double-lobectomy [
58]. The authors describe a muscular thickening of pulmonary arterioles in between an external diameter of 30 to 80 μm, while remodeling of pulmonary arteries of over 100 μm is not reported. Interestingly, our study reveals qualitative differences of arterial remodeling between the MCT- and the MCT/PE-group, concerning pulmonary arteries with an external diameter ranging from 101 to 450 μm. Medial and adventitial thickening in MCT/PE-animals exceeded values in MCT-treated rats significantly. In fact, exuberant adventitial cellularity was observed to a smaller extent in MCT, while consistently present in MCT-PE. In accordance with excessive wall dimensions, MCT/PE exhibited significantly lower surface areas of arterial lumina than MCT, underlining the obstructive character of vascular remodeling.
Dendritic cells and CD68+cells are massively recruited to remodeled arteries of hypertensive rats with increased blood-flow
We have previously demonstrated that DCs may be involved in hypertensive pulmonary vasculopathy, since they are recruited to arterial lesions in experimental and human PH. DCs are mainly known as antigen presenting cells connecting the innate immunity to the adaptive immune response when facing danger signals and inducing tolerance to self-antigens in a non alerted immunitary state [
59]. Under inflammatory conditions, DCs orchestrate the immune response through activation and initialization of T-lymphocytes, but may also communicate with B-cells and fibroblasts [
60]. In previous reports, DCs have been demonstrated to organize arterial inflammatory lesions in systemic atherosclerosis [
61]. Noteworthy, a recent study has demonstrated an increased DC resistance towards organic and hydrogen peroxides through overproduction of manganese superoxide dismutase [
62]. In systemic arteries DCs accumulate in regions displaying hemodynamic stress by turbulent flow conditions and being prone to hypertensive intimal lesions [
17]. We observed intense recruitment of DCs into the wall of affected pulmonary arteries in MCT/PE-rats and to a lesser extent in MCT-rats. DCs were uniformly present within the dense adventitial cell-accumulations in MCT/PE and frequently infiltrated medial smooth-muscle-cells, up to subintimal regions. Noteworthy, in all groups, including control animals, numerous DCs were detected within BALT, while interstitial DCs were sparse. This finding is supported by previous observations mainly locating myeloid dendritic cells within the range of isolated pulmonary lymphoid follicles and BALT in the context of pulmonary interstitial diseases, in particular rheumatoid arthritis (RA) and Sjögren's syndrome (SS) [
63,
64].
In MCT/PE-animals, adventitial thickening mainly consisted in cells staining for CD68. However, in arteries presenting intense adventitial cellularity, we detected a pronounced medial infiltration by CD68
+/vimentin
+ cells. These cells, to some extent, morphologically displayed a fusiform pattern and were uniformly arranged within the muscular compartment, explaining difficult differentiation on HE-stained slides. Frid and co-workers have reported the arterial recruitment of mesenchymal precursors of a monocyte/macrophage lineage, including circulating fibrocytes, in hypoxic lungs of calves and rats [
65]. Recently, Burke and colleagues have reported influx of immune cells, including dendritic cells, into the pulmonary arterial walls of Wistar-Kyoto rats during sustained hypoxia, and concomitant increase of pro-inflammatory cytokines, growth factors and adhesion molecules [
66]. More specifically, the investigators observed an increase in the appearance of OX62+ dendritic cells in the peri-adventitial region of vessels from chronically hypoxic animals on days 7 and 28. Interestingly, it has also been shown that activated macrophages express vimentin only in proinflammatory conditions under the influence of proinflammatory cytokines and that macrophagic vimentin secretion is an important source of oxidative metabolite generation [
67]. These features would support a self-supporting concept of inflammatory action/reaction and OS.