Heme oxygenase (HO) breaks down heme, yielding equimolar amounts of biliverdin, iron, and carbon monoxide (CO). Heme oxygenase induction protects in experimental models of inflammation, making the pathway a promising target in sepsis and organ failure (1). Among the isoenzymes cloned to date, only HO-1 can be induced primarily by oxidative stress (2, 3).
The gene encoding HO-1 (gene HMOX1, Ref_mRNA NM_002133) is located on chromosome 22 (4) and is organized by five exons (5). Of major interest is a gene-associated microsatellite (GT-dinucleotide repeat), which is highly polymorphic (6). It is reported that this microsatellite is located within the promoter and that variants are associated with different expression levels (7). Resulting genotypes are associated with severity, e.g., in critical illness, acute respiratory distress syndrome (ARDS), myocardial infarction, and failure of kidney and liver grafts (8–11). This microsatellite has been reported to be in linkage disequilibrium with a single-nucleotide polymorphism (SNP) rs2071746 A(-419)T (11).
Reported protective effects of upregulated HO activity are restricted to a narrow window of overexpression because all metabolites are potentially toxic (1). Fate of biliverdin, which is further metabolized by biliverdin reductase (BVR) to yield bilirubin, might contribute to net effects on outcome. Two constitutively expressed BLVRs, A (BVR-A, gene BLVRA, Ref_mRNA NM_000712) and B (BVR-B, gene BLVRB, Ref_mRNA NM_000713), located on chromosomes 7 and 19 convert biliverdin to bilirubin, and numerous SNPs are annotated (12, 13). Information regarding their impact on outcome is scarce (14, 15) and has not been studied in the critically ill. Recent evidence suggests, that both free heme (16) and higher-order metabolites, such as bilirubin oxidation end products (17), might contribute to end organ failure. Thus, in addition to HMOX1 and its well-established association with outcome under stress conditions, the BLVRs warrant attention.
Thus, we tested the hypothesis that genetic variation in the heme degradation pathway (either within the HMOX1 or BLVR gene loci) influences outcome in patients with severe sepsis.
MATERIALS AND METHODS
Cohorts and data collection
Blood samples were collected during the randomized, prospective, controlled, open-label, multicenter VISEP trial (VISEP) (18). After approval by the ethics committee, all patients or legal surrogates gave informed consent for the main trial with or without consent for additional genetic analyses. Sepsis-related Organ Failure Assessment (SOFA) scores were calculated daily to assess morbidity (19).
An independent multicentric cohort of septic patients obtained by the Hellenic Sepsis Study Group (HSSG; www.sepsis.gr) was used for replication. Samples used in this analysis came from four centers of the HSSG, and the protocol was approved by the ethics committees of the participating hospitals. Written informed consent was provided by the patients or next of kin.
Patients admitted to Jena University Hospital for elective abdominal surgery as well as healthy volunteers served as controls for both genetic analyses and HO-1 plasma levels. After written informed consent was provided, blood for genotyping and plasma samples were taken. Genotyping and HO-1 enzyme-linked immunosorbent assay (ELISA) were performed for all cohorts in the same laboratory by personnel blinded to the clinical phenotype.
DNA extraction and polymerase chain reaction
For DNA extraction from the collected EDTA blood, the QIAmp DNA Blood Kit (catalog no. 51104; Qiagen, Hilden, Germany) was used. Purified DNA was dissolved in water and immediately frozen until analysis.
Polymerase chain reaction (PCR) reactions were performed in 96-well microtiter plates, by using one ready-to-go PCR bead (Amersham, GE Healthcare Europe, Freiburg, Germany), 18 μL of sterile water, 1 μL of 10 μM forward and reverse primer each, and 5 μL (∼150–200 ng) DNA. Polymerase chain reaction was carried out on a Biometra-Cycler (Biometra, Goettingen, Germany) under the following conditions: one cycle for 5 min at 93°C, five cycles: 30 s at 95°C, 30 s at 59°C, and 1 min at 72°C, followed by 31 cycles: 30 s at 95°C, 30 s at 61°C, and 1 min at 72°C, and a final extension step of 1 h at 72°C and 2 min at 18°C. After PCR, Amplicon quality and correct product size were controlled on a 1% agarose/ethidium bromide gel using 5 μL of the PCR product.
Primer pairs for (GT)n microsatellite, rs2071746, rs699511, rs699512, rs10234057b (exon 2 of BLVRA) genotyping, rs1802846 (exon 4 of BLVRA) genotyping, rs45460395 (last exon of BLVRB) genotyping:
HMOX1-GT-f 5′-ATCAAGTCCCAAGGGGACAG-3′ as the forward and the FAM-labeled reverse primer HMOX1-GT-r 5′-GAAAGTGGGCATCAGCTGTT-3′, resulting in a PCR product size of around 300 base pairs (bp) depending on the (GT)n repeat length (300 bp obtained by in silico PCR with Human March 2006 Assembly [hg18] where a 32-unit repeat is annotated).
- 2. rs2071746 (HMOX1) genotyping:
Forward primer HMOX1-SNP-f 5′-CAGTTGTAGGGATGAACCATGA-3′ and reverse primer HMOX1-SNP-r 5′- AGGCTCTGGGTGTGATTTTG -3′, resulting in a PCR product size of 323 bp.
- 3. rs699511, rs699512, rs10234057b (exon 2 of BLVRA) genotyping:
Forward primer BLVRA_ex2-f 5′-GCACCAGTTTTGGGAAGTG-3′ and reverse primer BLVRA_ex2-r 5′-CACCAAGGAGACAATGCAGA-3′, resulting in a PCR product of 706 bp.
- 4. rs1802846 (exon 4 of BLVRA) genotyping:
Forward primer BLVRA_ex4-f 5′-TAGCCTGGAAGTGGGAGAGA-3′ and reverse primer BLVRA_ex4-r 5′-AGCTGGCACTTCCCTGTTAG-3′, resulting in a PCR product of 748 bp.
- 5. rs45460395 (last exon of BLVRB) genotyping
Forward primer BLVRB_last-f 5′-TGAATGAATGCATAACAAGCAA-3′ and reverse primer BLVRB_last-r 5′-TGAGCACCTGCATATGGAAA-3′, resulting in a PCR product of 506 bp.
SNP genotyping and fragment length analysis
Single-nucleotide polymorphisms were genotyped by direct Sanger sequencing of PCR products using both PCR primers to obtain sequence information from both strands. Genotypes were called by manually inspecting the sequence reads.
Five microliters of the final PCR product was mixed up 10 μL deionized formamide and 0.5 μL Gene Scan ROX-500 standard (Applied Biosystems, Carlsbad, Calif) and loaded on an ABI sequencer 3730 for capillary gel electrophoresis. Fragment length of each PCR reaction was analyzed using Gene Mapper 4.0 software (Applied Biosystems) after calibrating on known cloned fragment length of either 20, 30, or 43 (GT)n repeat units.
A human genomic DNA pool (Roche, Mannheim, Germany) served as control for (GT)n microsatellite distribution.
HO-1 plasma concentration
Heme oxygenase 1 plasma concentrations were measured in healthy controls and patients with severe sepsis within 24 h after sepsis onset (day 1) and on day 5 using human HO-1 ELISA kit (Enzo Life Science, Loerrach, Germany) according to the manufacturer’s instructions.
All data are reported as mean (SD), mean and 95% confidence interval, or absolute and relative frequencies. One-way analysis of variance, χ2 test, or Fisher exact test were applied for comparison of continuous or categorical data, as appropriate. The Bonferroni-Holm method was used for adjustment of multiple testing. Overall survival was estimated by the Kaplan-Meier method and compared by log-rank test. P ≤ 0.05 was considered statistically significant. Assuming a mean SOFA of 7.8 (SD, 3.84) and a 28-day mortality of 25%, a difference of 1.2 SOFA points and absolute difference of 14% in mortality associated with a given SNP could be observed with a power of 90%. Statistical analyses were performed using SPSS version 13 (SPSS Inc., Chicago, Ill) and SAS software version 9.1.3 (SAS Institute Inc., Cary, NC).
Patient cohorts and characteristics
Of 537 patients belonging to the intent-to-treat population of the VISEP trial, in 430 approval to use DNA samples was obtained (thereof, 421 with data on (GT)n repeat length and 419 with data on HMOX1/BLVRA/BLVRB SNPs). A cohort of 398 DNA samples from the HSSG was available for replication, and both cohorts fulfilled the criteria for severe sepsis/septic shock (20) on inclusion. Less than 1% of patients were nonwhite in both cohorts. Whereas the cohorts were comparable regarding Acute Physiology and Chronic Health Evaluation scores, patients of the Greek cohort were older and had a higher share of pulmonary and urogenital infections but less abdominal infections. Mortality of the HSSG cohort was approximately 20% higher, and patients were only partially treated in an intensive care unit (ICU) (Table 1).
Analysis of variability within the (GT)n repeat length of HMOX1 and coherence between (GT)n repeat genotype, organ dysfunction, and mortality
According to the trimodal distribution of the HMOX1 (GT)n microsatellite, lengths were grouped into subclasses: short (s) alleles, defined as GT-repeat length below 27 units; alleles with middle (m) length, defined as 27 to 33 units; and long (l) alleles included GT-repeat lengths of 34 and more units.
The distribution of the (GT)n repeat length frequencies within the study group is shown in Figure 1. One patient categorized (l/l) died on the first day. The pattern of the (GT)n repeat length distribution differed between the VISEP patients and n = 86 healthy controls. Although (GT)23-T and (GT)30-A were the most common haplotypes in both cohorts, their frequencies differed significantly (P = 0.002), preferring the (GT)30-A haplotype in VISEP patients (Fig. 1).
Patients with an m/m genotype did not show statistically significant higher mean SOFA scores during the ICU stay compared with other patient groups (P = 0.076). Regarding mortality, m/m patients had increased 28-day mortality rates compared with the other genotypes (P = 0.033) (Table 2 and Fig. 2). However, there was no difference in 90-day mortality rates.
Variability of SNP rs2071746, the human HMOX1 gene, and association of HO-1 plasma concentrations and rs2071746 genotypes
The T allele frequencies of 40.2% and 40.3% in the German and Greek cohorts, respectively, are lower than reported for the HapMap-CEU population (45.8%, ss8295263). Genotype distribution indicating deviation from the Hardy-Weinberg equilibrium by the χ2 test was tested using the method of Rodriguez et al. (21). Although below the 5% limit indicating genotype data at random (corresponding to a χ2 > 3.84), in both cohorts, χ2 values of 0.43 (VISEP) and 2.97 (HSSG) indicate some kind of ascertainment.
Details of rs2071746 genotyping results are depicted in Table 1. Carriers of two A alleles had higher 28-day mortality rates (P = 0.047) compared with the other genotypes of this SNP (Table 1 and Fig. 2) in the VISEP cohort. Whereas the A allele was observed more frequently than would be expected according to the Hardy-Weinberg equilibrium in both cohorts, the additional association of A-homozygocity with unfavorable outcome was restricted to the VISEP cohort. Regarding the 90-day mortality rates, no association of rs2071746 and outcome could be found.
Mean HO-1 plasma levels were increased in septic patients on days 1 and 5 of sepsis compared with healthy controls (Table 3). Differences among the rs2071746 genotypes and HO-1 plasma concentration could not be shown.
Linkage between rs2071746 and the (GT)n microsatellite of the human HMOX1 gene
Linkage of the T allele of rs2071746 with short (GT)n repeats (and vice versa) has been reported by Buis et al. (11). Using HAPMAX (22), we performed linkage analysis by calculating the maximum likelihood estimates of haplotype frequencies. For the VISEP group, the most common haplotype (47.4%) is composed of the rs2071746 A and the (GT)30 alleles, followed by 23- and 24-GT-unit-long repeats linked to the T allele (20.2% and 9.3%, respectively). Importantly, very long repeats are linked to the T allele, with T/(GT)37 being the most common haplotype (2.4%). This pattern could also be observed within the healthy controls, showing similar haplotype frequencies compared with the VISEP patients (P = 0.224). Detailed haplotype frequencies are displayed in Figure 1. Although our results on the allele distributions of rs2071746 and the microsatellite, respectively, and their linkage within haplotypes are very similar to the findings of Buis et al. (11), it has to be mentioned that, compared with our data, their absolute GT-repeat length estimates differ by one unit.
Association of the SNPs of BLVRA and BLVRB and outcome
We inspected exons and their flanking sequence in the two genes for which SNPs with reported allele frequencies were available and resequenced exons 2 and 4 of BLVRA and the last exon 5 of BLVRB for which sequence variation is known from dbSNP (http://www.ncbi.nlm.nih.gov/snp). Of the investigated SNPs of BLVRA and BLVRB, only rs699511, rs699512, rs10234057, rs1802846, and rs45460395 had minor allele frequencies greater than 5%, sufficient for further association analysis. All other SNPs were excluded because of genetic homogeneity of the population (monomorph or less than three minor alleles in the entire study group). All SNPs of BLVRA and BLVRB were screened for association with sepsis and related organ dysfunction, but failed to reach statistical significance (Table 4). Thus, HMOX1 combined BLVRA/B haplotypes were not further analyzed.
In the present study, we confirm and extend an association of HMOX1 genotype with outcome in severe human sepsis, whereas BLVR genotypes had no impact on outcome. Although the German and Greek cohorts differed regarding age, mortality, and source of sepsis, they both displayed an overrepresentation of carriers of the A allele that may be associated with unfavorable outcome among patients hospitalized for life-threatening infections. In addition, within the German cohort, an association of the AA genotype with 28-day mortality was observed. Interestingly, this was not seen in the HSSG patients, consistent with the notion that differences regarding the differing focus of infection or regarding patient factors, e.g., critical care management, may influence clinical outcome of patients sharing identical HMOX1 genotypes.
Heme oxygenase isoenzymes catalyze the oxidative cleavage of the α-mesocarbon of heme, yielding equimolar amounts of biliverdin-IXa, divalent iron, and CO. Bilirubin, a second-order metabolite, is a potent anti-inflammatory and antioxidant substance, but bilirubin may also induce inflammation and apoptosis (23). Consistently, a higher serum bilirubin level on ICU admission is associated with ARDS and mortality in sepsis (24). Most cellular biliverdin is generated from heme, and accordingly, HMOX1 was shown to play a role in various inflammatory states, including sepsis (25, 26). Furthermore, increased arterial CO and monocytic HO-1 expression (27) and upregulation of HMOX1 during early sepsis paralleled by downregulation of nitric oxide synthase (26) have been reported.
Most information regarding the role of HMOX1 and its byproducts in sepsis is derived from rodent models. Several groups described a positive impact of HMOX1 on sepsis outcome (28, 29), and Chung et al. (28) provided evidence that CO released by HMOX1 from heme plays a distinct role in the antimicrobial defense.
However, these results cannot easily be transferred into human clinical settings, as there are interspecies differences regarding the HMOX1 gene. Especially the upstream region of human HMOX1, where the (GT)n microsatellite and rs2071746 SNP are localized, differs from the corresponding rodent gene, and not even the closest phylogenetic relatives of man (chimpanzee and gorilla) share this genetic structure, which may affect gene expression (30).
Another genetic variation is in linkage disequilibrium with the (GT)n microsatellite (155 bp upstream). This SNP rs2071746 (A-419T) determines a genotype (TT), carriers of which have a higher risk of developing Alzheimer dementia (31) and an increased incidence of ischemic heart disease (32). It is noteworthy that, although strong linkage between the length of the (GT)n microsatellite and alleles of the rs2071746 SNP is observed, the two genetic variations are reported to associate independently of each other with clinical phenotypes (11, 33). This may in part be explained by different microsatellite length classifications. Our observation that long (l) alleles are in linkage with the same rs2071746 allele as short (s) (GT)n repeats is an argument to use a three-class repeat classification (see also Fig. 1).
Recent evidence regarding causes of death within the first weeks or months after sepsis indicates that mortality seems to depend on both altered immune function (34) and cardiac mortality (35), which are characteristically affected by the HMOX1 genotype (10).
The two polymorphisms within HMOX1 investigated here were already in the focus of previous studies including organ transplantation (11, 36) and coronary artery disease (32) and were shown to affect gene expression (7). One explanation for this could be the linkage of SNP rs2071746 and the (GT)n length polymorphism. Middle-long (GT)n repeats (27–33 repeat units) were found to be in linkage with A alleles of the rs2071746 in our cohort as well as in other investigations (32, 37). These (GT)n repeats could therefore lead to HMOX1 alternative transcript variants and, because of linkage, affect the A allele with negative impact on outcome. Recently, Sheu and coworkers (9) could demonstrate an association of the HMOX1 promoter polymorphism and ARDS.
The effects of HMOX1 genotypes on HO-1 plasma concentrations were already subject to investigation (7–9), resulting in somehow conflicting results: although Sheu and coworkers (9) used a trimodal (GT)n repeat classification, plasma HO-1 levels in homozygous T allele carriers of rs2071746 were lower compared with AA and AT genotypes. After i.v. administration of heme arginate to healthy volunteers, Doberer and coworkers (7) could not find differences in HO-1 protein expression regarding genotypes in a bimodal (GT)n repeat classification in peripheral blood mononuclear cells, but mRNA steady-state levels were associated with the genotype. Using the same trimodal (GT)n repeat classification, Saukkonen and coworkers (8) describe similar genotype distribution and haplotype frequencies in critical ill patients as were found in our patient population. Plasma HO-1 levels significantly increased in their cohort during ICU stay compared with healthy controls, whereas carriers of homozygous T alleles of rs2071746 had lower HO-1 plasma levels compared with the AA genotype. Because of the linkage disequilibrium between the (GT)n microsatellite and rs2071746, a trimodal (GT)n repeat classification based on rs2071746 genotypes should therefore be preferred. Different origins of sepsis, nonstandardized timepoints of blood sample collection in a human setting, and the small sample size may explain failure to detect association with the corresponding genotype in our cohort. But it has also to be mentioned that elevated HO-1 plasma levels in septic patients may reflect not only the cellular expression of HMOX1, but also cellular damage releasing the cytosolic enzyme into the plasma. This could distort a possible association of HMOX1 expression levels related to HMOX1 genotypes. Measuring direct HMOX1 and HO-1 expression from different tissues (i.e., liver or lung tissue) might therefore be of major impact, but was not available in our cohort. In this respect, the gastrointestinal tract plays a key role in initiating and perpetuating septic complications. Whereas VISEP patients were predominantly surgical patients with sepsis caused by (postoperative) gastrointestinal tract infection, HSSG patients mostly had respiratory tract infection. It is obvious that in abdominal sepsis aspects, such as anastomotic breakdown, are critical for the development of sepsis, which are independent of the genotype.
Heme and its degradation products, including biliverdin, are potentially cytotoxic. We therefore assessed possible associations of common SNPs in BVR A and BVR B genes with outcome. Genetic association studies of the BLVR system are so far scarce. Lin and coworkers (14) investigated common variants of four bilirubin metabolism genes and their association with serum bilirubin and coronary artery disease in the Chinese Han population. They provide genetic evidence for an association with bilirubin metabolism, bilirubin levels, and coronary artery disease. Association studies of BLVR and sepsis-related organ dysfunction are so far missing. We resequenced portions of the genes covering rs699511, rs699512, BLVRA), and rs45460395 (last exon of BLVRB). Allele frequencies were in the range of published population data, but without relation to outcome.
Taken together, we found an association between the (GT)n gene polymorphism and rs2071746 of HMOX1 and 28-day mortality rate of patients with severe sepsis. In contrast, association of the BVR system with outcome of sepsis could not be observed. Consistent with the complex role of heme and its metabolites in cellular homeostasis, there seems to be no simple association between genotype and outcome across patient populations. As therapeutic interventions such as CO inhalation are currently discussed, our data might shed light on potential target populations.
The authors thank their study nurse team for patient recruitment and blood sample collection, Edith Walther and Beate Szafranski, for technical support in performing genotyping and ELISA.
1. Bauer M, Huse K, Settmacher U, Claus RA: The heme oxygenase–carbon monoxide system: regulation and role in stress response and organ failure. Intensive Care Med
34: 640–648, 2008.
2. Scott JR, Chin BY, Bilban MH, Otterbein LE: Restoring HOmeostasis: is heme oxygenase-1 ready for the clinic? Trends Pharmacol Sci
28: 200–205, 2007.
3. Bauer I, Rensing H, Florax A, Ulrich C, Pistorius G, Redl H, Bauer M: Expression pattern and regulation of heme oxygenase-1/heat shock protein 32 in human liver cells. Shock
20: 116–122, 2003.
4. Kutty RK, Kutty G, Rodriguez IR, Chader GJ, Wiggert B: Chromosomal localization of the human heme oxygenase genes: heme oxygenase-1 (HMOX1) maps to chromosome 22q12 and heme oxygenase-2 (HMOX2) maps to chromosome 16p13.3. Genomics
20: 513–516, 1994.
5. Shibahara S, Sato M, Muller RM, Yoshida T: Structural organization of the human heme oxygenase gene and the function of its promoter. Eur J Biochem
179: 557–563, 1989.
6. Kimpara T, Takeda A, Watanabe K, Itoyama Y, Ikawa S, Watanabe M, Arai H, Sasaki H, Higuchi S, Okita N: Microsatellite polymorphism in the human heme oxygenase-1 gene promoter and its application in association studies with Alzheimer and Parkinson disease. Hum Genet
100: 145–147, 1997.
7. Doberer D, Haschemi A, Andreas M, Zapf TC, Clive B, Jeitler M, Heinzl H, Wagner O, Wolzy M, Bilban M: Haem arginate infusion stimulates haem oxygenase-1 expression in healthy subjects. Br J Pharmacol
161: 1751–1762, 2010.
8. Saukkonen K, Lakkisto P, Kaunisto MA, Varpula M, Voipio-Pulkki LM, Varpula T, Pettila V, Pulkki K: Heme oxygenase 1 polymorphisms and plasma concentrations in critically ill patients. Shock
34: 558–564, 2010.
9. Sheu CC, Zhai R, Wang Z, Gong MN, Tejera P, Chen F, Su L, Thompson BT, Christiani DC: Heme oxygenase-1 microsatellite polymorphism and haplotypes are associated with the development of acute respiratory distress syndrome. Intensive Care Med
35: 1343–1351, 2009.
10. Exner M, Minar E, Wagner O, Schillinger M: The role of heme oxygenase-1 promoter polymorphisms in human disease. Free Radic Biol Med
37: 1097–1104, 2004.
11. Buis CI, van der Steege G, Visser DS, Nolte IM, Hepkema BG, Nijsten M, Hepkema BG, Slooff MJ, Porte RJ: Heme oxygenase-1 genotype of the donor is associated with graft survival after liver transplantation. Am J Transplant
8: 377–385, 2008.
12. Parkar M, Jeremiah SJ, Povey S, Lee AF, Finlay FO, Goodfellow PN, Solomon E: Confirmation of the assignment of human biliverdin reductase to chromosome 7. Ann Hum Genet
48: 57–60, 1984.
13. Saito F, Yamaguchi T, Komuro A, Tobe T, Ikeuchi T, Tomita M, Tomita M, Nakajima H: Mapping of the newly identified biliverdin-IX beta reductase gene (BLVRB) to human chromosome 19q13.13–>q13.2 by fluorescence in situ
hybridization. Cytogenet Cell Genet
71: 179–181, 1995.
14. Lin R, Wang Y, Wang Y, Fu W, Zhang D, Zheng H, Yu T, Wang Y, Shen M, Lei R, et al.: Common variants of four bilirubin metabolism genes and their association with serum bilirubin and coronary artery disease in Chinese Han population. Pharmacogenet Genomics
19: 310–318, 2009.
15. Lin R, Wang X, Wang Y, Zhang F, Wang Y, Fu W, Yu T, Li S, Xiong M, Huang W, et al.: Association of polymorphisms in four bilirubin metabolism genes with serum bilirubin in three Asian populations. Hum Mutat
30: 609–615, 2009.
16. Larsen R, Gozzelino R, Jeney V, Tokaji L, Bozza FA, Japiassu AM, Bonaparte D, Cavalcante MM, Chora A, Ferreira A, et al.: A central role for free heme in the pathogenesis of severe sepsis. Sci Transl Med
2: 51ra71, 2010.
17. Clark JF, Sharp FR: Bilirubin oxidation products (BOXes) and their role in cerebral vasospasm after subarachnoid hemorrhage. J Cereb Blood Flow Metab
26: 1223–1233, 2006.
18. Brunkhorst FM, Engel C, Bloos F, Meier-Hellmann A, Ragaller M, Weiler N, Moerer O, Gruendling M, Oppert M, Grond S, et al.: Intensive insulin therapy and pentastarch resuscitation in severe sepsis. N Engl J Med
358: 125–139, 2008.
19. Vincent JL, Moreno R, Takala J, Willatts S, De Mendonca A, Bruining H, Reinhart CK, Suter PM, Thijs LG: The SOFA (Sepsis-related Organ Failure Assessment) score to describe organ dysfunction/failure. On behalf of the Working Group on Sepsis-Related Problems of the European Society of Intensive Care Medicine. Intensive Care Med
22: 707–710, 1996.
20. Bone RC, Balk RA, Cerra FB, Dellinger RP, Fein AM, Knaus WA, Schein RM, Sibbald WJ: Definitions for sepsis and organ failure and guidelines for the use of innovative therapies in sepsis. The ACCP/SCCM Consensus Conference Committee. American College of Chest Physicians/Society of Critical Care Medicine. 1992. Chest
136: e28, 2009.
21. Rodriguez S, Gaunt TR, Day IN: Hardy-Weinberg equilibrium testing of biological ascertainment for Mendelian randomization studies. Am J Epidemiol
169: 505–514, 2009.
22. Krawczak M, Konecki DS, Schmidtke J, Duck M, Engel W, Nutzenadel W, Trefz FK: Allelic association of the cystic fibrosis locus and two DNA markers, XV2c and KM19, in 55 German families. Hum Genet
80: 78–80, 1988.
23. Abraham NG, Kappas A: Pharmacological and clinical aspects of heme oxygenase. Pharmacol Rev
60: 79–127, 2008.
24. Zhai R, Sheu CC, Su L, Gong MN, Tejera P, Chen F, Gong MN, Wang Z, Convery MP, Thompson BT, Christiani DC: Serum bilirubin levels on ICU admission are associated with ARDS development and mortality in sepsis. Thorax
64: 784–790, 2009.
25. Schaer DJ, Schaer CA, Schoedon G, Imhof A, Kurrer MO: Hemophagocytic macrophages constitute a major compartment of heme oxygenase expression in sepsis. Eur J Haematol
77: 432–436, 2006.
26. Reade MC, Millo JL, Young JD, Boyd CA: Nitric oxide synthase is downregulated, while haem oxygenase is increased, in patients with septic shock. Br J Anaesth
94: 468–473, 2005.
27. Takaki S, Takeyama N, Kajita Y, Yabuki T, Noguchi H, Miki Y, Inoue Y, Nakagawa T, Noguchi H: Beneficial effects of the heme oxygenase-1/carbon monoxide system in patients with severe sepsis/septic shock. Intensive Care Med
36: 42–48, 2010.
28. Chung SW, Liu X, Macias AA, Baron RM, Perrella MA: Heme oxygenase-1–derived carbon monoxide enhances the host defense response to microbial sepsis in mice. J Clin Invest
118: 239–247, 2008.
29. Rensing H, Bauer I, Peters I, Wein T, Silomon M, Jaeschke H, Bauer M: Role of reactive oxygen species for hepatocellular injury and heme oxygenase-1 gene expression after hemorrhage and resuscitation. Shock
12: 300–308, 1999.
30. Bennett P: Demystified… microsatellites. Mol Pathol
53: 177–183, 2000.
31. Infante J, Rodriguez-Rodriguez E, Mateo I, Llorca J, Vazquez-Higuera JL, Berciano J, Combarros O: Gene-gene interaction between heme oxygenase-1 and liver X receptor-beta and Alzheimer’s disease risk. Neurobiol Aging
31: 710–714, 2008.
32. Ono K, Goto Y, Takagi S, Baba S, Tago N, Nonogi H, Iwai N: A promoter variant of the heme oxygenase-1 gene may reduce the incidence of ischemic heart disease in Japanese. Atherosclerosis
173: 315–319, 2004.
33. Rueda B, Oliver J, Robledo G, Lopez-Nevot MA, Balsa A, Pascual-Salcedo D, Gonzalez-Gay MA, Gonzalez-Escribano MF, Martin J: HO-1 promoter polymorphism associated with rheumatoid arthritis. Arthritis Rheum
56: 3953–3958, 2007.
34. Kaplan V, Clermont G, Griffin MF, Kasal J, Watson RS, Linde-Zwirble WT, Angus DC: Pneumonia: still the old man’s friend? Arch Intern Med
163: 317–323, 2003.
35. Musher DM, Rueda AM, Kaka AS, Mapara SM: The association between pneumococcal pneumonia and acute cardiac events. Clin Infect Dis
45: 158–165, 2007.
36. Geuken E, Buis CI, Visser DS, Blokzijl H, Moshage H, Nemes B, Leuvenink HG, de Jong KP, Peeters PM, Slooff MJ, et al.: Expression of heme oxygenase-1 in human livers before transplantation correlates with graft injury and function after transplantation. Am J Transplant
5: 1875–1885, 2005.
37. Ono K, Mannami T, Iwai N: Association of a promoter variant of the haeme oxygenase-1 gene with hypertension in women. J Hypertens
21: 1497–1503, 2003.
Heme oxygenase; biliverdin reductase; mortality; organ failure; genotypes; infection