Study participants’ demographic and clinical characteristics
The demographic and clinical characteristics of the 110 participants are summarised in Table 2. All participants were in child bearing age (median 28 years, 25th–75th percentile: 23–34). The group of women with a positive CMV DNA (CMV+) status (median 24 kg/m2, 25th–75th percentile: 22–27) had a significantly (p=0.006) lower body max index (BMI) than the group who tested negative for CMV DNA (CMV-) (median 26 kg/m2, 25th–75th percentile: 24–29. CMV+ participants also had significantly lower systolic blood pressure when compared with the CMV- participants. Age, gestational age, parity, gravidity, diastolic blood pressure, pulse rate, income, level of education and HIV status were comparable between CMV+ cases and CMV- controls.
Table 2 Participants’ demographic and clinical characteristics
Characteristic
|
Combined
(n=110)
|
CMV-
(n=74)
|
CMV+
(n=36)
|
P-value
|
Median age in years (25th-75th percentiles)
|
28 (23–34)
|
29 (23–34)
|
28 (23–33)
|
0.85
|
Mean gestational age, weeks ± sd
|
32.3 ± 4.4
|
32.4 ± 4.8
|
32.1 ± 3.5
|
0.73
|
Mean SBP, mmHg ± sd
|
112 ± 13
|
113 ± 14
|
109 ± 9
|
0.037
|
Mean DBP, mmHg ± sd
|
69 ± 10
|
70 ± 10
|
67 ± 9
|
0.13
|
Mean pulse rate, bpm ± sd
|
82 ± 10
|
82 ± 10
|
80 ± 12
|
0.33
|
Median BMI (25th-75th percentiles)
|
25.7 (23.4–28.4)
|
26.3 (24.3–28.8)
|
24.2 (21.7–27.3)
|
0.006
|
Median parity (25th-75th percentiles)
|
1 (0–2)
|
1 (0–2)
|
1 (0–2)
|
0.31
|
Median gravidity (25th-75th percentiles)
|
2 (1–3)
|
3 (2–4)
|
2 (1–3)
|
0.62
|
HIV infected n (%)
|
73 (66)
|
45 (61)
|
28 (78)
|
0.08
|
Median income in USD/month (25th-75th percentiles)
|
235 (168–300)
|
235 (171–300)
|
225 (153–332)
|
0.97
|
Education n (%)
|
|
|
|
0.30
|
Secondary
|
97 (88)
|
67 (91)
|
30 (83)
|
|
Primary
|
9 (8)
|
4 (5)
|
5 (14)
|
|
Tertiary
|
4 (4)
|
3 (4)
|
1 (3)
|
|
Key: CMV=Cytomegalovirus, CMV+=CMV infected, CMV-=CMV uninfected, BMI=Body mass index
Association between SNPs and CMV infection
Genotype data for the 20 SNPs genotyped was available for all 110 participants and the SNP rs113181057 on the IFNAR1 gene was monomorphic in the study population. There was a departure from Hardy-Weinberg equilibrium (HWE) for four of the 20 SNPs: TLR7 rs179008 in cases, TLR2 rs1816702 and IL-6 rs10499563 in the controls and IFNAR1 rs2843710 in both groups. An additional table file shows genotype frequencies in the CMV+ and CMV- negative groups and the univariable logistic regression analyses of SNPs and CMV status [see Additional table 1]. Using the univariate logistic regression analysis of codominant and log additive inheritance models, 4 SNPs (rs10499563 (p<0.001), rs179008 (p<0.001), rs1816702 (p=0.002) and rs352139 (p=0.003) were significantly associated with CMV DNA status (Supplementary Table1). The IL-6 rs10499563T>C polymorphism was significantly associated with lower risk of CMV infection. When compared to the IL-6 rs10499563T/T genotype, the rs10499563T/C was associated with a lower risk of CMV infection as the genotype was significantly (p<0.001) less frequent in the CMV+ group (14%) than the CMV- group (70%). Likewise, the TLR2 rs1816702C>T SNP was significantly associated with lower risk of CMV infection. Genotype rs1816702C/C genotype was significantly (p=0.002) higher in the CMV+ (47%) than the CMV- women (11%)
In contrast, TLR7 (rs179008A>T) and TLR9 (rs352139T>C) polymorphisms were associated with an increased risk of CMV infection. The TLR7 rs179008C/C genotype was significantly higher in the CMV+ group than the CMV- group (31% vs. 3%; p<0.001. With reference to the TLR9 rs352139T/T genotype, both the rs352139T/C and rs352139C/C genotypes were significantly (p=0.005) higher in the CMV+ women (28% and 58% respectively) than in the CMV- women (11% and 47% respectively). These associations remained significant after correction for multiple comparisons (Figure 1). When other models of genetic inheritance were considered, the association of IL-6 rs10499563 maintained significant association with CMV status after Bonferonni correction (BC) in dominant, and overdominant models. SNPs rs1816702 and rs179008 also maintained significance with CMV status after BC in the dominant and recessive models (Figure 1).
Fig. 1 Plot of log10 p-values for the association of gene with CMV DNA across models of genetic associations. For each figure panel, the lower dotted horizontal line is for the nominal p-value threshold for significance (0.05), while the upper dotted blue line for the Bonferroni corrected threshold p-value for significance.
Table 3 shows multivariable logistic regression analysis of SNPs (rs10499563, rs179008, rs1816702 and rs352139) that were associated with CMV status in the univariable analyses. BMI was also included in the model. All SNPs maintained significant association with CMV infection status in at least one of the models. However significant association of rs352139 with CMV status was borderline (p=0.049) in the log additive model while it was not significant in the other models. BMI’s association with CMV status also substantially attenuated in the multivariate logistic regression model (p=0.068).
Table 3 Multivariable adjusted models containing BMI and significant SNPs in univariable analysis
SNP
|
Model
|
Genotype
|
OR (95%CI)
|
p-value
|
TLR2 rs1816702
|
Codominant
|
C/T
|
0.09 (0.02-0.43)
|
0.001
|
T/T
|
0.06 (0.01-0.48
|
|
Dominant
|
C/T-T/T
|
0.08 (0.02-0.37)
|
0.0003
|
Recessive
|
T/T
|
0.32 90.07-1.50
|
0.133
|
Overdominant
|
C/T
|
0.29 (0.08-1.01)
|
0.044
|
Log additive
|
0,1,2
|
0.22 (0.08-0.62)
|
0.001
|
TLR7 rs179008
|
Codominant
|
A/T
|
3.67 (0.79-116.99)
|
0.011
|
T/T
|
18.69 (1.59-220.04)
|
|
Dominant
|
A/T-T/T
|
6.05 (1.53-23.94)
|
0.006
|
Recessive
|
T/T
|
13.15 (1.15-149.74)
|
0.013
|
Overdominant
|
A/T
|
2.27 (0.53-9.68)
|
0.262
|
Log additive
|
0,1,2
|
4.08 (1.46-11.39)
|
0.003
|
TLR9 rs352139
|
Codominant
|
T/C
|
2.87 (0.50-16.58)
|
0.144
|
C/C
|
8.13 (0.90-73.63)
|
|
Dominant
|
T/C-C/C
|
3.58 (0.65-19.66)
|
0.121
|
Recessive
|
C/C
|
3.65 (0.67-19.77)
|
0.124
|
Overdominant
|
T/C
|
1.05 (0.30-3.68)
|
0.938
|
Log additive
|
0,1,2
|
2.85 (0.95-8.58)
|
0.049
|
Il-6 rs10499563
|
Codominant
|
T/C
|
0.05 (0.01-0.25)
|
<0.001
|
C/C
|
0.42 (0.05-3.77)
|
|
Dominant
|
T/C-C/C
|
0.08 (0.02-0.31)
|
<0.001
|
Recessive
|
C/C
|
1.53 (0.21-11.39)
|
0.682
|
Overdominant
|
T/C
|
0.06 (0.01-0.27)
|
<0.001
|
Log additive
|
0,1,2
|
0.19 (0.06-0.58)
|
0.001
|
The IL-6 rs10499563T>C SNP was associated with low likelihood of CMV positivity in codominant (OR=0.05; 95%CI=0.01–0.25, p=0.001) as well as in the log additive, dominant and overdominant models. The result shows the association of the C allele with lower odds of CMV infection even in heterozygous state (rs10499563 T/C). For the TLR7 rs179008A>T SNP, the T allele was significantly associated with higher odds of CMV infection in the codominant model (OR=3.67; 95%CI=0.79–116.99; p<0.011) which, was maintained in the log additive, recessive and dominant models. Hence T allele will likely be associated with CMV+ status in both homozygous and heterozygous states (rs179008T/T and rs179008A/T). The TLR2 rs1816702T>C was significantly associated with decreased risk of CMV positivity both in the codominant (OR=0.05; 95%CI=0.01–0.25, p=0.001) as well as in the log additive, dominant and overdominant models. Hence, risk of CMV infection will be decreased in the homozygous state, rs1816702C/C. The TLR9 rs352139 was significantly associated with likelihood of CMV infection, only in the log additive model (OR=2.85: 95%CI=0.95-8.58; p=0.049).
Comparison of variant allele frequencies from this study with other populations
The variant allele frequencies of the genotyped SNPs were compared with data from two other populations: Asians and Europeans. Table 4 gives variant allele frequencies for the genotyped SNPs in this study as well as for Asians and Europeans as reported on dbSNP.
Table 4 Comparison of variant allele frequencies of genotyped SNPs with other populations
Gene
|
SNP
|
Variant allele
|
Zimbabwean (This study)
|
Other Africans
|
Europeans
|
Asians
|
TLR2
|
rs4696480
|
A
|
0.31
|
0.37
|
0.52
|
0.57
|
TLR2
|
rs3804099
|
T
|
0.45
|
0.36
|
0.56
|
0.72
|
TLR2
|
rs1816702
|
T
|
0.47
|
0.43
|
0.12
|
0.00
|
TLR4
|
rs1554973
|
T
|
0.20
|
0.21
|
0.77
|
0.86
|
TLR4
|
rs2737190
|
A
|
0.14
|
0.16
|
0.33
|
0.37
|
TLR4
|
rs10759932
|
C
|
0.19
|
0.25
|
0.85
|
0.76
|
TLR4
|
rs7856729
|
T
|
0.38
|
0.33
|
0.13
|
0.10
|
TLR7
|
rs179008
|
T
|
0.23
|
0.12
|
0.23
|
0.00
|
TLR9
|
rs352139
|
C
|
0.42
|
0.61
|
055
|
0.40
|
TLR9
|
rs5743836
|
G
|
0.36
|
0.42
|
0.13
|
0.00
|
TLR9
|
rs187084
|
G
|
0.30
|
0.29
|
0.43
|
0.40
|
TLR9
|
rs352140
|
T
|
0.32
|
0.29
|
0.55
|
0.39
|
IL-6
|
rs10499563
|
C
|
0.31
|
0.27
|
0.23
|
0.16
|
IL-6R
|
rs4537545
|
C
|
0.26
|
0.34
|
0.37
|
0.32
|
IL-10
|
rs1800872
|
T
|
0.40
|
0.44
|
0.24
|
0.68
|
IL-10
|
rs1878672
|
C
|
0.25
|
0.26
|
0.45
|
0.05
|
IL-28B
|
rs12979860
|
G
|
0.70
|
0.82
|
0.86
|
0.97
|
IFNAR1
|
rs2843710
|
G
|
0.30
|
0.31
|
0.41
|
0.36
|
IFNAR1
|
rs113181057
|
C
|
0.00
|
N/A
|
N/A
|
N/A
|
IL-1A
|
rs1800587
|
C
|
0.65
|
0.60
|
0.71
|
0.93
|
Key: SNP=Single nucleotide polymorphism, TLR=Toll-like receptor, IL=Interleukin, IFNAR=Interferon α