Lack of association of Klotho gene variants with valvular and vascular calcification in Caucasians: a candidate gene study of the Framingham Offspring Cohort

Abstract

Background. Valvular and vascular calcification are important early aging phenotypes and represent risk factors for cardiovascular morbidity and mortality. Klotho is a gene primarily expressed in the kidney that has an important role in calcium–phosphate homeostasis. The functional KL-VS variant of Klotho has been associated with aging and cardiovascular disease in human studies, but its role in valvular and vascular calcification remains unknown. We performed a candidate gene study in the Framingham Offspring Cohort to evaluate the effect of KL-VS variant of the Klotho gene on valvular calcification.

Methods. We analyzed the Klotho KL-VS genotype (rs9536314) from the Affymetrix 550K genome-wide dataset, distributed by dbGAP, on 1389 cases and 2139 controls from the Framingham Heart Study Offspring Cohort. Allele and genotype frequencies were compared between cases and controls. Valvular calcification was defined as presence of calcification on the mitral annulus or the aortic valve as determined by echocardiography. A sensitivity analysis of coronary artery calcification by electron beam computed tomography was performed on 1363 patients.

Results. The frequency of the TT versus the TG allele was not different between the cases and the controls (39 versus 41%). The KL-VS variant of Klotho was not associated with valvular or vascular calcification, despite adequate power to detect association (86% for odds ratios ≥1.2). In sensitivity analyses, no association (P > 0.001) between other common variants of Klotho, β-Klotho or fibroblast growth factor-23 and the end points of valvular or vascular calcification was observed.

Conclusions. In our adequately powered candidate gene study, we did not observe an association with the functional KL-VS variant of Klotho and presence of valvular or vascular calcification. Future studies aimed at combining cohorts with echocardiographic phenotypes need to be conducted to identify genetic variants associated with valvular calcification.

Background

Valvular calcification, particularly mitral annular calcification, remains an independent predictor of cardiovascular and cerebrovascular events [1, 2, 3]. Risk factors for calcification include environmental factors such as cigarette smoking and comorbidities like hypertension, diabetes and the presence of chronic kidney disease [4, 5, 6]. More recently, several studies have demonstrated a significant heritable component to vascular and valvular calcification, and a strong genetic predisposition to developing adverse outcomes related to valvular disease [5, 7, 8]. While these previous studies have used both candidate gene approaches and genome-wide linkage approaches, most of the candidate genes studied have been involved in either lipid metabolism or vascular function and have not focused on calcium–phosphate homeostasis and early aging, both of which may be significant contributors to the pathophysiology of valvular calcification [6, 9, 10].

The human Klotho gene spans 50 kb over Chromosome 13 and encodes a secreted form and a single-pass transmembrane protein. Mice with the Klotho gene knocked out exhibit a syndrome that resembles accelerated aging including atherosclerosis, osteoporosis and extensive vascular and ectopic calcification. In humans, Klotho is primarily expressed in the kidney and interacts with β-Klotho and fibroblast growth factor-23 (FGF-23) and plays a key role in calcium–phosphate homeostasis [11, 12, 13].

The KL-VS functional variant of Klotho represents a common allele that has been associated with early onset coronary disease, elevated systolic blood pressure, serum cholesterol and advanced aging [4, 14, 15, 16]. Given the role of Klotho in calcium–phosphate homeostasis and the presence of vascular and ectopic calcification in the mouse model, it is plausible that the functional variant of the Klotho gene may be related to valvular and vascular calcification [17, 18].

To test this hypothesis, we performed a candidate gene study of the KL-VS variant of the Klotho gene with mitral annular or aortic valve calcification. Data from the Framingham Offspring Study Cohort of healthy Caucasians were utilized in the present investigation [19].

Materials and methods

Study population

The Framingham Offspring Study enrolled 5124 men and women, consisting of children (and their spouses) of the original Framingham Heart Study Cohort in the early to mid 1970s. Details regarding the design and methods of this study have previously been described [19]. The population used for this study included participants who attended the sixth examination cycle between 1995 and 1998, since both cardiac echocardiograms and electron beam computed tomography (EBCT) scans were performed during this examination.

This study was limited to the 3528 Caucasian subjects with genotype data and cardiac echocardiograms with reports of mitral or aortic valve and annular calcification. A subset of patients (n = 1363) had EBCT scans.

Covariates and outcomes

Cardiac echocardiography was performed as part of routine visits according to the Framingham Offspring Cohort protocol. Two-dimensional echocardiograms were performed with a Hewlett-Packard Sonos1000 ultrasound machine and 2.5 MHz transducer. The ultrasound operators were blinded to clinical data [20].

Moderate or severe mitral annular calcification was considered to be present when an echo-dense band around the mitral annulus was >0.3 cm thick on the M-mode or when there was more than one-third circumferential calcification of the annulus in the parasternal short-axis view on two-dimensional study. Moderate or severe aortic annular calcification was considered to be present if more than one half of the annulus demonstrated increased echogenicity and thickening. Moderate or severe aortic sclerosis was present when the aortic cusps had diffuse thickening. The presence of any calcification or the mitral annulus or aortic valve defined case status [20].

Coronary artery calcification was determined using 8-slice multidetector computed tomography (Lightspeed Ultra; GE, Milwaukee, WI), as previously described. The degree of calcification is expressed as an Agatston score [21]. Cystatin C was measured using particle-enhanced immunonephelometry (Dade Behring BN 100 nephelometer), which has an interassay coefficient of variation (CV) of 3.3% and intra-assay CV of 2.4%. Glomerular filtration rate (GFR) was estimated using a validated cystatin C-based equation: estimated eGFR (eGFR) = 76.7*(cystatin C)^−1.18 [22].

Genotyping

The KL-VS variant along with other Klotho, β-Klotho and FGF-23, single nucleotide polymorphisms (SNPs) were genotyped using the Affymetrix GeneChip Human Mapping 500K Array Set supplemented with matched analysis on Affymetrix Human Gene Focused 50k Arrays for a total of ∼550k genotypes. In order to minimize false-positive associations due to genotyping artifact, we limited our analyses to SNPs with a genotyping call rate ≥80% and a Hardy–Weinberg Equilibrium P ≥ 0.001. Given lower statistical power to detect associations with rarer SNPs, we limited our results to SNPs with a minor allele frequency ≥1%.

Statistical analysis

Patients with presence of any calcification on the mitral annulus or the aortic valve on echocardiography were defined as cases. Controls, conversely, had absence of calcification on the mitral and aortic valve. Power calculations were performed using Quanto Software v1.2. Based on the number of cases and controls and using an additive genetic model, our study was adequately powered (86%) to detect odds ratios (OR) of ≥1.2 for the role of this common variant in valvular calcification.

Differences in baseline characteristics between cases and controls were tested using the chi-square and t-tests, wherever appropriate. Unadjusted and age and sex-adjusted logistic regression models were then employed to examine the association between KL-VS variants and case status, using additive, recessive or dominant genetic models.

Sensitivity analyses

A sensitivity analysis for vascular calcification using EBCT criteria was performed. In the analysis, cases were defined based on the presence of coronary or aortic calcification, as calculated by the Agatston score [21]. These scores were log transformed due to their highly skewed distribution. A logistic regression model was used to examine the association between the KL-VS variant and vascular calcification. In sensitivity analyses, we also performed unadjusted and age-adjusted linear regression models to evaluate the association between KL-VS variants and the log-Agatston score.

We also examined the association between the KL-VS variant and the intermediate phenotypes of serum calcium and serum phosphate levels as well as pulse pressure. Additional sensitivity analyses included examining the association between presence of valvular calcification and other common gene variants for the Klotho, β-Klotho and FGF-23 genes. The interaction between the KL-VS variant and serum cystatin C, vitamin D levels or a cystatin-based eGFR was also examined with the end point of valvular calcification.

Finally, we also performed a conditional haplotype analysis using plink, as previously described [23]. Briefly, haplotypes were phased and assembled based on observed multi-SNP identities across regions of interest for each individual in the case and control cohorts. Significance of case/control group observations were adjusted per haplotype using frequency of observation relative to the probability of any given haplotype combination, which is based on the individual SNP allele frequencies as observed within our population.

Results

The clinical characteristics of cases and controls that were obtained from participants of the Framingham Offspring Cohort are presented in Table 1. The average age of the study sample was 58 years, 48% were male and 38% had evidence of mitral annular or aortic valve calcification on echocardiography. Individuals with echocardiographic evidence of mitral annular or aortic valve calcification (e.g. cases) were older, more likely to diabetic, had higher physician measured systolic blood pressure and had a slightly lower eGFR and higher serum cystatin C concentrations.

Table 1.

Study sample characteristics^a

	Cases, N = 1389	Controls, N = 2139	P-value
Ageb, years	62 (54, 69)	56 (50, 63)	<0.001
Male gender, %	48	47	0.61
Smoking status (cigarettes past year)
Yes	206	291	0.52
No	1175	1549	0.52
Diabetes, %	14	7	<0.001
Systolic blood pressure, mmHg	130 (118, 142)	124 (112, 136)	<0.001
Diastolic blood pressure, mmHg	75 (69, 81)	75 (69, 81)	0.59
Body mass index (kg/m2)	27.6 (24.8, 31.0)	26.8 (24.1, 30.2)	<0.001
Total cholesterol (mmol/L)	5.23 (4.66, 5.93)	5.28 (4.68, 5.96)	0.62
High-density lipoprotein cholesterol (mmol/L)	1.22 (0.98, 1.58)	1.29 (1.06, 1.61)	<0.001
Cystatin C (mg/L)	0.95 (0.86, 1.09)	0.91 (0.82, 1.01)	<0.001
eGFRcys, mL/min/1.73m2	82 (69, 92)	86 (76, 97)	<0.001
Calcium (mmol/L)	2.40 (2.33, 2.45)	2.38 (2.32, 2.43)	0.11
Phosphate (mmol/L)	1.14 (1.06, 1.26)	1.13 (1.06, 1.23)	0.05
1-25 Hydroxy vit D (ng/mL)	19 (14, 24)	19 (15, 24)	0.40

^aData are expressed as proportions or median (interquartile range).

^bData available for 3221 subjects.

Table 2.

The association of the Klotho KL-VS variant with echocardiographic evidence of mitral annular or aortic valve calcification^a

SNP	Genotype	Cases, N (%)	Controls, N (%)	OR (95% CI)	OR (95% CI) adjusted for age and sex
rs9536314	TT	900 (65)	1404 (66)	Ref	Ref
	TG	455 (33)	663 (31)	0.93 (0.81–1.08)	0.90 (0.77–1.05)
	GG	34 (2)	72 [22]	1.36 (0.90–2.06)	1.29 (0.84–2.01)
	TT	900 (65)	1404 (66)	Ref	Ref
	TG/GG	489 (35)	735 (34)	0.96 (0.84–1.11)	0.93 (0.80–1.08)
	TT/TG	1355 (98)	2069 (97)	Ref	Ref
	GG	34 (2)	72 [22]	1.39 (0.92–2.10)	1.34 (0.87–2.07)

^aCI, confidence interval.

The KL-VS variant was genotyped and the minor allele frequency was observed in 18% of the study sample. No deviation from Hardy–Weinberg equilibrium was found. (P = 0.81). There was no difference in the frequency of the KL-VS variants between cases and controls (Table 2). The frequency of the TT/TG/GG genotypes were not different between cases and the controls (0.66/0.31/0.03 versus 0.65/0.33/0.02, P = 0.19).

There was no association found between KL-VS variants and mitral valve annular calcification, whether an additive, recessive or dominant genetic model was used. Similarly, no association was found across genotypes with respect to other echocardiographic measures of valvular calcification, intermediate phenotypes of serum calcium and phosphate and pulse pressure or with any measure of valvular, coronary or aortic artery calcification score (Table 3). There was also no evidence of an interaction between cystatin C or vitamin D level, KL-VS genotype status and valvular calcification.

Table 3.

The Klotho KL-VS variant and its effects on cardiac and aortic calcification^a

	TT	TG	GG	Additive P-value	Recessive P-value	Dominant P-value
Echocardiographic calcification
N	2004	1003	99
% Aortic valve calcification	0.6	0.8	2	0.24	0.11	0.33
N	2051	452	32
% Mitral annular calcification	44	46	34	0.81	0.06	0.33
Calcification scores [21]
N	886	423	38
Aortic valve	0	0	0	0.99	0.88	0.99
IQR	0–193.7	0–197.2	0–188.6
N	886	423	38
Mitral valve	0	0	0	0.98	0.89	0.40
Range	0–494.3	0–674.1	0–312.1
N	840	407	38
Coronary artery	36.2	41.3	27.5	0.81	0.77	0.61
IQR	0–257.8	0–235.4	0–448.4
N	895	364	38
Aortic artery	616.8	812.5	406.9	0.38	0.18	0.93
IQR	19.9–2962.3	52.2–2263.3	0–1684.2

^aIQR, interquartile range.

Examining additional SNPs within the Klotho β-Klotho or FGF-23 gene, we identified 26 SNPs that had a minor allele frequency >1% and where genotype data was missing in <10% of subjects (Table 4). No differences in the frequency of these genotypes were noted between cases and controls. In addition, our haplotype analysis failed to identify a single haplotype that was consistently associated with the outcome.

Table 4.

Additional SNPs examined for Klotho β-Klotho and FGF-23^a

			HWE
Gene	SNP	MAF	Chi-square	P-value
Klotho	rs562020	0.34	0.31	0.58
	rs385564	0.33	1.20	0.27
	rs576404	0.40	1.13	0.29
	rs571118	0.47	0.15	0.70
	rs516306	0.38	1.04	0.31
	rs9526983	0.08	6.09	0.01
	rs684492	0.32	0.60	0.44
	rs684868	0.32	0.58	0.44
	rs553791	0.32	0.55	0.46
	rs554634	0.32	0.50	0.48
	rs9527024	0.17	0.15	0.70
	rs9536313	0.17	0.11	0.74
	rs9527025	0.17	0.0007	0.98
	rs648202	0.14	0.18	0.67
β-Klotho	rs900563	0.19	1.24	0.27
	rs12647895	0.19	4.63	0.03
	rs1982738	0.24	0.10	0.75
	rs11940694	0.40	0.24	0.63
	rs1979283	0.34	1.56	0.21
	rs2608819	0.12	0.83	0.36
	rs6854452	0.29	5.27	0.02
	rs1458254	0.15	3.80	0.05
	rs4975017	0.29	4.17	0.04
FGF-23	rs11063112	0.30	0.43	0.51
	rs7955866	0.11	2.68	0.10
	rs13312789	0.12	5.94	0.01

^aSNP, single nucleotide polymorphisms; MAF, minor allele frequency; HWE, Hardy–Weinberg equilibrium; SNPs rs9527023 and rs649964 were genotyped but not analyzed due to >10% missing genotype data; SNPs rs3752472, rs649964 and rs649964 were genotyped but not analyzed due to a MAF ≤0.2%.

Discussion

The Klotho–β-Klotho–FGF-23 pathway plays an active role in calcium–phosphate homeostasis and as a result, represents an important gene candidate for the study of valvular and vascular calcification. In our well-powered candidate gene study for OR ∼1.2, we failed to find an association between the KL-VS variant of Klotho and mitral annular or aortic valve calcification. Our secondary and sensitivity analysis also did not find any associations between common variants of Klotho β-Klotho and FGF-23 and the presence of valvular or vascular calcification. Our findings suggest that the KL-VS variant does not have a major effect on valvular calcification in a relatively healthy Caucasian population.

The Klotho gene and its KL-VS functional variant have been implicated in multiple age-related and cardiovascular disease processes. Studies in multiple populations have suggested that variations in Klotho polymorphisms can lead to early aging phenotypes such as osteoporosis and osteoarthritis. Previous studies in the Bohemian Czech population have also suggested an association with human aging, systolic blood pressure, high-density lipoprotein cholesterol and early onset coronary artery disease and stroke [4, 14, 15]. More recently, there has been increasing recognition that the Klotho gene may also play a role in the progression of chronic kidney disease, carotid atherosclerosis and early mortality on dialysis [24, 25, 26]. While these findings are yet to be replicated in larger studies, little attention has been paid to calcification phenotypes despite the predominant role of the Klotho protein in renal handling of calcium and phosphate, and the strong links between kidney disease, calcification and adverse cardiac outcomes.

Our negative findings regarding the lack of association between the Klotho gene and valvular or vascular calcification have important clinical implications. The lack of association between any of the Klotho SNPs and any end points of valvular and vascular calcification, as well as intermediate phenotypes of serum calcium and phosphate suggests that common variants of the Klotho gene may not affect ectopic calcification in humans in a similar manner to the animal models. Although we were not able to examine Klotho gene expression in our study, there is evidence from cell lines that the SNPs involved in our candidate gene study, have functional significance, and likely do alter the expression of the Klotho protein [4, 27]. However, the function of the Klotho protein, particularly its relationship with β-Klotho and FGF-23, may be too complex for a single variant effect and the true relationship between Klotho and calcification may lie in gene–gene or gene–environment interactions. In our study, we did not observe an interaction between the KL-VS variant status and kidney function as measured by cystatin C; nonetheless, limitations in population size and current statistical methods do not allow us to examine all possible interactions.

To the best of our knowledge, our study is the first to examine the association between the KL-VS functional variant of Klotho and phenotypes of valvular and vascular calcification. Although our primary aim was to evaluate the relationship between the functional variant of Klotho and valvular calcification, we also found no relationships between any SNPs of Klotho/β-Klotho/FGF-23 for any valvular or vascular calcification phenotypes in our study cohort. Our secondary analysis involving phenotypes of more severe mitral annular calcification as well as coronary artery and abdominal aortic calcification and the intermediate phenotypes for serum calcium and phosphate also failed to show moderate–large effects for the candidate gene polymorphisms.

Our analysis has some limitations. We used the Affymetrix 500K chip in a relatively healthy Caucasian population. As a result, these findings cannot be generalized to non-Caucasians. Secondly, if the functional KL-VS variant of Klotho only acts in the presence of certain genetic or environmental factors we would not have captured the relationship, as we are not powered to detect gene–gene and gene–environment interactions. More specifically, we did not have information on vitamin D therapy and parathyroid hormone (PTH) levels in our cohort, if the effect of Klotho is dependent on vitamin D and PTH status, we have may missed the association. Finally, we are also not adequately powered to rule out small effects OR ∼1.2 for the other studied SNPs and as a result, we may have missed associations for other nonfunctional or rare variants.

Conclusions

In summary, we present an adequately powered candidate gene study of the KL-VS variant of the Klotho gene and the phenotype of cardiac valvular calcification. Despite the role of Klotho in calcium–phosphate homeostasis and the previous association studies of the functional SNP with early aging, we did not find a significant genetic effect (OR < 1.2) for mitral annular or aortic valve calcification. Alternate candidate genes for vascular and valvular calcification deserve validation in this well-designed cohort of healthy middle-aged and elderly Caucasians.