Cough associated with angiotensin-converting enzyme inhibitors: the role of pharmacogenetics

Introduction

Angiotensin-converting enzyme inhibitors (ACE inhibitors) have been used to treat cardiovascular and metabolic diseases for over 30 years. In addition to the antihypertensive effect, ACE inhibitors have been shown to reduce the mortality and complications of diseases such as coronary heart disease, heart failure, diabetes, and diabetic nephropathy.

Hypotension, hyperkalemia, dizziness, headache, and persistent dry cough are some of the common side effects of ACE inhibitors. The cough is usually dry, with a tickling or scratchy sensation in the throat, and may occur within a few hours of the first dose or even weeks or months later. The cough reflex may become more sensitive, which affects the severity of chronic cough due to other causes [1]. Cough caused by ACE inhibitors may improve within 1–4 weeks after stopping the drug, but in some cases it may take up to 3 months.

The incidence of dry cough in patients treated with ACE inhibitors is approximately 1.5%–11% and even up to 35% [2]. In fact, not all ACE inhibitor studies included cough as an endpoint. In addition, these studies were limited by small sample sizes and lack of long-term follow-up with some events, leading to marked differences in the study results for cough incidence. Moreover, the incidence of cough varies across individual ACE inhibitor drugs and a large 2019 network meta-analysis (135 RCTs, n=45,420) found a wide range of relative risks versus placebo: RR=1.8 for spirapril, RR=2.46 for trandolapril, RR=2.9 for enalapril, RR=3.11 for captopril, RR=3.18 for perindopril, RR=3.41 for quinapril, RR=3.94 for benazepril, RR=4.39 for lisinopril, RR=5.79 for ramipril, RR=8.5 for fosinopril, RR=13.27 for moexipril [2]. Only a few ACE inhibitors have data from dedicated pharmacoepidemiological studies in real-life clinical practice demonstrating the incidence of cough. In this context, perindopril is an ACE inhibitor for which there is extensive evidence from both RCTs and real-world data. In a series of real-world studies such as PAINT, PIANIST, PROOF, and PETRA, the incidence of cough was very low (range from <0.001 to 0.8%) even with the maximum dose of perindopril. Moreover, in a large pooled analysis of 27,492 patients randomized to perindopril, the incidence of ACE inhibitor-induced cough was 3.9%, and only 3.1% of patients discontinued treatment because of this side effect [3]. In another prospective pharmacoepidemiological study involving 10,380 patients receiving ramipril for a period of no more than 8 weeks, the incidence of cough associated with the use of ramipril was 7.1% [4]. Metadata from 23 RCTs on arterial hypertension and myocardial infarction using zofenopril (n=7249) also showed a low incidence of cough in patients - 2.6% [5], while according to data from 55 RCTs using enalapril (n=23,559), the incidence of cough reached 11.48% [6].

Mechanism of cough caused by ACE-inhibitors

Cough is a well-described class effect of ACE inhibitors. Although the exact mechanism of ACE-induced cough remains unclear, there are several theories about how cough occurs.

ACE inhibition blocks the renin-angiotensin-aldosterone cascade. ACE inhibitors are competitive inhibitors of ACE and prevent the conversion of angiotensin I to angiotensin II. ACE is also responsible for the degradation of bradykinin. Active bradykinin is produced by its precursor kininogen, which is degraded by kallikrein. Bradykinin has a short half-life because it is rapidly degraded by ACE. Thus, the half-life of bradykinin can be prolonged by ACE inhibitors when ACE is inhibited, and its activity and concentration can be increased. The degradation of bradykinin and substance P by ACE and their subsequent accumulation in the upper and lower airways by ACE is the most widely accepted theory. Bradykinin sensitizes the airway sensory nerves by rapidly adapting stretch receptors and C-fiber receptors, which release neurokinin A and substance P. This causes the contraction of the airway smooth muscle, leading to bronchoconstriction and cough [7].

Other proposed mechanisms include bronchial hyperreactivity, asthma, history of drug-induced toxicoderma, chronic heart failure, increased sensitivity of bradykinin-dependent sensory nerve fibers of the airways, increased sensitivity of the cough reflex, and mechanisms involving ACE insertion/deletion polymorphism and bradykinin receptor gene polymorphism [8–10].

Role of the ACE gene polymorphism and bradykinin receptor gene expression in the development of ACE-induced cough

The first step in studying the role of pharmacogenetic factors in the development of cough associated with the use of ACE inhibitors was to analyze the association of polymorphisms of the ACE gene and the bradykinin B2 receptor gene (BDKRB2), including polymorphism in one nucleotide (single nucleotide polymorphism, SNP) or in several fragments of the ACE gene (insertion/deletion, I/D), which may be responsible for the development of this side effect.

Early studies focused on the association of ACE inhibitor-associated cough with the I/D polymorphism in the ACE gene based on data on the role of ACE gene polymorphism in the activity of the renin-angiotensin-aldosterone system. ACE or kininase II is a metallopeptidase whose main known functions are to convert angiotensin I to the vasoactive and aldosterone-stimulating peptide angiotensin II and to inactivate bradykinin. The ACE gene has a frequent I/D polymorphism and is associated with increased activity of this enzyme in plasma and tissues [11]. It has been established that patients with genotype II have the lowest serum ACE levels compared with patients with genotypes ID or DD [11]; therefore, genotype II is associated with an increased risk of cough. It has also been established that the D allele of the ACE gene is associated with increased degradation of bradykinin: the half-life of bradykinin is significantly reduced in the serum of patients with the DD genotype compared to the II genotype (26.3 s versus 42.1 s, p=0.029) [12]. The role of bradykinin in the development of cough associated with ACE inhibitors has also been previously established [7]. Accumulation of bradykinin can cause activation of proinflammatory peptides (e.g., substance P), increase sensitization of the sensory nerves of the respiratory tract, and increase cough.

In a 2000 study in Japan, a retrospective analysis of the bradykinin B2 receptor gene polymorphisms (rs1799722 or 58T/C) in connection with the development of cough induced by ACE inhibitors was found for the T allele (p=0.001) [13].

In two exploratory studies in 2011, the role of various genes involved in bradykinin metabolism was analyzed among the European population of Spain [14, 15]. The first study examined different loci of the ACE gene, bradykinin B1 and B2 receptor genes, and the angiotensin II receptor gene in a group of 281 patients [14]. Among all the genes of interest, only 4 loci of the BDKRB2 gene (rs4900312, rs8016905, rs20695759, rs8013400) showed the involvement of polymorphism in the development of cough caused by ACE inhibitors, with a significance level of p<0.05. At the same time, the only SNP of the BDKRB2 gene (rs8016905) was associated with the development of cough caused by ACE inhibitors and genotype analysis showed an increased risk for the AG genotype versus the GG+AA genotypes (51% of cough cases versus 34.1%, OR=2.21, p=0.003) [14]. Another study examined an even broader set of genes involved in kinin metabolism as well as genes involved in inflammation (nitric oxide synthase genes, individual prostaglandin synthase, and prostaglandin receptor genes) in 249 patients with arterial hypertension [15]. The results revealed the significance of the rs8012552 locus of the BDKRB2 gene (p=0.012), for which the presence of the C-allele is associated with a significant increase in the risk of cough (OR=1.609, 95% CI 1.107–2.340).

Of interest are the results of some meta-analyses that assessed the impact of ACE and bradykinin B2 receptor gene polymorphisms. The first meta-analysis in 2011, which included 6 studies assessing the ACE I/D gene polymorphism (n=1121) and 3 studies assessing the bradykinin B2 receptor gene polymorphism 58 T/C (rs1799722) (n=300), made it possible to establish the presence of a relationship for both polymorphisms for the Asian population (OR=1.49, 95% CI 1.11–2.02 and OR=2.25, 95% CI 1.42–3.57, respectively) [16]. The data on the association between the ACE I/D gene polymorphism and ACE inhibitor-induced cough were confirmed in another meta-analysis that combined 14 studies (n=2623) and showed that the I allele significantly increases the risk of cough in the Asian population (OR=1.40, 95% CI 0.93–2.11) [17].

A 2012 meta-analysis combined data from 11 studies on arterial hypertension (for the period 1992–2009), including 906 cases of cough and 1175 controls without cough [18]. The distribution of the II and DD genotypes of the ACE gene (rs4646994) differed significantly between Asian and Caucasian populations in the studies: the recessive II genotype was 36.4% versus 24.9% (p=0.0001), and the dominant DD genotype was 16.3% versus 28.6% (p<0.0001), respectively. No difference was found in the distribution of the ID genotype between Asians and Caucasians (47.3% versus 46.5%). Pooled analysis revealed a significant 1.5-fold increase in the risk of cough development with ACE inhibitor use in the recessive model II (OR=1.61, 95% CI 1.18–2.20, p=0.003), whereas in the dominant model DD, a nonsignificant 16% increase in the risk of cough development was noted (OR=1.16, 95% CI 0.78–1.74, p=0.46). Comparison of the role of I/D polymorphism also showed a significant effect of the I allele in association with cough (OR=1.33, 95% CI 1.04–1.70, p=0.02), demonstrating that a higher frequency of the I allele leads to increased susceptibility to ACE inhibitor-associated cough. However, subgroup analysis revealed the influence of ethnicity and age for the association between the ACE gene polymorphism and cough: the highest association of the I allele with the development of cough was observed only among Asians over 60 years of age (OR=2.08, 95% CI 1.61–2.69, p<0.00001), while there was no association for the Caucasian race (OR=0.99, 95% CI 0.76–1.29). Thus, the results of the meta-analysis confirmed the opinion that the development of cough is associated with the recessive genotype II and leads to higher concentrations of bradykinin and substance P, provoking the development of cough.

In 2019, the results of a more modern meta-analysis of 26 studies (for the period 1993–2013) assessing the impact of ACE and bradykinin B2 receptor gene polymorphisms were published, in which 1641 cases of cough and 2436 control cases were noted [19]. The role of the I/D polymorphism of the ACE gene was analyzed in 15 studies in European, Asian, and African patients. An association of the recessive I allele with an increased risk of developing cough was confirmed: when comparing the I allele against the D allele, OR=1.45 (95% CI 1.14–1.84, p<0.001). Thus, in patients with genotype II versus ID+DD, the OR was 1.72 (95% CI 1.25–2.37, p=0.001), while when comparing genotype II versus DD, the OR was 1.81 (95% CI 1.13–2.89, p=0.013). Analysis of ethnic subgroups revealed a significant relationship between the I allele and the D allele only for the Asian population (OR=1.74, 95% CI 1.28–2.35, p<0.001). Within the framework of this meta-analysis, 5 studies examined the BDKRB2 gene polymorphism (-58T/C or rs1799722), and the population included only Asians. An association was found for the dominant T-allele versus the C-allele in the development of cough (OR=1.21, 95% CI 0.89–1.63, p=0.031) and for the dominant TT genotype versus the CC genotype (OR=1.45, 95% CI 0.77–2.77, p=0.037).

Other polymorphic variants in the mechanism of cough caused by ACE-inhibitors

The generally accepted pathogenesis of ACE inhibitor-induced cough is associated with a decrease in plasma ACE activity and the accumulation of substance P and bradykinin in the airways [7]. Therefore, previous pharmacogenomic studies have focused mainly on genes associated with ACE and related pathways, including bradykinin receptor genes, prostaglandins, and various metalloproteinases. However, other potential risk genes for ACE inhibitor-induced cough may encode transporter proteins that affect plasma ACE inhibitor concentrations or membrane transport proteins, such as ABO and SCLO1B1 (OATP1B1).

A series of studies have revealed the role of ABO and SLCO1B1 gene polymorphisms in the genesis of cough. An association of the ABO gene polymorphism (rs495828 or G/T) with a significant risk in the presence of the T allele [14] and for the TT genotype (OR=2.69, 95% CI 1.22–5.94, p=0.008) [20] has been shown. Previously, genomic studies have already established a close relationship between the ABO gene and ACE activity [21]. It turned out that the T allele leads to a decrease in ACE activity, and on the one hand, reduces the risk of developing hypertension, and on the other hand, can lead to an increase in bradykinin levels, causing cough in patients receiving ACE inhibitors. In a genome-wide association study (GWAS), the SLCO1B1 polymorphism was strongly associated with an increased risk of enalapril-induced cough. It was shown that carriers of the SLCO1B1 521T>C (rs4149056) C-allele had a doubled risk of developing enalapril-induced cough than carriers of the T-allele (OR=2.02, 95% CI 1.34–3.04), while no significant correlation was found between the SLCO1B1 388A>G (rs2306283) polymorphism [20]. Haplotype analysis showed that compared with SLCO1B1 *1b/*1b carriers, SLCO1B1 *15/*15 carriers had an almost seven-fold higher risk of cough (OR=6.94, 95% CI 1.30–37.07, p=0.020).

The search for candidate genes responsible for the development of cough associated with the use of ACE inhibitors has been carried out in a number of modern genome-wide GWAS studies and has allowed the identification of a number of new genes responsible for the development of cough caused by the use of ACE inhibitors. 

In Sweden, a study was conducted to search for candidate genes for 124 cases of cough from the SWEDEGENE database and a control group (n=1345) [22]. An association of the CLASP1, TGFA and MMP16 gene polymorphism with an increased risk of cough was revealed, but the CLASP1 (cytoplasmic linker-associated protein 1, related to the microtubule group) locus rs62151109 (T/C) polymorphism was found to be significant – OR=3.97 (95% CI 2.39–6.59, p=9.44×10^-8).

In a genome-wide association study in the USA (1595 cases of cough and 5485 controls), the role of the KCNIP4 gene polymorphism was established (association value p <5×10⁻⁶) [23]. The strongest association was at the rs145489027 locus for the minor A allele (OR=1.3, 95% CI 1.2–1.4, p=1.0×10⁻⁸) and was noted for the Caucasian and African American populations. The gene encoding KCNIP4 is expressed predominantly in neuronal tissue and plays an important role in neuronal excitability, being responsible for the activity of voltage-dependent potassium channels in sensory nerves. Thus, the KCNIP4 gene may contribute to susceptibility to the development of ACE inhibitor-induced cough due to innate changes in vagal ion channel activity and airway hyperreactivity. 

The role of this gene KCNIP4 was confirmed in another large international genomic study searching for candidate genes associated with the development of cough associated with the use of ACE inhibitors and their subsequent withdrawal in three large cohorts (Great Britain, Denmark, Iceland, n=78,000) [24]. The association for the KCNIP4 gene also showed high significance (p<0.004). For the rs16870989 locus and the minor A allele of the KCNIP4 gene, an increased risk of cough was observed (OR=1.12, 95% CI 1.09–1.14, p=2.0×10⁻²²), and there was also an increased risk of ACE inhibitor withdrawal due to the development of cough (OR=1.11, 95% CI 1.08–1.14, p=1.4E^-12). For the first time, additional candidate genes that may play a role in the development of ACE inhibitor-associated cough were reported: PREP gene (p <0.004), NTSR1 gene (p<0.004), L3MBTL4 gene (p<0.004), SRBT1 gene (p<0.05), and a role for SLCO1B1 gene polymorphism was confirmed. PREP gene (rs12210271) encodes prolyl endopeptidase responsible for the formation and degradation of several vasoactive and nephron peptides (in addition to bradykinin), including angiotensin, substance P, vasopressin, and neurotensin. Reduced PREP expression may have a cumulative effect on cough risk via the bradykinin mechanism. The NTSR1 gene (rs6062847) encodes the neurotensin 1 receptor, which promotes mast cell degranulation with the release of proinflammatory mediators such as histamine and leukotrienes and the development of bronchoconstriction [25]. The SPTBN1 gene encodes βII-spectrin, which is expressed in large quantities in the brain and is involved in the development of various neurological disorders.

A small Chinese GWAS (n=391) identified other candidate genes for ACE inhibitor-induced cough, PNPT1 and PCGF3 [26]. PNPT1 is expressed in the brain, spinal cord, and lung tissue and encodes for polyribonucleotide nucleotidyl transferase 1. Overexpression of PNPT1 increases cellular production of reactive oxygen species, leading to activation of the NF-κB signaling pathway and release of proinflammatory cytokines. Reactive oxygen species-induced oxidative stress is a stimulus and activator of nociceptive sensory nerve endings that innervate the airways and may trigger the cough reflex. Polymorphic variants of PNPT1 may contribute to ACE inhibitor-induced cough by stimulating the sensory (afferent) nerves. The candidate gene PCGF3 is also expressed mainly in the brain and is involved in neuronal differentiation.

Thus, the role of the neurobiological and/or inflammatory mechanisms in the genesis of ACE inhibitor-associated cough development is demonstrated.

Conclusion

Taken together, the results of all pharmacogenomic studies provide insight into the pathophysiological processes underlying ACE inhibitor-associated cough. In addition, GWAS may provide polygenic prediction of the risk of developing cough and discontinuing ACE inhibitors, which may in the future identify patients at risk of developing ACE inhibitor-associated cough. Assuming that genotyping becomes the standard of care in healthcare systems, the integration of genetic risk data may facilitate more differentiated drug selection and improved treatment efficacy.