HMOX1 genetic polymorphisms and outcomes in infectious disease: a systematic review

Abstract

Introduction: Heme-oxygenase 1 (HMOX1) is a critical stress response gene that catalyzes the multistep oxidation of heme. A GT(n) repeat of variable length in the promoter in has been associated with a wide range of human diseases, including infections. This paper aims to summarise and systematically review associations between the length of the HMOX1 GT(n) promoter and infectious disease in humans. Methods: A search using relevant terms was performed in PubMED and EMBASE through to 15/01/21 identifying all research that studied an association between the HMOX1 GT(n) repeat polymorphism and the incidence and/or outcome of any human infectious disease. Citations were screened for additional studies. Potential studies were screened for inclusion by two authors. Data was extracted on allele frequency, genotype, strength of association, mechanism of genotyping, and potential biases. A narrative review was performed across each type of infection. Results 1,533 studies were identified in the search, and one via citation screening. Sixteen studies were ultimately included, seven in malaria, three in HIV, three in sepsis, and one each in pneumonia, hepatitis C, and acute respiratory distress syndrome (ARDS). Sample sizes for nearly all studies were small (biggest study, n = 1,646). Allelic definition was different across all included studies. In malaria, three studies suggested that longer alleles were associated with reduced risk of severe malaria, particularly malaria-induced renal dysfunction, with four studies identifying a null association. In sepsis, two studies suggested an association with longer alleles and better outcomes. Conclusions Despite the importance of HMOX1 in survival from infection, and the association between repeat length and gene expression, the clinical data supporting an association between repeat length and incidence and/or outcome of infection remain inconclusive. The most promising data supports a potential association with protection from severe malaria, although this was not found in all studies.

Phenotypic associations with the HMOX1 GT(n) promoter repeat in European populations

“…As the impact of HMOX1 repeat polymorphisms might only be present in specific environmental associations such as in severe cellular stress, and there is some evidence that HMOX1 polymorphisms alter survival from severe infection, 14,30 we undertook a survival analysis for 11 coded infections within UK Biobank (Table 4).…”

“…Given this variation; and with no empiric evidence to support previous definitions, we chose to simply quartile the alleles and include them in the model. 30…”

“…As the impact of HMOX1 repeat polymorphisms might only be present in specific environmental associations such as in severe cellular stress, and there is some evidence that HMOX1 polymorphisms alter survival from severe infection, 14,30 we undertook a survival analysis for 11 coded infections within UK Biobank (Table 4). Infections with a low 28-day mortality were excluded from the analysis and only three infections were taken forward for formal modelling (endocarditis, pneumonia, and sepsis).…”

“…Although this technical limitation should be recognised, it is important to note that all prior associations with this repeat have been with repeat length, with no data that we are aware of suggesting the effect is related to a particular allele rather than the total length of the repeat. 1,[7][8][9][10]14,16,30,37 . Secondly, a genome-wide analysis of 2,060 expression short tandem repeats found linear associations to be the most common association between STR's and gene expression, a finding also identified in other studies of human STRS.…”

“…In a previous review of HMOX1 promoter polymorphisms and infection, we showed that this classification was inconsistent, and creates a significant risk of bias. 30 For example, in the previous meta-analysis on type 2 diabetes, one study was excluded as allelic definition was too different from the other studies, while they called short alleles as under 25-27 repeats, and long alleles as 25 to 27 repeats. 7 However, given that i) the trimodal distribution peaks at 27 repeats, ii) definitions of repeat length vary and iii) there are generally small errors and variability in genotyping by fragment length polymorphism, these definitions are highly questionable and are at significant risk of biasing studies.…”