Human TMPRSS2 and ACE2 Genetic Variability on COVID-19 Outcomes in Patients From Brazil

1. Introduction

The first coronavirus disease 2019 (COVID-19) case was reported in Wuhan, China [1] almost two years ago. Since then, COVID-19 has afflicted 281,808,270 patients in 210 nations and territories worldwide and resulted in 5,411,759 cumulative deaths until January 03, 2022. In Brazil, the number of COVID-19 cases is among the high-ranking countries in terms of incidence and mortality over these two years, with 22,246,276 positive cases and 618.534 total fatalities through January 03, 2022.

As the literature describes, coronaviruses' spike (S) protein promotes viral entrance into human target cells. The binding of the S protein to a cellular receptor, which allows viral attachment to the surface of target cells, is required for entry. Furthermore, the viral entry involves S protein priming by cellular proteases, which requires S protein cleavage at the S1/S2 and S2' sites enabling the fusing of viral and cellular membranes, a process mediated by the S2 subunit [2,3,4].

In this context, angiotensin-converting enzyme 2 (ACE2) and the transmembrane serine protease 2 (TMPRSS2) are recognized as human entry proteins of the severe acute respiratory syndrome coronavirus 2 (SARS-CoV-2) [3]. Recently, single nucleotide polymorphisms (SNP) on ACE2 and TMPRSS2 genes have been studied in different populations to elucidate the impact on coronavirus disease (COVID-19) establishment. Two SNPs are constantly described in the literature with clinical significance in infectious diseases. The rs35803318 is a synonymous variant (C>T) in the exon 18 of the ACE2 gene, whereas the rs2070788 is an intro variant (G>A) which does not impact the protein [5,6,7,8].

Therefore, in this study, it was investigated the single nucleotide polymorphisms of ACE2 (rs35803318) and TMPRSS2 (rs2070788) genes in COVID-19 patients from Brazil, their association with clinical parameters, and the Expression Quantitative Trait Locus (eQTL) for these SNPs.

2. Materials and Methods

2.1. Study Population and Sample Collection

COVID-19-positive patients from two hospitals: José Frota Institute and São José Hospital, Ceará state (Brazil), were selected. The samples were collected with a nasopharyngeal swab, and SARS-CoV-2 was confirmed by a real-time reverse transcription-polymerase chain reaction (rRT-PCR). RNA was extracted for detection with the 2019-nCoV qPCR Diagnostic Panel from The Centers for Disease Control and Prevention (CDC) following the protocol's instructions. Then, hospitalized COVID-19 patients in the Intensive Care Unit (ICU) with an rRT-PCR quantification cycle below 25 had their blood samples collected for genomic DNA extraction. Therefore, the population-based study was conducted on 56 ICU-admitted patients. Clinical characteristics were also collected, such as blood pressure, heart rate, respiratory rate, Glasgow scale, leukocyte count, platelets count, creatinine, Simplified Acute Physiology Score III (SAPS3), patient's outcome, sepsis occurrence, and necessity of mechanic ventilation.

2.2. Population study

A total of 56 COVID-19 patients admitted to ICU from Brazil were included in the Hardy and Weinberg equilibrium analysis. Moreover, 3,115 samples from 26 different populations concentrated in five continent groups: African, Americas, East Asian, European, and South Asia, were obtained from the 1000 Genomes Phase 3 database (1000G) (www.internationalgenome.org) for comparison purposes.

2.3. Ethics Approvals

The ethical approvals were obtained from the local ethical committee of the Federal University of Ceará (Ref. No. 4.024.941) following human subjects and complied with the principles of the Helsinki Declaration. Informed consent was obtained before collecting samples from all patients.

2.4. SARS-CoV-2 detection

The samples were collected with a nasopharyngeal swab and placed in a 3 mL tube of viral transportation solution. SARS-CoV-2 RNA was extracted using the QIAamp® Viral RNA kit (Qiagen) according to the manufacturer's protocol for one-step SARS-CoV-2 detection. A 2019-nCoV qPCR Diagnostic Panel from The Centers for Disease Control and Prevention (CDC) with specific primers was used to confirm the presence of SARS-CoV-2 in all samples collected. The primer-probe mixes in this kit target two sections of the Nucleocapsid gene (N1 and N2), as well as the human endogenous control (RNase P gene), which serves as a sample integrity control.

2.5. Genomic DNA Extraction

All COVID-19 patients had a 3 mL peripheral blood sample taken in an EDTA tube. Blood samples were promptly kept at temperatures ranging from 20 Cº to 25 Cº. Only patients with positivity for SARS-CoV-2 detection were considered COVID-19 patients and therefore enrolled in this study. DNeasy Blood K (Qiagen, Hilden, Germany) extracted genomic DNA according to the manufacturer's instructions. The isolated DNA was dissolved in nuclease-free water and kept at 4 degrees Celsius until needed. NanoDropTM was used to verify the quality and integrity of DNA (Thermo Scientific, Waltham, MA, USA). By measuring optical density (OD) at 260 nm (OD260) and 280 nm (OD280), all COVID-19 and control DNA samples were checked for purity (OD280). The 260/280 ratios varied from 1.83 to 1.99, suggesting that the DNA was acceptable.

2.6. ACE2 and TMPRSS2 Genotyping

Genotyping was performed using TaqMan SNP (Single Nucleotide Polymorphism) genotyping assays (ThermoFisher Scientific) using ACE2 SNP probe (SNP ID: rs35803318) and TMPRSS2 SNP probe (SNP ID: rs2070788) also obtained from ThermoFisher® Scientific. Real-time TaqMan PCR and genotyping analyses were performed on an Applied Biosystems Quant Studio 5 system (Applied Biosystems, Foster City, CA, USA) according to the manufacturer's standard PCR protocol.

2.7. Expression quantitative trait locus (eQTL)

We used the Genotype-Tissue Expression v7 (GTEx) tool to explore accessible data for expression quantitative trait loci (eQTL) analysis further to understand the potential functions of the strongly related loci. The GTEx was designed to define human transcriptomes and has generated a reference library of non-diseased tissue gene expression levels, including genotype and gene expression. The multi-tissue eQTL and expression level data were obtained from the GTEx Portal V8 (https://gtexportal.org/home) with 49 different tissues from 15,201 samples. The adaptive permutations mode was used with the setting permutation of 1000, and a false discovery rate (FDR) threshold of ≤0.05 was applied to identify genes with a significant eQTL.

2.8. Statistical analysis

Deviations from Hardy–Weinberg disequilibrium (HWD) were calculated by the chi-square (χ2) goodness-of-fit test. Differences in the ACE2 rs35803318 and TMPRSS2 rs2070788 allele frequencies between population groups were evaluated using Fisher's Exact test. Clinical characteristics and genotype association were assessed using chi-square test and Kruskal-Wallis's test. A p-value < 0.05 was considered significant

Results 

General characteristics of the study subjects 

Fifty-six patients with documented COVID-19-positive PCR admitted to the Intensive Care Unit (ICU) were first analyzed by their clinical characteristics (Table 1). Patients were grouped into a discharged group (patients dismissed from ICU) corresponding to 45 individuals and a deaths group (patients who died at ICU) corresponding to 11 individuals. There were significant differences between discharged and death groups in the following variables: systolic blood pressure, respiratory rate, Glasgow scale, SAPS3, sepsis occurrence, and mechanical ventilation support (Table 1). The prediction power of the prognostic system, SAPS3, was the most exciting result for the death group, in which the score median was 71 against 57 of the discharged group (p<0.0003).

ACE2 and TMPRSS2 SNP distribution in the Brazilian population 

Fifty-six ICU-admitted patients from Brazil were enrolled in this study and compared with 3,115 samples from 26 populations concentrated in five continent groups: African, Americas, East Asian, European, and South Asian. Table 2 and Table 3 display the genotypes and allele frequency distribution for human TMPRSS2 (rs2070788) and ACE2 (rs35803318) polymorphisms that varied significantly among the continental populations and our Brazilian cohort.

It was evidenced that the SNP present in the sequences of ACE2 and TMPRSS2 was in agreeance with the Hardy-Weinberg Equilibrium (HWE) in most all populations (p-value>0.05), except for the ACE2 SNP rs35803318 in the Brazilian population (p-value<0.05) (Table 3). For ACE2 rs35803318, we found 92.6% CC, 3.4% CT, and 4.0% TT genotype carriers in SARS-CoV-2-positive patients, with an altered allele frequency of 5.7% for the T-allele. For TMPRSS2 rs2070788, we found 22.3% GG, 50.7% AG, and 27% AA genotype carriers in SARS-CoV-2-positive patients, with an altered allele frequency of 52.4% for the A-allele (Table 2 and Table 3).

Comparing the allelic frequencies of TMPRSS2 (rs2070788) polymorphism, it was found no significant differences when the Brazilian population was compared with Americas, South Asian, or European populations (p>0.05) (Table 2). Comparing the allelic frequencies of ACE2 (rs35803318) polymorphism, no differences, only between Brazil and the Americas or Europe (p>0.05), were observed. On the other hand, the comparison analysis showed that the Brazilian population had distinct frequencies from African and East Asian populations (p<0.05) (Table 3).

2.9. Expression quantitative trait locus (eQTL)

We further evaluated the eQTL profile of TMPRSS2 and ACE2 using the Genotype-Tissue Expression (GTEx) database. We found only one eQTLs associated with TMPRSS2 SNP (rs2070788) expression in the lung tissue (p = 8.9 x 10^-9) (Figure 1). Interestingly, there were 28 specific tissues associated with SNP of ACE2 (rs35803318), and PIR gene expression was the most significant alteration observed, including in testis (NES=-0.321, p-value=3.9x10^-7), as well as esophagus, skin, heart, lung, and pituitary (Figure 2). Moreover, several system nervous tissues (hypothalamus, cerebellum, amygdala, tibial nerve) had significant eQTL with negative normalized effect size (NES<-0.310, p-value<3.9x10^-4).

ACE2 and TMPRSS2 SNPs were not associated with COVID-19 severity and mortality 

This paper section presents the genotype and allele distributions of different SNPs considering the clinical variables. Table 4 and Table 5 show the association of the SNPs (ACE2 and TMPRSS2) with a categorial variable such as outcome, sepsis occurrence, and mechanical ventilation requirement. Figure 3 and Figure 4 presented the comparison among SNPs genotypes based on interval variables, such as the Glasgow scale, respiratory rate, heart rate, diastolic and systolic pressure, leukocyte and platelets count, and creatinine levels. No association between the clinical variables presented and the SNPs was demonstrated (P>0.05).

3. Discussion

Since COVID-19 was declared a pandemic by the World Health Organization (WHO) on March 11, 2020, the SARS-CoV-2 interaction with human target cells has been studied to comprehend and provide a method for preventing these interactions [9,10]. The first case in Brazil occurred in February from a man who arrived from Italy and presented the characteristic symptoms of the disease. Two days later, around 182 suspect cases occurred in 16 Brazilian states. From April, the contagion was widely distributed. In May, Brazil had already trespassed the number of cases in China, with Ceara State being the third state out of 26 with more cases and the first in the Northeast with 48,489 cases, according to the Ministry of Health [11].

As we already know, SARS-CoV-2 enters the targeted cells using the viral spike (S) protein to interact with a cellular receptor. The virus entry depends on the binding of the S protein to the TMPRSS2 for S protein priming and engages ACE2 as the entrance receptor [2,3]. Furthermore, distinct variations in the genes ACE2 and TMPRSS2 have been observed in various populations worldwide, indicating that differences in COVID-19 infection and severity may be influenced by host genetic backgrounds [6,12,13].

The genetic background of the Brazilian population is quite complex to analyze, given the significant admixture between European, African, and indigenous individuals and the history of invasions, migrations, and forced slavery [14]. Moreover, Ceará state was the first state that freed slaves in Brazil, with a new flow of migration and interaction. This genetic complexity can be an advantage in mapping genes for diseases, as it allows the observation of clinical and biological effects resulting from such an expansive admixture.

Among several SNPs analyses gaining attention in the literature [15,16,17], two SNPs studied, rs2070788 (TMPRSS2) and rs35803318 (ACE2), have been studied in the European population, Peruvian native American and Amazon native American [6,7]. This study aims to describe the distribution of these SNPs in the Brazilian population compared to other continental populations and the impact of these variants on gene expression and clinical parameters. It is worth noticing that our study is the first to discuss these SNPs in the northeast region of Brazil, a populational background with strong African and European ancestry [18].

Our results revealed that rs35803318 (ACE2) was the only SNP in Hardy-Weinberg (HWE) disequilibrium for the Brazilian population. Several factors cause deviations of HWE in a population, such as mutation, natural selection, non-random mating, genetic drift, and gene flow [19]. In a vast population with multiple ethnicities, as in Brazil, non-random mating related to geographical location may be a prevalent source of departures from HWE due to heterozygous deficiency [20]. An Italian study with rs35803318 (ACE2) also revealed significant differences in the frequency distribution in the Italian cohort compared to the American population [5]. Additionally, one of the most important aspects for consideration about ACE2 SNP is its absence in Asian and African Populations (Table 3) which indicates the European contribution for Brazilian polymorphism distribution in this case, same results observed by another study [7].

Moreover, regarding rs35803318 (ACE2) polymorphism, there were no significant differences between Brazil, the Americas, and European populations, which can be explained by the highly heterogeneous and inter-ethnic admixture in the Brazilian population [21]. In contrast, the comparison analysis showed that the Brazilian population had distinct frequencies from the African and Asian populations even though the Africans extensively arrived in Brazil between the 17th and 19th centuries [22]. We also hypothesized that the comparison differences of ACE2 (rs35803318) SNP's allelic frequencies might occur due to the H-W disequilibrium observed [23].

The estimated effect size of an eQTL allele on gene expression is recorded in the GTEx database, which enables the identification of genes whose expression is influenced by genetic variation [24]. Previously, a study found that the presence of the rs2070788 GG genotype, which corresponds to higher TMPRSS2 expression in human lung tissues, increased the risk of severe H1N1 influenza [25]. Then, we first hypothesized that the genotype AA, which corresponds to lower TMPRSS2 expression in the lung, can protect against COVID-19 for populations such as East Asia and Africa based on the eQTL results and the genotype frequencies. In this sense, we also suggested that Brazil and the Americas might be the most susceptible population to SARS-CoV-2 infections since the AA genotype is low frequency.

The trans-regulatory role of SNP rs35803318 (ACE2) was another key finding of our investigation. The presence of the altered allele (T) reduces the expression of the PIR gene, which encodes for the Pirin protein [26]. Pirin causes a significant activation of the NF-B/p65 pathway, which is linked to various chronic inflammatory illnesses [27]. As a result, NF-B p65 pathway hyperactivation causes cytokine storm syndrome and enhances COVID-19 severity [28]. Shoily et al. [29] had already reported another SNP rs4646156 (ACE2) as an eQTL locus that can modify the expression of the PIR gene in various tissues, including the esophagus, thyroid, adipose, and artery. Interestingly, the highest frequency of the TT genotype was found in the Brazilian population analyzed.

Our preliminary results and a few studies have hypothesized that rs2070788 (TMPRSS2) and rs35803318 (ACE2) SNPs are statistically associated with severe illness and poor prognosis in different populations [6,7,30,31]. However, despite it, there was no significant association between clinical parameters and SNPs’ genotype distribution in the Brazilian population. Therefore, on the opposite side of other countries' studies, the Brazilian patient's outcome or severity was not affected by the presence or absence of ACE2 and TMPRSS2 SNPs.

In conclusion, this study showed the genotype and allele frequencies of rs2070788 (TMPRSS2) and rs35803318 (ACE2) SNPs in different populations. Since the presence of polymorphisms with a protective feature against SARS-CoV-2, the Hardy-Weinberg imbalance of polymorphisms in the studied population may have a relevant impact on disease behavior for the Brazilian northeast population. However, in contrast to other studies, there was no significant association between genotype distribution and patient’ outcome and severity.