Severe acute respiratory syndrome (SARS) is the first pandemic of the 21st century
(1). Since its recognition, 8437 individuals have been affected and 813 have died
(2). Approximately 20–30% of patients required intensive care admission (1). Although
there was a slight predominance of female SARS patients, possibly because of the overrepresentation
of female healthcare workers (1), male SARS patients were more likely to suffer poor
outcomes (3). In a major hospital outbreak in Hong Kong (4), 32% of male and 15% of
female SARS patients required intensive care or died. Remarkably, similar demographic
data were seen among SARS patients in the greater Toronto area, Canada, where 32%
of males and 14% of females with SARS required intensive care or died (5). Karlberg
et al. (3) studied the case fatality rates among all confirmed SARS patients documented
in the Hong Kong SARS epidemic in 2003. The authors concluded that the mortality rates
differed significantly between males and females, being 21.9% and 13.2%, respectively.
The relative risk for death in males was 1.62 after adjustment for age. It is thus
an intriguing coincidence that ACE2, the gene for the newly identified functional
receptor for the SARS coronavirus, angiotensin-converting enzyme 2, maps to the X-chromosome
(Xp22) (6).
ACE2 was first identified as a homolog of angiotensin-converting enzyme with zinc
metalloproteinase activity (7). Many of its activities differ from those of angiotensin-converting
enzyme (8). ACE2 has been found to be an important regulator of cardiac function (9).
Since the identification of ACE2 as the functional receptor for the SARS coronavirus
(6), efforts have been spent on characterizing its molecular interaction with the
virus (10)(11). On the other hand, studies on mouse hepatitis virus, a group 2 coronavirus
(12), demonstrated that allelic variants of viral receptor were associated with altered
virus-binding activity, which mediated host susceptibility (13). Hence, it is plausible
that genetic variants of ACE2 may moderate the effects of SARS coronavirus infection
and, possibly, gender-specific effects.
For this study, we obtained institutional ethics approval. We identified 103 single-nucleotide
polymorphisms (SNPs) in ACE2, using the University of Washington and Fred-Hutchinson
Cancer Research Center Variant Discovery Resource, SeattleSNPs (14). Two of the identified
SNPs were located within the coding regions [dbSNP identification nos. rs4646116 and
rs4646179 (15)], whereas the remainder were located within the introns of ACE2. SNP
validation by direct sequencing of the 101 noncoding SNPs was conducted with use of
buffy coat DNA obtained from 10 female Chinese volunteers. This validation strategy
allows the detection of SNPs with minor allele frequencies of at least 10% at 90%
power. The two coding SNP loci were verified on buffy coat DNA from 20 female Chinese.
Forty-eight pairs of primers were designed to facilitate direct sequencing of ACE2
regions spanning all of the SNPs. Buffy coat DNA was extracted according to the Blood
and Body Fluid Spin protocol in manufacturer’s instructions for the QIAamp DNA Blood
Mini Kit (Qiagen). DNA sequencing was performed by the dideoxy dye terminator method
on an automated DNA sequencer (3100 Genetic Analyzer; Applied Biosystems) based on
capillary electrophoresis. Sequences were edited and aligned, and comparisons were
made with the SeqScape software (Applied Biosystems).
SNP validation confirmed sequence variations at five sites (Table 1 ). All five SNP
loci were noncoding. Both of the coding SNPs were shown to be nonpolymorphic among
the 20 females. The positions and orientations of the five verified SNPs are illustrated
in Fig. 1 . A case–control study was conducted to compare the frequencies of the five
polymorphisms among 168 SARS patients (81 males and 87 females, among whom 30 males
and 16 females had poor outcomes) from the Prince of Wales Hospital, Hong Kong (4),
and 328 healthy volunteers (174 males and 154 females). All of the individuals studied
were unrelated individuals of Chinese ethnicity. Genotype characterization was performed
with TaqMan (Applied Biosystems) allelic discrimination assays on an ABI Prism 7900HT
sequence detection system (Applied Biosystems). Each assay consisted of two allele-specific
minor groove binding probes with different fluorescent labels, i.e., 6-carboxyfluorescein
(FAM) or VIC™, designed for the discrimination of the two respective alleles at each
SNP locus. The assays were set up according to the manufacturer’s instructions (TaqMan
Core PCR Kit; Applied Biosystems) in a reaction volume of 10 μL. The primers and fluorescent
probes were used at concentrations of 900 and 200 nM, respectively. We used 10 ng
of buffy coat DNA for amplification. The thermal profile consisted of an initial denaturation
period for 10 min at 95 °C, followed by 40 cycles of denaturation at 92 °C for 15
s, and 1 min of combined annealing and extension at either 62 °C (SNPs rs2285666,
rs4646142, and rs714205) or 65.5 °C (SNPs rs2106809 and rs2074192; Table 1 ). The
genotypes were scored with the SDS2.1 software.
Table 1.
Allele and genotype frequencies for the studied groups.
A. Allele frequencies and χ2 analyses (df = 1)
SNP locus1 (dbSNP ID)
1848 (rs2106809)
95701 (rs2285666)
16854 (rs4646142)
360242 (rs714205)
37138 (rs2074192)
Frequency, n (%)
χ2
Frequency, n (%)
χ2
Frequency, n (%)
χ2
Frequency, n (%)
χ2
Frequency, n (%)
χ2
Minor allele
T
G
C
G
A
Controls (174 M/154 F)
228 (47)
224 (46)
223 (46)
215 (45)
206 (43)
SARS patients3 (81 M/87 F)
114 (45)
χ2 = 0.354; P = 0.552
111 (44)
χ2 = 0.470; P = 0.493
110 (43)
χ2 = 0.539; P = 0.463
111 (44)
χ2 = 0.047; P = 0.840
105 (41)
χ2 = 0.109; P = 0.741
SARS patients with poor outcome3 (30 M/16 F)
30 (48)
χ2 = 0.000; P = 0.979
27 (44)
χ2 = 0.090; P = 0.765
27 (44)
χ2 = 0.072; P = 0.788
26 (42)
χ2 = 0.069; P = 0.793
25 (40)
χ2 = 0.051; P = 0.821
Male SARS patients with poor outcome4 (n = 30)
13 (43)
χ2 = 0.590; P = 0.443
12 (40)
χ2 = 0.488; P = 0.485
12 (40)
χ2 = 0.410; P = 0.522
11 (37)
χ2 = 0.746; P = 0.388
11 (37)
χ2 = 0.401; P = 0.527
Male controls (n = 174)
92 (53)
85 (49)
84 (48)
82 (47)
78 (45)
Female SARS patients4 (n = 87)
78 (45)
χ2 = 0.002; P= 0.962
77 (44)
χ2 = 0.008; P = 0.928
77 (44)
χ2 = 0.008; P = 0.928
78 (45)
χ2 = 0.064; P = 0.799
73 (42)
χ2 = 0.000; P = 0.991
Female controls (n = 154)
136 (44)
139 (45)
139 (45)
133 (43)
128 (42)
B. Genotype frequencies and χ2 analyses (df = 2)
Genotype
CC/CT/TT
AA/AG/GG
GG/GC/CC
CC/GC/GG
GG/AG/AA
Female controls (n = 154)
48/76/30
χ2 = 0.280; P = 0.878
44/81/29
χ2 = 0.099; P = 0.952
44/81/29
χ2 = 0.099; P = 0.952
44/87/23
χ2 = 1.102; P = 0.576
48/84/22
χ2 = 2.315; P = 0.314
Female SARS patients4 (n = 87)
25/46/16
25/47/15
25/47/15
26/44/17
1/39/17
1
The SNP loci are identified according to the University of Washington ACE2 sequence
coordinates (14), and the relevant identification numbers at the SNP database (15)
are in parentheses.
2
Genotyping performed with Applied Biosystems Assays-on-Demand™, C2551616_1 and C2551626_1.
3
The combined male (M) and female (F) control group is used as the comparison group.
4
The control group of the same gender is used as the comparison group. Because males
are hemizygous for ACE2, the number of alleles is equivalent to the number of males.
Because females have two X chromosomes, the number of alleles is twice the number
of females. At each SNP locus, only the minor allele frequency is listed.
Figure 1.
Schematic illustration of the genomic organization of ACE2 and positions of the studied
SNP loci.
ACE2 contains 18 exons. Each exon is represented by an open box. The arrows mark the
positions of the five verified SNPs. All five SNPs are located within noncoding regions
of ACE2. The individual SNPs are named according to their identification numbers registered
at the SNP database (dbSNP) (15).
The allele frequencies for the SARS and control groups are listed in Table 1 . When
we used the allele frequencies obtained from the control group, the group sample size
provided a power of at least 80% for the determination of a genetic factor that contributes
50% increased likelihood toward the development of SARS or poor outcome with 95% confidence.
Statistical significance among groups was examined by the χ2 test for each SNP locus
(SigmaStat, Ver. 3.0; SPSS). Statistical significance was denoted by a two-tailed
P <0.05. No significant difference was observed in the allele distributions between
the female and male controls (data not shown), between the SARS cases and controls,
between SARS cases with poor outcomes and controls, between the male SARS patients
with poor outcome and the male controls, or between the female SARS patients and female
controls (Table 1 ). The observed genotype distributions for each of the five loci
among the female controls did not deviate significantly from those expected from the
Hardy–Weinberg equilibrium. The genotype frequencies for each of the five SNP loci
were not statistically significantly different between the female SARS patients and
the female controls. Because males are hemizygous for ACE2, the genotype frequency
is equivalent to the allele frequency.
We therefore conclude that although ACE2 serves functionally as the receptor for entry
of the SARS coronavirus into human host cells, the evidence provided by this study
does not support an association between its common genetic variants and SARS susceptibility
or outcome. Despite its X-chromosome location, poor outcomes in male SARS patients
do not appear to be related to genetic variants of ACE2.