Acute respiratory distress syndrome (ARDS) is a life-threatening disease characterized by an acute-onset, progressive, hypoxic condition with radiographic bilateral lung infiltration, e.g., due to various pathogens, but not due to hydrostatic pulmonary edema. Thus, acute respiratory distress syndrome (ARDS) results in a highly impaired pulmonary function and hypoxemia [1
]. A draft definition proposed three mutually exclusive categories of ARDS based on the degree of hypoxemia: mild (PaO2
≤ 300 mmHg), moderate (PaO2
≤ 200 mmHg), and severe (PaO2
≤ 100 mmHg). ARDS is most often due to direct injury to the lung (e.g., pneumonia and gastric aspiration) or indirect injury (e.g., sepsis and pancreatitis) [2
]. However, ARDS can also develop following trauma or inhalation of toxic gases like ozone [3
]. Thus, ARDS is characterized by a strong inflammatory response and impaired oxygenation due to a ventilation-perfusion mismatch [4
], resulting in pulmonary vasoconstriction caused by the Euler-Liljestrand mechanism. This mechanism, while being helpful in compensating for regional ventilation abnormalities, aggravates hypoxemia in ARDS due to a further increase in pulmonary artery pressure. Of note, in ARDS patients severe pulmonary hypertension can even lead to right heart failure and death. Whereas the Euler-Liljestrand mechanism in ARDS increases pulmonary artery hypertension and hypoxemia, counter regulating mechanisms are activated as well [5
]. Key regulators of the hypoxic response are the so-called hypoxia inducible factors (HIF) and HIF degrading prolylhydroxylases (PHD). HIF and PHD are of importance both in hypoxia and inflammation i.e., important during human hypoxemia and profound inflammatory response in ARDS [6
]. In particular, the HIF-2 pathway is involved in a multitude of biological processes impacting on pulmonary hypertension, erythropoietin synthesis, iron metabolism, bone marrow microenvironment, and tumor progress [8
]. Alterations of constitutively activated HIF-2α are associated with an increased risk for neuroendocrine tumors [12
]. Furthermore, HIF-2α is implicated in the pulmonary regulation of thrombo-spondin-1 and contributes to pulmonary artery hypertension-driven vascular remodeling and vasoconstriction [14
]. Thus, differences in HIF-2α activity or HIF degrading prolylhydroxylases might influence the pulmonary artery hypertension counterregulating effects of hypoxemia, and thus alter patients’ outcome.
Recently, functionally active genetic variants were identified both in the HIF-2α
(C/T; Single Nucleotide Polymorphism SNP C/G [ch2: 46441523(hg18)] and the PHD2
genes (C/T; SNP rs516651 and T/C; SNP rs480902). These genetic variants were found to be associated with erythrocytosis, pulmonary hypertension, and chronic mountain sickness, respectively [10
]. However, it is unknown whether these genetic variants impact on the outcome in ARDS patients.
Accordingly, we tested the hypotheses that (1) SNPs in HIF-2α or PHD2 genes are common in Caucasians, and (2) they are an independent risk factor for 30-day mortality in ARDS.
In this study, we show that the functionally active PHD2
SNP rs516651 [18
], located in the key pathway for the hypoxic-inflammatory response, is associated with increased 30-day mortality in ARDS patients. In contrast, the PHD2
SNP rs480902 is not. Furthermore, the HIF-2α
SNP [ch2: 46441523(hg18)] GG-genotype was neither present in our ARDS patients of Caucasian heritage nor in healthy Caucasian blood donors.
Hypoxia-inducible factors and HIF degrading PHDs are key regulators of the human response to low ambient oxygen [19
] and also provide a link between hypoxia and the inflammatory response [9
]. HIF degrading prolylhydroxylase 2 SNPs are known to be associated with erythrocytosis in humans [10
] and PHD2
deficiency evokes erythrocytosis by activating the renal erythropoietin pathway in mice [20
]. Furthermore, the lack of PHD2
leads to defective vascular growth [21
]. Moreover, recent studies have shown that genetic variants in HIF-2α
gene and PHD2
influence the hypoxic inflammatory response, resulting in altered heart rate, arterial oxygen saturation, incidence of chronic mountain sickness or erythrocytosis [6
Our study is the first to analyze the occurrence of these three genetic variants (HIF-2α SNP: rs46441523; PHD2 SNP rs516651 and rs480902) in Caucasians with ARDS and their potential association with death from ARDS.
T-allele carriers of the PHD2
SNP rs516651 are quite common in Caucasian ARDS patients (TT-genotype: 2.3%; CT-genotype 18.9%) and with homozygous TT-genotype carrying a significantly greater mortality compared to CC-genotypes. As homozygous TT-genotypes were all male patients and older compared to the other genotypes, in a next step we performed a multivariate Cox-regression analysis, including age, gender, need for dialysis as well as other clinical confounders known to carry a higher risk for death from ARDS to our analysis. Thus, even when adjusting for these variables, the TT-genotype of the PHD2
SNP rs516651 is an independent predictor for 30-day mortality in ARDS. Of note, increased pulmonary artery pressure, often present in ARDS patients and to a varying extent due to hypoxic pulmonary vasoconstriction (the Euler-Liljestrand-mechanism), also is a risk factor for pulmonary edema in high-altitude residents [24
]. Of note, genotyping of high-altitude residents adapted to hypoxia, like Tibetans, did not reveal the presence of T-alleles, possibly due to an evolutionary selection bias [15
]. Thus, the PHD2
SNP rs516651 polymorphism may impact on adaptation to hypoxia both in high-altitude-residents and patients with ARDS. However, further studies are needed to confirm our results and to analyze the underlying pathomechanisms. Showing that genetic variants depict a specific phenotype not only in high-altitude residents but also in ARDS patients is crucial before finally judging the relevance of this particular SNP on ARDS pathogenesis.
We also analyzed a further genetic variant in the PHD2
gene (SNP rs480902) but this genetic variant did not impact on ARDS mortality in our cohort. Wu et al. showed a correlation between single nucleotide polymorphisms in hypoxia-related genes like the PHD2
(SNPrs480902) and susceptibility to acute high-altitude pulmonary edema (HAPE). They found that the HAPE cases had a significant higher T-allele frequency than the control group [25
]. Buroker et al. described phenotypical differences between Han Chinese with acute mountain sickness (AMS) in the rs480902 SNP. They found a significant correlation between (rs480902) SNP (C/T) and heart rate, arterial oxygen saturation of hemoglobin, and the hematocrit in the AMS-group. The CC- and TT-genotypes had a significantly higher heart rate compared to the CT-genotype while patients with the CC-genotype had a significantly greater arterial oxygen saturation of hemoglobin than those with either the CT- or TT-genotypes in the Han AMS study group. AMS Chinese with CT- and TT-genotypes had a significantly higher hematocrit than those with CC-genotypes. It seems that the various genotypes have different mechanisms of compensation in hypoxia. In our study, however, we did not find such differences in the ARDS cohort.
However, mice with PHD2
-deficiency show increased angiogenesis due to upregulated vascular endothelial growth factor-A (VEGF-A) serum concentrations. Furthermore, erythropoietin, which stimulates angiogenesis and erythropoiesis, was dramatically overexpressed in PHD2
-deficient mice. Thus, PHD2
is a major negative regulator for vascular growth. Additionally, PHD2
knockout in mice inhibits tumor necrosis factor α (TNFα) and intercellular Adhesion Molecule 1 (ICAM-1) expression and decreases both cell apoptosis and macrophage infiltration [21
]. Even when PHD2
alterations impact on the inflammatory response in mice or the hypoxic response in high-altitude residents, analyzed PHD2
genetic variants did neither impact on serum inflammatory variables like procalcitonin (PCT) or C-reactive Protein (CRP) serum concentrations nor on 30-day mortality from ARDS in our patient cohort. It is known that both hypoxia and inflammation alter the hypoxic-inflammatory response, thus alternative pathways might be induced and could be an effect of those SNPs. To further elucidate this analysis of the genotype dependent PHD2
protein activity and target gene expression in patients with and without ARDS should be done in a subsequent study.
Third, we analyzed the frequency of the HIF-2α
gene genetic variant rs46441523. Surprisingly, GG-genotypes did not exist in our Caucasians with ARDS, and only a single (surviving) individual carrying a CG-genotype was observed. Since the frequency distribution of alleles in ARDS patients and blood donors were similar, these data indicate that, at least in our region, this HIF-2α
polymorphism does not exist in Caucasians either with or without ARDS. It is already know about the HIF-2α
polymorphism that it changes very fast. Between Tibetan and Han samples one SNP at Endothelial PAS domain-containing protein 1 (EPAS1) shows a 78% frequency difference, representing the fastest allele frequency change observed at any human gene to date [26
]. Thus, further studies are warranted to analyze allele frequency distributions in different ethnicities and diseases.
Our study has limitations. First, although we included as many as 272 ARDS patients, our cohort might still be considered small. Second, we included ARDS patients over a long time period, and systemic changes in patient care cannot be entirely ruled out. However, this likely is of minor importance as changes in the standard of care would have influenced all patients similarly irrespective of their genotype, and the intensity in charge was blind to specific patients’ genotypes. In fact, considering this timing, the finding that the homozygous TT-genotype of the PHD2 SNP rs516651 carried a significantly greater mortality compared to the CC-genotype is even more robust. However, the most important limitation of our study is that we cannot provide mechanistic explanations. Thus, further studies are necessary to analyze genotype-dependent immune cell function, intracellular signaling cascades, and pulmonary vascular tone in ARDS.
The rs480902 SNPs did not meet the Hardy-Weinberg equilibrium, with a p Value of 0.01. According to Chen, deviations from the Hardy-Weinberg equilibrium proportions suggest that at least one of the standard underlying assumptions for the test (non-overlapping generations, large population size with random mating, no mutation, no migration, and no selection) may be violated. Thus, the distribution of this genetic variant should be analyzed in a larger cohort.