Genome-wide association studies have led to identification of >160 variants with significant association to coronary artery disease (CAD).¹ One of the variants identified (rs7678555) is located on chromosome 4q27 in proximity of PDE5A and linked to enhanced PDE5A mRNA levels in aorta.² The mechanisms by which the variant affects CAD risk remain unclear.

In an extended family with high prevalence of premature CAD (Figure [A]), we identified by exome sequencing a variant at the PDE5A locus cosegregating with the disease (ie, NM_033430.3:c.19A>T [p.Lys7Ter; gnomAD: 4-120548312-T-A, MAF 2.1·10⁻⁴]). Study approval was obtained by the institutional ethics committee; participating individuals gave informed consent. The data that support the findings of this study are available from the corresponding authors on reasonable request. PDE5A encodes phosphodiesterase 5A. The enzyme degrades the second messenger cyclic guanosine monophosphate (cGMP), which is produced by the soluble guanylyl cyclase, an enzyme that is also genome-wide significantly associated with CAD and was found to be mutated in a large CAD family.³ The variant is located in the first intron of isoforms PDE5A1 and PDE5A3 and the first exon of PDE5A2. In PDE5A2, the variant leads to the exchange of lysine to a stop codon at the seventh amino acid position (Figure [B] and [C]). Thus, we first assumed a loss of PDE5A2 activity. To this end, we cloned the coding PDE5A2 wild-type sequence into an expression vector containing a C-terminal V5-tag and introduced the p.Lys7Ter variant using in vitro mutagenesis. We used HEK 293 cells stably overexpressing soluble guanylyl cyclase to also overexpress PDE5A2^p.Lys7Ter. After activation of soluble guanylyl cyclase using S-nitrosoglutathione, we unexpectedly observed active PDE5A2, as demonstrated by reduced cGMP levels. Using immunoblotting, we found that the variant did not lead to a loss of PDE5A2 protein; rather, we detected a band with reduced protein size (Figure [D] and [E]). To identify putative alternative starts of translation, we performed serial in vitro mutagenesis of the methionine residues downstream of the p.Lys7Ter variant and identified the third methionine (amino acid 92) as an alternative start of translation (Figure [D]). Of note, an alternative start of translation was previously described for PDE3A⁴ but not PDE5A. RNA sequencing of 3 carriers and matched controls did not reveal alternative spliced PDE5A2 transcripts (data not shown). Because the variant is furthermore located in an alternative PDE5A promoter region, we next investigated whether it influences PDE5A2 expression. We performed reporter gene assays using constructs that contained wild-type and mutated alternative promoter sequences (±300 nucleotides 3′ and 5′ of the variant) upstream of the firefly luciferase. The construct containing the p.Lys7Ter alternative promoter displayed increased reporter gene activity compared with the wild-type alternative promoter (Figure [F]). Using droplet digital polymerase chain reaction and allele-specific probes, we validated increased expression of the p.Lys7Ter PDE5A2 allele compared with the wild-type allele in available family members carrying the variant on 1 allele. In available whole blood RNA samples of 3 carriers, more copies of the mutated than the wild-type allele were detected (Figure [G]). Taken together, these data indicate that the variant leads to enhanced expression of an N-terminally truncated PDE5A2 isoform that still is capable of degrading cGMP. To study the effect of the p.Lys7Ter variant on the regulation of PDE5A2 expression, we aimed at identifying transcription factors that specifically bind to the wild-type or the mutated allele. In the eukaryotic promoter database, the zinc finger transcription factor ZFX (Zinc Finger Protein X-linked) was predicted to bind to several sites within the region of interest (Figure [H]). Chromatin immunoprecipitation experiments confirmed binding of ZFX to the alternative PDE5A2 promoter region. Pull-downs using oligonucleotide baits carrying either the wild-type or the mutant allele in HeLa cell nuclear lysates revealed preferential binding of ZFX to the wild-type compared with the mutant allele in western blot analysis (Figure [I–K]). We next silenced ZFX expression using RNA interference with specific small interfering RNAs in vascular smooth muscle cells and observed increased PDE5A2 mRNA levels by quantitative polymerase chain reaction (Figure [L]). Taken together, these data point to a repressive effect of ZFX on PDE5A2 expression in vascular smooth muscle cells, which seems altered in carriers of the mutant p.Lys7Ter allele.

Figure. Association of a PDE5A (phosphodiesterase 5A) variant with coronary artery disease/myocardial infarction in an extended family.A, Family pedigree. The prevalence of traditional risk factors was as follows: diabetes, 0%; hypertension, 39%; low-density lipoprotein cholesterol level, median 143 mg/dL (interquartile range, 126.3–176.3 mg/dL); and body mass index, 23.8 kg/m² (range, 21.9–25.3). The arrow denotes the index patient (III-1). Individuals III-6, III-23, and the index patient underwent exome sequencing. Other family members were genotyped. Variant carrier status is displayed as +/– or –/– in case of heterozygous carrier status or wild-type (WT) PDE5A, respectively. The maximum LOD score was 3.16 at a recombination fraction θ=0.17. B and C, Location of the variant at the PDE5A locus (B) and in the secondary structure of PDE5A2 (C). D, Immunoblot of HEK293 cells overexpressing (pDEST40; Thermo Fisher Scientific) WT PDE5A2 (PDE5A2^WT) and the mutant PDE5A2 (PDE5A2^p.Lys7Ter). To identify alternative starts of translation, 2 methionines downstream of p.Lys7Ter were mutated (ΔMet2: second methionine mutated; ΔMet3: third methionine mutated; ΔMet2/3: second and third methionine mutated) using in vitro Quik Change mutagenesis (Thermo Fisher Scientific). E, Levels of cyclic guanosine monophosphate (cGMP) in HEK 293 cells overexpressing the soluble guanylyl cyclase and PDE5A2^WT, PDE5A2^p.Lys7Ter, or control (no PDE5A). Radioimmunoassay to determine cGMP formation was analyzed after 60 seconds stimulation with S-nitrosoglutathione. Nine experiments and 1-way repeated measures analysis of variance with the Tukey multiple comparisons test were performed. F, Reporter gene assay (pGL4.10; Promega) containing the PDE5A alternative promoter sequence of the WT or p.Lys7Ter alleles. Five experiments and paired t test were performed. G, Copy numbers of PDE5A2 mRNA in whole blood of heterozygous p.Lys7Ter variant carriers. Copy numbers were determined using droplet digital polymerase chain reaction and allele-specific oligonucleotide primers with equal binding affinity to both alleles (Bio-Rad). Copy numbers of the p.Lys7Ter allele were higher in available individuals (n=3) compared with the WT allele (133.5±2.8 vs 116.7±3.4 [copies/µL], respectively; P=0.01). Paired t test was performed. H, Predicted ZFX binding sites (red boxes) at the PDE5A locus (https://epd.epfl.ch/cgi-bin/get_doc?db=hgEpdNew&format=genome&entry=PDE5A_2; accessed November 23, 2020; cutoff P=0.001). I, Differential binding of the transcription factor ZFX to oligonucleotide baits carrying either the WT or mutant allele. The pull-down experiment was carried out in HeLa nuclear lysate and ZFX binding was confirmed in western blot using an anti-ZFX antibody (5419S; Cell Signaling Technology). J and K, Chromatin immunoprecipitation after overexpression of ZFX-V5 in HeLa cells. J, Agarose gel electrophoresis of PDE5A2 alternative promoter sequences after immunoprecipitation of either V5 (αV5) or control (immunoglobulin G [IgG]). K, Quantification of chromatin immunoprecipitation experiments. Five experiments and paired t test were performed. Data are mean and SEM. L, Secondary to silencing of ZFX, PDE5A mRNA levels (as determined by quantitative polymerase chain reaction) were increased in vascular smooth muscle cells (PromoCell). Five experiments and paired t test were performed. Data are mean and SEM. M, Translational perspective: genes encoding the α1 subunit of the soluble guanylyl cyclase (sGC) and PDE5A were found to be associated with coronary artery disease by genome-wide association studies and studies in extended families. Whereas the second messenger cGMP has inhibitory effects on smooth muscle cells and platelets, for example, its formation and degradation can be modulated by sGC stimulators and PDE5A inhibitors, respectively.

We report identification of a variant affecting PDE5A2 expression that is linked to CAD in an extended family. This PDE5A2 gain-of-function mutation adds to the short list of monogenic forms of CAD. This finding strengthens the assumption that PDE5A is the gene mediating CAD risk at the 4q27 locus identified by genome-wide association studies.² Furthermore, it highlights the importance of nitric oxide signaling and the second messenger cGMP in atherosclerosis. Whereas genetic predisposition to enhanced nitric oxide signaling is associated with reduced risk of CAD,⁵ dysfunctional nitric oxide signaling has been shown to increase risk.³ Pharmacologic elevation of cGMP levels via increased production or decreased degradation might represent a promising strategy for prevention or treatment of CAD (Figure [M]). In the clinical setting, the mechanistic role and the therapeutic potential of targeting particularly PDE5A in atherosclerosis remain to be investigated.

Nonstandard Abbreviations and Acronyms

CAD	coronary artery disease
cGMP	cyclic guanosine monophosphate
PDE5A	phosphodiesterase 5A

coronary artery disease

cyclic guanosine monophosphate

phosphodiesterase 5A

This work was supported by Deutsche Forschungsgemeinschaft (SFB1123, B02), the Corona Foundation, the Fondation Leducq (CADgenomics, 12CVD02), the German Federal Ministry of Education and Research (e:AtheroSysMed, 01ZX1313A-2014), the European Union Seventh Framework Program FP7/2007 to 2013 (CVgenes-at-target, HEALTH-F2-2013-601456), and ERA-CVD (Druggable-MI-genes, 01KL1802).

Disclosures Dr Schunkert has received personal fees from MSD Sharp & Dohme, Amgen, Bayer Vital GmbH, Boehringer Ingelheim, Daiichi-Sankyo, Novartis, Servier, Brahms, Bristol-Myers-Squibb, Medtronic, Sanofi Aventis, Synlab, Pfizer, and Vifor T as well as grants and personal fees from AstraZeneca outside the submitted work. Dr Schunkert and Dr Kessler are named inventors on a patent application for prevention of restenosis after angioplasty and stent implantation outside the submitted work. The other authors report no conflicts.

https://www.ahajournals.org/journal/circ

*T.A. Dang and T Kessler contributed equally.

For Sources of Funding and Disclosures, see page 664.

中文翻译：

---

在扩展型心肌梗塞家族中的染色体 4q27 冠状动脉疾病位点鉴定功能性 PDE5A 变体

全基因组关联研究已鉴定出 160 多种与冠状动脉疾病 (CAD) 显着相关的变异。¹已鉴定的变异之一 (rs7678555) 位于染色体 4q27 上PDE5A附近，与主动脉中PDE5A mRNA 水平升高有关。²该变异影响 CAD 风险的机制尚不清楚。

在一个早发 CAD 患病率高的大家族中（图 [A]），我们通过外显子组测序鉴定了与疾病共分离的PDE5A基因座的变异（即 NM_033430.3:c.19A>T [p.Lys7Ter; gnomAD : 4-120548312-TA, MAF 2.1·10 ^-4 ])。研究获得机构伦理委员会批准；参与人员知情同意。相应作者可应合理要求提供支持本研究结果的数据。PDE5A编码磷酸二酯酶 5A。该酶降解由可溶性鸟苷酸环化酶产生的第二信使环鸟苷酸 (cGMP)，该酶在全基因组范围内也与 CAD 显着相关，并被发现在大型 CAD 家族中发生突变。³变体是位于同种型的第一内含子PDE5A1和PDE5A3和第一外显子PDE5A2。在 PDE5A2 中，该变体导致赖氨酸交换为第七个氨基酸位置的终止密码子（图 [B] 和 [C]）。因此，我们首先假设 PDE5A2 活性丧失。为此，我们将编码PDE5A2野生型序列克隆到包含 C 端 V5 标签的表达载体中，并使用体外诱变引入 p.Lys7Ter 变体。我们使用稳定过表达可溶性鸟苷酸环化酶的 HEK 293 细胞也过表达 PDE5A2 ^p.Lys7Ter. 使用 S-亚硝基谷胱甘肽激活可溶性鸟苷酸环化酶后，我们意外地观察到活性 PDE5A2，如 cGMP 水平降低所证明的。使用免疫印迹，我们发现该变体不会导致 PDE5A2 蛋白的丢失；相反，我们检测到蛋白质大小减小的条带（图 [D] 和 [E]）。为了确定可能的替代翻译起点，我们对 p.Lys7Ter 变体下游的甲硫氨酸残基进行了连续体外诱变，并确定了第三个甲硫氨酸（氨基酸 92）作为替代的翻译起点（图 [D]）。值得注意的是，先前为PDE3A ⁴而不是PDE5A描述了另一种翻译开始。3 位携带者和匹配对照的 RNA 测序未显示选择性剪接PDE5A2转录本（数据未显示）。因为该变体还位于替代的PDE5A启动子区域，我们接下来研究了它是否影响PDE5A2表达。我们使用包含萤火虫荧光素酶上游的野生型和突变的替代启动子序列（变异的 3' 和 5' 的±300 个核苷酸）的构建体进行报告基因分析。与野生型替代启动子相比，包含 p.Lys7Ter 替代启动子的构建体显示出更高的报告基因活性（图 [F]）。使用液滴数字聚合酶链反应和等位基因特异性探针，我们验证了 p.Lys7Ter PDE5A2 的表达增加等位基因与携带 1 个等位基因变异的现有家族成员中的野生型等位基因进行比较。在 3 个携带者的可用全血 RNA 样本中，检测到比野生型等位基因更多的突变拷贝（图 [G]）。总之，这些数据表明该变体导致 N 端截短的PDE5A2同种型的表达增强，该同种型仍然能够降解 cGMP。研究 p.Lys7Ter 变体对PDE5A2调控的影响表达，我们旨在鉴定与野生型或突变等位基因特异性结合的转录因子。在真核启动子数据库中，预测锌指转录因子 ZFX（Zinc Finger Protein X-linked）与感兴趣区域内的几个位点结合（图 [H]）。染色质免疫沉淀实验证实了 ZFX 与替代PDE5A 2 启动子区域的结合。在 HeLa 细胞核裂解物中使用携带野生型或突变等位基因的寡核苷酸诱饵进行下拉显示，与蛋白质印迹分析中的突变等位基因相比，ZFX 优先结合野生型（图 [I-K]）。我们接下来让ZFX 保持沉默在血管平滑肌细胞中使用特定小干扰 RNA 的 RNA 干扰表达，并通过定量聚合酶链反应观察到PDE5A2 mRNA 水平增加（图 [L]）。总之，这些数据表明 ZFX 对血管平滑肌细胞中PDE5A2表达的抑制作用，这似乎在突变 p.Lys7Ter 等位基因的携带者中发生了改变。

数字。 大家族中 PDE5A（磷酸二酯酶 5A）变体与冠状动脉疾病/心肌梗塞的关联。A、家庭血统。传统危险因素的患病率如下：糖尿病，0%；高血压，39%；低密度脂蛋白胆固醇水平，中位数 143 mg/dL（四分位距，126.3–176.3 mg/dL）；和体重指数，23.8 kg/m ²（范围，21.9-25.3）。箭头表示索引患者 (III-1)。个体 III-6、III-23 和索引患者接受了外显子组测序。其他家庭成员进行了基因分型。在杂合携带者状态或野生型 (WT) PDE5A 的情况下，变异携带者状态分别显示为 +/– 或 –/– 。在重组分数 θ=0.17 下，最大 LOD 得分为 3.16。B和C，变体在PDE5A基因座 ( B ) 和 PDE5A2 ( C )二级结构中的位置。D，过表达（pDEST40；Thermo Fisher Scientific）WT PDE5A2（PDE5A2 ^WT）和突变体PDE5A2（PDE5A2 ^p.Lys7Ter）的HEK293细胞的免疫印迹。为了确定翻译的替代起点，p.Lys7Ter 下游的 2 个蛋氨酸发生突变（ΔMet2：第二个蛋氨酸突变；ΔMet3：第三个蛋氨酸突变；ΔMet2/3：第二个和第三个蛋氨酸突变）使用体外 Quik Change 诱变（Thermo Fisher Scientific） . 乙，过表达可溶性鸟苷酸环化酶和 PDE5A2 ^WT、PDE5A2 ^p.Lys7Ter或对照（无 PDE5A）的HEK 293 细胞中环磷酸鸟苷 (cGMP) 的水平。在用 S-亚硝基谷胱甘肽刺激 60 秒后分析放射免疫测定法以确定 cGMP 的形成。进行了九次实验和使用 Tukey 多重比较检验的 1 向重复测量方差分析。F，包含WT或p.Lys7Ter等位基因的PDE5A替代启动子序列的报告基因测定（pGL4.10；Promega）。进行了五次实验和配对t检验。G , PDE5A2 的拷贝数杂合 p.Lys7Ter 变异携带者全血中的 mRNA。使用液滴数字聚合酶链反应和对两个等位基因具有相同结合亲和力的等位基因特异性寡核苷酸引物 (Bio-Rad) 确定拷贝数。与 WT 等位基因相比，p.Lys7Ter 等位基因的拷贝数在可用个体 (n=3) 中更高（分别为 133.5±2.8 和 116.7±3.4 [拷贝/μL]；P = 0.01）。进行了配对t检验。H，PDE5A位点预测的 ZFX 结合位点（红框）（https://epd.epfl.ch/cgi-bin/get_doc?db=hgEpdNew&format=genome&entry=PDE5A_2；2020 年 11 月 23 日访问；截止P =0.001） . 一世, 转录因子 ZFX 与携带 WT 或突变等位基因的寡核苷酸诱饵的差异结合。下拉实验在 HeLa 核裂解物中进行，ZFX 结合使用抗 ZFX 抗体（5419S；Cell Signaling Technology）在蛋白质印迹中得到确认。J和K，ZFX-V5 在 HeLa 细胞中过表达后的染色质免疫沉淀。J，V5 (αV5) 或对照（免疫球蛋白 G [IgG]）免疫沉淀后PDE5A2替代启动子序列的琼脂糖凝胶电泳。K，染色质免疫沉淀实验的量化。进行了五次实验和配对t检验。数据是平均值和 SEM。升，继ZFX沉默后，血管平滑肌细胞(PromoCell) 中PDE5A mRNA 水平（通过定量聚合酶链反应确定）增加。进行了五次实验和配对t检验。数据是平均值和 SEM。M，转化观点：通过全基因组关联研究和大家族研究，发现编码可溶性鸟苷酸环化酶 (sGC) 和 PDE5A α1 亚基的基因与冠状动脉疾病有关。而第二信使 cGMP 对平滑肌细胞和血小板具有抑制作用，例如，其形成和降解可以分别通过 sGC 刺激剂和 PDE5A 抑制剂进行调节。

我们报告了影响PDE5A2表达的变体的鉴定，该变体与大家族中的 CAD 相关。这种PDE5A2功能获得性突变增加了 CAD 单基因形式的短名单。这一发现强化了这样一种假设，即PDE5A是介导 4q27 位点 CAD 风险的基因，这一点由全基因组关联研究确定。²此外，它强调了一氧化氮信号和第二信使 cGMP 在动脉粥样硬化中的重要性。一氧化氮信号增强的遗传倾向与 CAD 风险降低有关，⁵功能失调的一氧化氮信号已被证明会增加风险。³通过增加产量或减少降解来提高 cGMP 水平的药理学可能代表了预防或治疗 CAD 的有希望的策略（图 [M]）。在临床环境中，特别针对 PDE5A 在动脉粥样硬化中的机制作用和治疗潜力仍有待研究。

非标准缩写和首字母缩略词

计算机辅助设计	冠状动脉疾病
GMP	环磷酸鸟苷
PDE5A	磷酸二酯酶5A

冠状动脉疾病

环磷酸鸟苷

磷酸二酯酶5A

这项工作得到了 Deutsche Forschungsgemeinschaft (SFB1123, B02)、Corona Foundation、Leducq 基金会 (CADgenomics, 12CVD02)、德国联邦教育和研究部 (e:AtheroSysMed, 01ZX1313A-2014)、欧盟第七框架计划的支持FP7/2007 至 2013（CVgenes-at-target，HEALTH-F2-2013-601456）和 ERA-CVD（Druggable-MI-genes，01KL1802）。

披露Schunkert 博士已从 MSD Sharp & Dohme、Amgen、Bayer Vital GmbH、Boehringer Ingelheim、Daiichi-Sankyo、Novartis、Servier、Brahms、Bristol-Myers-Squibb、Medtronic、Sanofi Aventis、Synlab、Pfizer 和 Vifor T 收取个人费用以及阿斯利康在提交作品之外的赠款和个人费用。Schunkert 博士和 Kessler 博士被任命为一项专利申请的发明人，该专利申请用于在提交的工作之外预防血管成形术和支架植入后的再狭窄。其他作者报告没有冲突。

https://www.ahajournals.org/journal/circ

*TA Dang 和 T Kessler 贡献相同。

有关资金来源和披露信息，请参见第 664 页。