key: cord-0884342-gcyj5dvq authors: Zhao, Jie V.; Schooling, C. Mary; Leung, Gabriel M. title: Using genetics to understand the role of antihypertensive drugs modulating angiotensin‐converting enzyme in immune function and inflammation date: 2020-10-23 journal: Br J Clin Pharmacol DOI: 10.1111/bcp.14572 sha: de64dd1d6d23dabfd4781fbc6aabc7cd93df76cd doc_id: 884342 cord_uid: gcyj5dvq AIM: Angiotensin‐converting enzyme 2 (ACE 2) is the binding domain for severe acute respiratory syndrome coronavirus (SARS‐CoV) and SARSCoV‐2. Some antihypertensive drugs affect ACE2 expression or activity (ACE inhibitors and angiotensin II receptor blockers [ARBs]), suggesting use of other hypertensives might be preferable, such as calcium channel blockers (CCBs). Given the limited evidence, the International Society of Hypertension does not support such a policy. METHODS: We used a Mendelian randomization study to obtain unconfounded associations of antihypertensives, instrumented by published genetic variants in genes regulating target proteins of these drugs, with immune (lymphocyte and neutrophil percentage) and inflammatory (tumour necrosis factor alpha [TNF‐α]) markers in the largest available genome‐wide association studies. RESULTS: Genetically predicted effects of ACE inhibitors increased lymphocyte percentage (0.78, 95% confidence interval [CI] 0.35, 1.22), decreased neutrophil percentage (−0.64, 95% CI −1.09, −0.20) and possibly lowered TNF‐α (−4.92, 95% CI −8.50, −1.33). CCBs showed a similar pattern for immune function (lymphocyte percentage 0.21, 95% CI 0.05 to 0.36; neutrophil percentage −0.23, 95% CI −0.39 to −0.08) but had no effect on TNF‐α, as did potassium‐sparing diuretics and aldosterone antagonists, and vasodilator antihypertensives. ARBs and other classes of hypertensives had no effect on immune function or TNF‐α. CONCLUSION: Varying effects of different classes of antihypertensives on immune and inflammatory markers do not suggest antihypertensive use based on their role in ACE2 expression, but instead suggest investigation of the role of antihypertensives in immune function and inflammation might reveal important information that could optimize their use in SARSCoV‐2. Angiotensin-converting enzyme 2 (ACE2) is the binding domain of the severe acute respiratory syndrome coronaviruses (SARS-CoV) and SARSCoV-2. 1 A key concern is that ACE inhibitors and angiotensin II receptor blockers (ARBs), commonly used antihypertensives, may increase ACE2 expression or activity, and thereby increase the risk of COVID-19 infection. 2 Correspondingly, calcium channel blockers (CCBs), which do not affect ACE2 expression or activity, have been proposed as an alternative treatment. 2 In contrast, it has also been suggested that upregulation of ACE2 expression might protect against infection if binding of the coronavirus spike protein to ACE2 leads to ACE2 downregulation, 3 but the mechanism has not been assessed. Given the unclear role of ACE inhibitors and ARBs in infection, the International Society of Hypertension has stated "there is no good evidence to change the use of ACE-inhibitors or ARBs for the management of raised blood pressure in the context of avoiding or treating COVID-19 infection". 4 Consistently, limited evidence from a small observational study suggests patients using ACE inhibitors or ARBs had higher CD3 and CD8 T cell counts. 5 Recent observational studies also show no association of use of ACE inhibitors and ARBs with risk of in-hospital death in patients with COVID-19. 6-8 A potential benefit was seen with ACE inhibitor use, 6 but this "may be due to residual confounding" 9 and needs to be confirmed in clinical trials, as well as contextualized by mechanistic insight. In these circumstances when experimental evidence is lacking from drug testing, Mendelian randomization (MR) provides an alternative approach by exploiting genetic variants, randomly allocated at conception, that mimic drug effects. This study design has been successfully applied to assess the efficacy of several medications. 10, 11 Published genetic variants corresponding to the effects of a range of antihypertensives exist. [12] [13] [14] Here, to be comprehensive we used these genetic variants to assess the effects of a comprehensive range of antihypertensives on key markers of immune function and inflammation related to COVID-19, ie, lymphocyte percentage, neutrophil percentage and tumour necrosis factor alpha (TNF-α). Severe COVID-19 is associated with a major immune inflammatory response with abundant lymphocytes, neutrophils and excess inflammation. 15 Lymphocyte percentage is an established predictor of the severity of COVID-19, 16 neutrophils are a modulator of immune response. 17 TNF-α, an amplifier of inflammation, is important in acute inflammatory reactions; anti-TNF therapy has recently been proposed as a promising COVID-19 treatment strategy. 15 2 | METHODS We used an MR study to obtain unconfounded associations of the effects of antihypertensive drug use largely from published sources with lymphocyte percentage, neutrophil percentage and TNF-α. Specifically, we used as instruments published genetic variants predicting the effects of the use of different classes of antihypertensive drugs from genes regulating the drug-target proteins 12,13 which were related to systolic blood pressure (SBP) in the UK Biobank. For ACE inhibitors, we also replicated our findings using genetic variants related both to ACE concentration and to SBP in the UK Biobank as instrument. As we used several different sets of instruments, for ease of comparison of the MR estimates, we used their genetic associations with SBP from the UK Biobank in 361 194 white British as a proxy for their effects on antihypertensives. We obtained genetic instruments predicting the effects of the use of ACE inhibitors, ARBs and CCBs, as well as other classes of antihypertensives, specifically alpha-adrenoceptor blockers, adrenergic neurone blocking drugs, beta-adrenoceptor blockers, centrally acting antihypertensive drugs, loop diuretics, potassium sparing diuretics (PSDs) and aldosterone antagonists, renin inhibitors, thiazides and related diuretics, and vasodilator antihypertensives from published sources. 12, 13 Specifically, these published studies gave the genetic variants regulating the expression of the relevant drug target genes What is already known about this subject • There are safety concerns about antihypertensives in SARSCoV-2 regulating ACE2 expression or activity. • Observationally, the use of ACE inhibitors is not related to higher risk of COVID-19 events, but might have potential benefits. • These observations have not been confirmed in randomized controlled trials and the relation to immune function and inflammation is unclear. What this study adds Genetic associations with lymphocyte percentage and neutrophil percentage were obtained from UK Biobank summary statistics provided by Neale Lab (http://www.nealelab.is/uk-biobank/). The UK Biobank is a large, ongoing, prospective cohort study with median follow-up time of 11.1 years. 18 MR estimates were based on the SNP-specific Wald estimates (genetic association with outcome divided by genetic association with the exposure), meta-analysed using inverse variance weighting with multiplicative random effects, as necessary. In sensitivity analysis, we used different methods with different assumptions about potential bias from horizontal pleiotropy, including Mendelian randomization pleiotropy residual sum and outlier (MR-PRESSO), a mode-based method, and contamination mixture method. MR-PRESSO is able to identify outliers with potential horizontal pleiotropy amongst multiple genetic variants and provide a corrected estimate after removing these outliers. 20 The modebased method assumes the true causal effect is the value taken by the largest number of genetic variants, 21 so it is robust to outliers, 22 but the estimates are generally conservative. 21 The contamination mixture method is similar but less conservative than the mode-based method. 21, 23 All statistical analyses were conducted using R version 3.6.2 (R Foundation for Statistical Computing, Vienna, Austria), and the R package "MendelianRandomization". All the estimates of genetic associations were taken from publicly available summary statistics, obtained from studies previously collected with informed consent, without any personal information in the study. We used one SNP (rs4968783 in ACE) for ACE inhibitors, one SNP for ARBs (rs118123032 in AGTR1) and 12 SNPs for CCBs scaled to SBP (effect sizes) using UK Biobank summary statistics. 12 We also used one SNP (rs4291 in ACE) for ACE inhibitors and 24 SNPs for CCBs derived based on a GWAS meta-analysis of the UK Biobank and the International Consortium of Blood Pressure. 13 The genetically predicted effects of the use of ACE inhibitors and CCBs both increased lymphocyte percentage and decreased neutrophil percentage, with a larger effect size for ACE inhibitors ( Table 1 ). The estimates for ACE inhibitors were robust to using genetic variants predicting ACE concentration ( (Table 3) . Genetically predicted ACE inhibitors may lower TNF-α (Tables 4 and 5 ), especially when using genetic variants predicting ACE concentration ( Table 5 ). The estimates were robust to using all the genetic variants predicting ACE concentration and different analysis methods (Supporting Information Tables S3 and S4 ). CCBs did not clearly affect TNF-α, and nor did other hypertensives ( Table 6 ). The estimates were robust to different analysis methods (Supporting Information Tables S5 and S6 ). Using genetic proxies for drug effects, we found ACE inhibitors, which increase ACE2 expression, increased lymphocyte percentage, decreased neutrophil percentage and may also lower TNF-α. CCBs, PSDs and aldosterone antagonists, and vasodilator antihypertensives similarly increased lymphocyte percentage and decreased neutrophil percentage, but were unrelated to TNF-α. However, other antihypertensives, including ARBs, which also increase ACE2 expression, had no effect on immune markers or inflammation. As such, consistent with the statement from the International Society of Hypertension 4 and previous observational studies, 6 Lower lymphocyte percentage is predictive of higher severity of COVID-19 infection. 16 Anti-TNF drugs have been hypothesized as a potential treatment for COVID-19 infection. 15 As such, the associa- The target domain of COVID-19, ACE2, has high expression in the testes. 26,27 Sex hormones modulate immune response and inflammation in animals. 28 Testosterone is generally immunosuppressive, while oestrogen tends to be immune-promoting. 29, 30 Genetically predicted testosterone was associated with lower lymphocyte percentage in a recent MR study. 31 Statins, which lower testosterone, 32 have been hypothesized to be protective for COVID-19 33 by modulation of NF-κB, 34 mediated by TNF-α. 35 Further investigation of these mechanistic pathways might help find a unifying explanation for differences in patterns of COVID-19 by sex and setting, similar to differences in other hormone-modulated conditions by setting. 36, 37 Despite consistency across genetic instruments, this study has several limitations. First, MR relies on three assumptions, ie, the genetic instruments are related to the exposure, are not related to potential confounders and the effect of the genetic instrument on the outcome is exclusively through the exposure. 38 To satisfy these assumptions, we used SNPs related to the expression of genes regulating the drug target proteins. We also checked that these SNPs are not directly related to immune function, although we cannot exclude the possibility that unidentified pleiotropic association may exist, which is a common limitation of MR studies. However, we compared the estimates using different SNP selections, which gave consistent findings. Given the possibility of unidentified pleiotropy, we used several different analytic methods that are based on different assumptions. The consistent directions of associations across these methods add confidence to the findings. These methods may differ in precision, for example the estimates from mode-based methods are generally more conservative than the contamination mixture method, 21 so they are used as sensitivity analysis supplementary to the main analysis. Second, measurement error might exist in the single time-point assay of lymphocyte percentage, neutrophil percentage and TNF-α. However, any measurement error should be nondifferential, thus bias towards the null, rather than give positive associations with lymphocyte percentage and inverse associations with neutrophil percentage. Third, the genetic associations with TNF-α were obtained from a relatively small GWAS, which might explain the wide confidence intervals in the association of ACE inhibitors with TNF-α. The associations were also adjusted for body mass index, which can cause a bias in some situations, 39 but is unlikely to do so here. 40 The study was also limited by the few large GWAS of immunity, thus replication in other large GWAS when they are available will be worthwhile. Fourth, T A B L E 2 Associations of ACE inhibitors with lymphocyte and neutrophil percentage using ACE SNPs as instrument in immune function and inflammation, but this may not be due to, or at least not totally due to, ACE2 expression because ARBs which also affect ACE2 expression did not affect lymphocyte percentage, neutrophil percentage or TNF-α. Exploring the underlying pathways, especially the pathways that differ between these antihypertensives, would be worthwhile. The effect of ACE inhibitors and the null effect of ARBs on key markers of immune function and inflammation support the current International Society of Hypertension statement that there is no evidence to indicate the use of antihypertensive drugs based on their role in ACE2 expression. However, concern about the effects of ACE inhibitors on immune function has revealed a complex pattern of effects of different classes of antihypertensive drugs whose elucidation might be relevant to both infectious diseases and optimization of the use of antihypertensives which should be further explored. Abbreviations: CI, confidence interval; PSD, potassium sparing diuretics; SNP, single nucleotide polymorphism; TNF-α, tumour necrosis factor alpha. Receptor recognition by the novel coronavirus from Wuhan: an analysis based on decadelong structural studies of SARS coronavirus Are patients with hypertension and diabetes mellitus at increased risk for COVID-19 infection? Angiotensin receptor blockers as tentative SARS-CoV-2 therapeutics Renin-angiotensin system inhibitors improve the clinical outcomes of COVID-19 patients with hypertension Cardiovascular disease, drug therapy, and mortality in Covid-19 Renin-angiotensin-aldosterone system blockers and the risk of Covid-19 Renin-angiotensinaldosterone system inhibitors and risk of Covid-19 Inhibitors of the renin-angiotensin-aldosterone system and Covid-19 Association of Triglyceride-Lowering LPL variants and LDL-C-lowering LDLR variants with risk of coronary heart disease Mendelian randomization study of ACLY and cardiovascular disease Repurposing antihypertensive drugs for the prevention of Alzheimer's disease: a Mendelian randomization study Use of genetic variants related to antihypertensive drugs to inform on efficacy and side effects ACE and type 2 diabetes risk: a Mendelian randomization study Trials of anti-tumour necrosis factor therapy for COVID-19 are urgently needed Lymphopenia predicts disease severity of COVID-19: a descriptive and predictive study How neutrophils shape adaptive immune responses UK biobank: an open access resource for identifying the causes of a wide range of complex diseases of middle and old age Genome-wide study for circulating metabolites identifies 62 loci and reveals novel systemic effects of LPA Detection of widespread horizontal pleiotropy in causal relationships inferred from Mendelian randomization between complex traits and diseases A comparison of robust Mendelian randomization methods using summary data Robust inference in summary data Mendelian randomization via the zero modal pleiotropy assumption A robust and efficient method for Mendelian randomization with hundreds of genetic variants A SARS-CoV-2 protein interaction map reveals targets for drug repurposing The effect of angiotensin converting enzyme inhibitors and angiotensin receptor blockers on death and severity of disease in patients with coronavirus disease 2019 (COVID-19): A meta-analysis ACE2 Expression in Kidney and Testis May Cause Kidney and Testis Damage After 2019-nCoV Infection Sex differences in immune responses: hormonal effects, antagonistic selection, and evolutionary consequences Effect of hormones on lympatic tissue The response of castrated male rats to the injection of testicular hormone Relevance to human health of growth and reproduction trading-off against longevity The effect of statins on testosterone in men and women, a systematic review and meta-analysis of randomized controlled trials Rationale for Consideration of Statins for COVID-19 Patients Statins may decrease the fatality rate of Middle East respiratory syndrome infection Inhibition of NF-kappaB-mediated inflammation in severe acute respiratory syndrome coronavirus-infected mice increases survival Population differences in the testosterone levels of Young men are associated with prostate cancer disparities in older men Secular trends of breast cancer in China, South Korea, Japan and the United States: application of the age-period-cohort analysis Reading Mendelian randomisation studies: a guide, glossary, and checklist for clinicians Bias in two-sample Mendelian randomization by using covariable-adjusted summary associations Bias from competing risk before recruitment in Mendelian Randomization studies of conditions with shared etiology Comparison of sociodemographic and health-related characteristics of UK biobank participants with those of the general population Mendelian randomization: using genes as instruments for making causal inferences in epidemiology Sick individuals and sick populations Mendelian randomization estimates may be inflated Using genetics to understand the role of antihypertensive drugs modulating angiotensin-converting enzyme in immune function and inflammation The authors would like to thank Neale Lab (http://www.nealelab.is/ uk-biobank/) for conducting the GWAS of biomarkers and providing the genetic associations in the UK Biobank. The authors would also like to thank the authors of the GWAS of inflammatory biomarkers (http://www.computationalmedicine.fi/data#NMR_GWAS) for sharing the data.