To the editor,

The novel Coronavirus disease 2019 (COVID-19) is the causative agent of a severe respiratory infection, which is of global public health concern [1, 2]. To date (17 April 2020), COVID-19 has resulted in a total of 2,175,460 laboratory-confirmed human infections, including 146,536 deaths. Moreover, cancer has been already identified as an individual risk factor for COVID-19 [3, 4]. However, a systematic analysis of ACE2 aberration remained uncharacterized in human cancers.

We then curated a pan-cancer analysis of ACE2 in malignancies. In TCGA pan-cancer panel, the most frequent DNA alteration is mutation. Mutations were mainly distributed in UCEC, SKCM, UCS and STAD (Figure S1A). In another pan-cancer panel, the most frequent DNA alteration is amplification. Amplifications of ACE2 were observed in NEPC and PRAD patients (Figure S1B). In addition, ACE2 mutations in malignancies were distributed across all exons of ACE2 without hot spot mutation site (Figure S2AC, Table S1). The most frequent mutation was H195Y/X195_splice (Figure S2B) and X555_splice (Figure S2D). Two most frequent mutations were distributed in the peptidase domain of ACE2, which were predicted to be inactivating mutations.

We next compared ACE2 expression in tumor and its normal control tissue. ACE2 expression was upregulated in six tumors while downregulated in three tumors (Table S2, Fig. 1a). Notably, COAD, KIRP, PAAD, READ, STAD and LUAD presented with increased ACE2 expression (Fig. 1b). Notably, because COVID-19 was mainly transmitted through air-way, we focus on respiratory system tumors. ACE2 was significantly upregulated in LUAD while remained unchanged in lung squamous cell carcinoma (Figure S3). Three tumors, including TGCT, THCA, and KICH, presented with decreased ACE2 expression (Fig. 1c). We then confirmed the relevance of genetic disorders and ACE2 expression. We found that mutations were not relevant to RNA expression (Figure S4A). In addition, we found DNA copy variation were neither statistically relevant to ACE2 expression (Figure S4B). Thus, it is possible that the upregulation of ACE2 expression was not resulted from genetic variation. We then examined the epigenetic disorders of ACE2 in cancers.

Fig. 1
figure 1

RNA expression aberration of ACE2 in tumors. a RNA expression aberration of ACE2 in tumors using GEPIA2 (upper panel) and Ualcan (lower panel) database. Red, tumor samples; grey, normal control samples. b Colon adenocarcinoma (COAD), kidney renal papillary cell carcinoma (KIRP), pancreatic adenocarcinoma (PAAD), rectum adenocarcinoma (READ), lung adenocarcinoma (LUAD), and stomach adenocarcinoma (STAD) presented increased ACE2 expression. *p < 0.05. This data was obtained using GEPIA2. Red, tumor samples; grey, normal control samples. c Testicular germ cell tumors (TGCT), thyroid carcinoma (THCA), and kidney chromophobe (KICH) presented decreased ACE2 expression. *p < 0.05. This data was obtained using GEPIA2. Red, tumor samples; grey, normal control samples.

Full size image

Four probes in ACE2 promoter were used for detecting DNA methylation level of ACE2 (Fig. 2a). We have found that the five highly ACE2 express tumors presented with decreased DNA methylation level of ACE2, including COAD, KIRP, PAAD, READ, and LUAD (Fig. 2b). Accordingly, an ACE2 downregulated tumor, TGCT, presented increased DNA methylation level (Fig. 2c). Since there is no available DNA methylation dataset for normal control of KICH, we could not compare global DNA methylation level of KICH. Instead, we compared DNA methylation level of KICH across different tumor stages, and we found that the DNA methylation was not significantly changed (Figure S5). In addition, DNA methylation level of ACE2 in THCA and STAD remained unchanged (Figure S6AB), which suggested DNA methylation might be not the only reason of abnormal ACE2 expression. Other possibilities, such as histone modifications [5] and glycosylation [6] may give rise in the abnormal expression of ACE2, which requires further explorations.

Fig. 2
figure 2

DNA methylation aberration of ACE2 in tumors. a Probes for detecting DNA methylation of ACE2 promoter. b Four highly ACE2 express tumors presented with decreased DNA methylation level of ACE2, including colon adenocarcinoma (COAD), kidney renal papillary cell carcinoma (KIRP), pancreatic adenocarcinoma (PAAD), and rectum adenocarcinoma (READ). *p < 0.05. This data was obtained using Ualcan. c An ACE2 downregulated tumor, testicular germ cell tumors (TGCT), presented increased DNA methylation level. *p < 0.05. This data was obtained using Ualcan.

Full size image

We then explored the impact of ACE2 expression in overall survival (OS) and disease-free survival (DFS). Six tumors (COAD, KIRP, PAAD, READ, STAD, and LUAD) presented with elevated ACE2 expression while three other tumors (TGCT, THCA, KICH) presented with decreased ACE2 expression; however, ACE2 expression was not statistically relevant to patients’ prognosis, neither in DFS (Figure S7) nor OS (Figure S8). For DFS, higher ACE2 expression predicted better outcome in KIRC, LIHC, LUSC, UCS, and OV groups (Figure S9A). For overall survival, higher ACE2 expression indicated better prognosis in KIRC, LIHC, and OV groups. However, higher ACE2 expression in LGG refers to unfavorable outcome, which indicated the ACE2 might function as a dual-edged sword for patients’ outcome (Figure S9B). Notably, ACE2 has been proven to be an important regulator in tumorigenesis. For instance, ACE2 inhibits breast cancer angiogenesis through suppressing VEGFa/VEGFR2/ERK pathway [7] and reduces cell invasion and migration in NSCLC cells [8].

Among these COVIDs with malignancies, LUAD was the most frequent type [3]. Moreover, patients with lung cancer were confirmed to harbor a higher incidence of COVID-19, with severer symptoms [3, 4]. Here, we have also proved that the ACE2 RNA expression was upregulated in LUADs. Since our study is only a database analysis, further validation in larger clinical cohort is warranted.