A defining feature of the COVID-19 pandemic has been the wide spectrum of clinical presentations, ranging from asymptomatic viremia to life-threatening illness.1 Factors such as advancing age and co-morbid status have been shown to strongly correlate with risk of developing severe COVID-19, but that is unlikely to be the whole story.2 Although the death rate in older patients is significantly higher than in their younger counterparts, a large proportion of elderly patients recover without having developed severe disease.
Beyond age and co-morbid status, what additional factors might protect an individual from severe disease once exposed? One suggestion that emerged early in the pandemic was that the patient’s own genetics might play a role in the way their body responds to infection.3 There has been a staggering international effort to investigate this and last week, Nature published findings from the Genomics of Mortality in Critical Care (GenOMICC) study.4 As always, here is a link to the full paper, but here we will take a look at what the work shows and the potential implications for COVID-19 and beyond.
From the Beginning
Studies which investigate the genetics of acutely unwell patients are few and far between but could be possible based on two basic rules, explained here:
1. We know what the human genome looks like and how it varies between individuals
The human genome is organised as a long string of nucleotides, around 3 billion in total. We have 4 types of nucleotide, guanine (G), adenine (A), cytosine (C) and thymine (T). Variation within the genome is normal, very common, and typically harmless. We all have approximately 4 million differences compared to the reference genome. These differences typically take the form of Single Nucleotide Polymorphisms (SNPs), where there is a substitution from one nucleotide to another (i.e. G>A).
2. We can pinpoint genomic variation between individuals that increases susceptibility to disease
Genomic variation exists on a spectrum (Figure 1) of frequency (how common it is in the population) and its effect size (how likely that genetic difference is to cause disease). Traditionally in clinical genetics, we focus on rare genetic variation which has a high effect size. For example, rare mutations in the FBN1 gene can cause Marfan syndrome. Clearly, most genetic variation is not associated with rare syndromic disease, with the vast majority of SNPs representing relatively benign variation within the genome. However, it is now increasingly recognised that at least some of this more common variation is relevant to a patient’s health. In recent years it has been shown that some of this variation can be used to predict an individual’s probability of developing common disease, the chance they will respond to their medication, and the risk of developing an adverse drug reaction.5,6
What is the GenOMICC Study?
The GenOMICC study has been actively recruiting since 2016 and aims to understand the genetic factors which influence outcome in critical illness, such as sepsis. As the SARS-CoV-2 pandemic emerged, the group – led by Kenneth Baillie in Edinburgh – pivoted to recruit critically ill patients with COVID-19. This manuscript presents an early analysis of part of that cohort, exploring whether there are any genetic changes associated with life-threatening COVID-19.
What did they do?
In this manuscript they start with 2,734 unique individuals with confirmed COVID-19. 2,636 patients were deemed to require continuous cardio-respiratory monitoring and recruited directly via the GenOMICC study from ICUs across the UK. An additional 135 patients admitted to hospital with confirmed COVID-19 were recruited through the ISARIC4C study.
DNA samples were taken from all participants and were tested using a methodology which is designed to detect 730,059 SNPs across the genome. After quality control (QC), genetic data were available for 2,244 patients for analysis. This type of QC filtering is quite normal in genetic studies.
This data then fed into a Genome Wide Analysis Study (GWAS). This is an established approach which examines whether individual’s with a given trait, in this case life threatening COVID-19, share genetic signals when compared to the wider population. Because the human genome is so large and so many SNPs are examined, the analysis approach is extremely stringent, with a validation process in replication datasets, to minimise the possibility of spurious associations.
In this study, genetic data from the life threatening COVID-19 cohort was compared against ancestry matched controls from the UK-Biobank (UKB). The majority of patients recruited to the study, and the vast majority of patients in the UKB, are of white European ancestry. Therefore, it was not hugely surprising that data for other ancestral cohorts were less robust. For the primary analysis, the data from 1,676 individuals of European decent were included. This is understandable from a methodological perspective but should be remembered when interpreting the results.
What did the study show?
The manuscript reports five robust and replicable genomic changes in the human genome which are associated with life threatening COVID-19 (Figure 2). This means that patients with life threatening COVID-19 were significantly more likely to have these specific SNPs. From this, the authors can infer that the genes in or near that region may play a role in predisposing to severe disease.
Figure 2 in tweet
Significant signals were spread across the genome. The signal on chromosome 3 has previously been reported but the fact it has been replicated in this study is encouraging.7There were four other significant signals associated life threatening COVID-19 which are entirely novel, one on chromosome 12, two on chromosome 19 and one on chromosome 21. The authors then highlighted the nearest candidate genes in each region. These were OAS1 (Chr 12), DPP9 (Chr 19), TYK2 (Chr 19), and IFNAR2 (Chr 21).
Great, but why does this matter?
This GenOMICC study provides an insight into potential disease mechanisms of life threatening COVID-19. The genes identified in this manuscript can be broadly split into those related to innate antiviral defences, known to be important early in disease (IFNAR2 and OAS1), and those related to host-driven inflammatory lung injury (DPP9 and TYK2), which is a key later stage mechanism of life-threatening Covid-19.
A relatively immediate impact of this work is to support the identification of drug candidates. Take TYK2 for example, which encodes for the Tyrosine Kinase-2 protein. The gene change near TYK2 identified in the study is hypothesised to cause increased expression of the TYK2 protein, and this signal was associated with an odds ratio of 1.3 for developing life threatening COVID-19. JAK-inhibitors, such as baricitinib, which are currently being tested in COVID-19, are hypothesised to inhibit TYK2 activity and therefore represent exciting therapeutic candidates. By remarkable coincidence, on the same day the GenoMICC study was printed, the NEJM published a study showing a moderate improvement in time to recovery for hospitalized adults with Covid-19 who received baricitinib plus remdesivir, over remdesivir alone.8
Increased expression of IFNAR2 meanwhile appears to be protective for life threatening COVID-19. This gene codes for a subunit of the interferon receptor subunit, and rare genetic changes in IFNAR2 are known to cause an immunodeficiency syndrome.9,10 The results from the GenoMICC study suggest a protective role for interferons, although the authors note that exogenous interferon treatment did not reduce mortality in hospitalised patients in a large-scale clinical trial.11 However, the clear genetic signal detected in this work suggests that interferons do play a role in the pathogenesis of COVID-19 and further trials may be warranted. It may be that interferons are more useful earlier in the disease course when viral loads are higher.
What’s the bottom line?
This GenOMICC study has identified a number of new genetic associations with life threatening COVID-19. This has the potential to help identify credible drug candidates and highlight those patients at particular risk. However, it’s critical to remember that genetics isn’t a panacea. Just because an individual has one of these genetic signals does not mean they will definitely go onto develop severe disease. Nor does it mean an individual without one of these changes is entirely protected. These genetic differences will nudge the odds ratio for life threatening COVID-19 slightly one way or another, but must be reviewed in the context of the patient as a whole.
What Next?
To be clear, although technology now exists to detect genetic variation rapidly in the acute setting, there is no suggestion that we’re going to start testing patients for these SNPs as they present to ED.12,13 If you knew an individual carried one of these variants, would it change your clinical practice or your decision regarding escalation? I expect not, although this is a fascinating and novel concept for debate.
The idea of using genetic risk scores to inform clinical practice in the ED might be viewed, to put it mildly, as a somewhat mad idea. However, it may not be as far away as you might think. Probably not in relation to COVID-19, but perhaps around the detection of Acute Coronary Syndrome or for identifying poor-responders to medication. Take a listen to Rick Body’s lecture on “The Future of Diagnostics” from St Emlyn’s Live in 2019 14, where he talks about just that idea.
In conclusion, the GenOMICC team and their collaborators should be exceptionally proud of this work. In a matter of months, they’ve identified robust genomic associations with life-threatening COVID-19. This work will help to prioritise drug targets and, with over 6,500 patients now recruited, many more insights are likely to come from this study.
vb
John McDermott @John_H_McD.
References
1. Meyerowitz, E. A., Richterman, A., Bogoch, I. I., Low, N. & Cevik, M. Towards an accurate and systematic characterisation of persistently asymptomatic infection with SARS-CoV-2. The Lancet Infectious Diseases 0, (2020).
2. Ruan, S. Likelihood of survival of coronavirus disease 2019. Lancet Infect Dis 20, 630–631 (2020).
3. The COVID-19 Host Genetics Initiative. The COVID-19 Host Genetics Initiative, a global initiative to elucidate the role of host genetic factors in susceptibility and severity of the SARS-CoV-2 virus pandemic. European Journal of Human Genetics 28, 715–718 (2020).
4. Pairo-Castineira, E. et al. Genetic mechanisms of critical illness in Covid-19. Nature (2020) doi:10.1038/s41586-020-03065-y.
5. Roden, D. M. Clopidogrel Pharmacogenetics – Why the Wait? N Engl J Med 381, 1677–1678 (2019).
6. Khera, A. V. et al. Genome-wide polygenic scores for common diseases identify individuals with risk equivalent to monogenic mutations. Nat Genet 50, 1219–1224 (2018).
7. Severe Covid-19 GWAS Group et al. Genomewide Association Study of Severe Covid-19 with Respiratory Failure. N Engl J Med 383, 1522–1534 (2020).
8. Kalil, A. C. et al. Baricitinib plus Remdesivir for Hospitalized Adults with Covid-19. N Engl J Med (2020) doi:10.1056/NEJMoa2031994.
9. Zhang, Q. et al. Inborn errors of type I IFN immunity in patients with life-threatening COVID-19. Science 370, (2020).
10. Duncan, C. J. A. et al. Human IFNAR2 deficiency: Lessons for antiviral immunity. Sci Transl Med 7, 307ra154 (2015).
11. WHO Solidarity Trial Consortium et al. Repurposed Antiviral Drugs for Covid-19 – Interim WHO Solidarity Trial Results. N Engl J Med (2020) doi:10.1056/NEJMoa2023184.
12. Llibre, A. et al. Development and clinical validation of the Genedrive point-of-care test for qualitative detection of hepatitis C virus. Gut 67, 2017–2024 (2018).
13. Duffy, D. et al. An in vitro diagnostic certified point of care single nucleotide test for IL28B polymorphisms. PLoS ONE 12, e0183084 (2017).
14. Rick Body, “The Future of Diagnostics #stemlynsLIVE,” in St.Emlyn’s, August 31, 2019, https://www.stemlynsblog.org/the-future-of-diagnostics-stemlynslive/
15 Genomicc recruitment tracker https://genomicc.org/uk/recruitment/