Patients With Coronary Microvascular Dysfunction Have Less Circulating α‐Klotho

Abstract

Background

Coronary microvascular dysfunction (CMD) represents an early functional characteristic of coronary vascular aging. Klotho (α‐klotho) is a circulating protein inversely linked to physiological aging. We examined low klotho as a potential marker for vascular aging in patients with CMD and no coronary artery disease.

Methods and Results

Patients undergoing nonurgent angiogram for chest pain who had no coronary artery disease underwent invasive coronary microvascular and endothelial function testing. CMD was defined by ≤50% increase in coronary blood flow (percentage change in coronary blood flow) in response to intracoronary acetylcholine or coronary flow reserve ≤2. Fresh arterial whole blood was used to analyze circulating endothelial progenitor cells with flow cytometry. Stored arterial plasma was used for klotho analysis by ELISA. Participants with CMD (n=62) were compared with those without CMD (n=36). Those with CMD were age 55±10 years (versus 51±11 years; P=0.07) and 73% women (versus 81%; P=0.38). Traditional risk factors for coronary artery disease were similar between groups. Patients with CMD had less klotho (0.88±1.50 versus 1.75±2.38 ng/mL; P=0.03), and the odds of low klotho in CMD were significant in a logistic regression model after adjusting for traditional cardiovascular risk factors (odds ratio [OR], 0.80 [95% CI, 0.636–0.996]; P=0.05). Higher klotho was associated with higher numbers of endothelial progenitor cells with vascular regenerative potential (CD34⁺ and CD34⁺CD133⁺KDR⁺). Among a subgroup of patients with atherosclerotic cardiovascular disease risk <5% (n=58), CMD remained associated with lower klotho (OR, 0.80 [95% CI, 0.636–0.996]; P=0.047).

Conclusions

Klotho may be a biomarker for CMD and may be a therapeutic target for groups of patients without significant traditional cardiovascular risk.

Nonstandard Abbreviations and Acronyms

cEVA

early coronary vascular aging

CFR

coronary flow reserve

CMD

coronary microvascular dysfunction

EPC

endothelial progenitor cell

FGF‐23

fibroblast growth factor‐23

klotho

α‐klotho

noCAD

nonobstructive coronary artery disease

Clinical Perspective.

What Is New?

• Klotho (α‐klotho), a protein inversely correlated with aging, is more likely to be low in patients with coronary microvascular disease in the absence of obstructive coronary artery disease (a feature of early coronary artery aging).

• Higher levels of klotho were associated with higher levels of circulating endothelial progenitor cells expressing vasculoprotective markers.

What Are the Clinical Implications?

• Klotho may be a marker for early coronary artery disease in patients who may be considered at low cardiovascular risk by traditional risk‐stratification measures.

Vascular aging is a key characteristic of cardiovascular disease, which remains the leading cause of death worldwide. ¹ Patients with early coronary vascular aging (cEVA) may first develop subclinical coronary artery disease (CAD) characterized by pathophysiological vascular changes that are typically associated with advanced chronologic age. ² One of the earliest detectable functional features of CAD is coronary microvascular dysfunction (CMD). ³ Patients with CMD in the absence of obstructive coronary artery disease (noCAD) may present with signs and symptoms of myocardial ischemia. ⁴ Despite the absence of traditional risk features, patients with CMD are at significant age‐dependent risk for developing myocardial perfusion abnormalities, adverse cardiovascular events, stroke, and heart failure. ⁵ , ⁶ , ⁷ , ⁸ Thus, identification of vascular injury and repair markers are critical to evaluating vascular health in cEVA.

Endothelial progenitor cells (EPCs) and protein klotho (α‐klotho) have been associated with vascular injury and repair. EPCs likely mobilize from bone marrow to act on vascular endothelium by paracrine mechanisms. Their role in vascular injury and repair may be related to cell marker expression. Circulating EPCs expressing CD34 and KDR (vascular endothelial growth factor receptor 2) have been linked to in vitro vascular repair, angiogenesis, and better cardiovascular outcomes. ⁹ We hypothesized that risk for cEVA may also be linked to klotho. Klotho has piqued significant interest in biological and clinical aging research as it exhibits significant age‐related pleiotropy in cardiovascular, renal, and metabolic systems, and others. ¹⁰ Although it is considered physiological for soluble klotho levels to decrease with natural aging in humans, lower circulating klotho levels have been observed in disease states that are associated with cEVA, such as coronary artery disease and arterial stiffness, cardiovascular mortality, and all‐cause mortality. ¹¹ Higher circulating klotho in humans has been associated with longevity, less frequent cardiovascular‐associated and all‐cause mortality, improved health outcomes, and fewer macrovascular complications in some populations. ¹²

Although observational studies have demonstrated an association between low klotho levels and advanced coronary vascular aging, the association of klotho and CMD in humans remains unknown. Thus, the current study aims to investigate the potential association between circulating klotho and CMD without CAD.

METHODS

Supportive data may be shared by the corresponding author on reasonable request.

Ethical Statement

This study was approved by the Mayo Clinic Institutional Review Board (approval number 22‐008605‐14). Written informed consent was obtained from all participants before a previously scheduled, clinically indicated invasive coronary angiogram. All research conduct was aligned with the Declaration of Helsinki guidelines.

Study Design and Recruitment

A retrospective observational study examined consecutive patients enrolled in a prospective database at Mayo Clinic (Rochester, MN), as previously described. ¹³ Adult patients who were previously scheduled for a clinically indicated invasive coronary angiogram for chest pain were recruited to voluntarily participate in this study. Exclusion criteria included acute coronary syndromes (ie, acute myocardial infarction or unstable angina), heart failure with reduced left ventricular ejection fraction <50%, severe liver disease, severe renal disease, Paget disease, or bone fracture within 5 years of enrollment. Eligible participants with noCAD observed during coronary angiogram (<40% coronary artery stenosis) then underwent invasive coronary microvascular assessment. Participants selected from the study database met all eligibility criteria, had previous EPC analysis performed using fresh cells obtained during study enrollment, and had stored plasma available for analysis. Participants meeting criteria between the years 2003 and 2013 were identified (n=115) in a preexisting retrospective database as participants with CMD who were exact matched 2:1 for age and sex (match tolerance of 0) with references who were recruited in the same manner and did not have CMD. Of these eligible participants retrospectively selected, a total of 98 had stored plasma available for duplicate immunoassay testing.

Invasive Coronary Function Testing

The invasive coronary function testing protocol has been described in detail elsewhere. ⁸ , ¹⁴ Patients with noCAD observed by diagnostic angiography first underwent endothelium‐dependent coronary vasoreactivity testing. A Doppler guidewire (0.014‐inch FloWire, Philips/Volcano, Inc) advanced just distal to the tip of a coronary‐infusion catheter (Ultrafuse, SciMed Life Systems) was positioned within the mid left anterior descending artery. Consecutive intracoronary boluses of increasingly concentrated acetylcholine were selectively infused within the left anterior descending artery. Doppler measurements were obtained following each infusion. The percentage change in epicardial coronary artery diameter was recorded. Coronary blood flow was calculated as follows: π×average peak velocity×coronary artery radius. The percentage change in coronary blood flow was considered the percentage difference between coronary blood flow at maximal hyperemia following acetylcholine injection and basal flow. Endothelium‐independent coronary vasoreactivity was assessed by coronary flow reserve (CFR). CFR equated to the ratio of maximal hyperemia following intracoronary adenosine (36–72 μg) and basal flow. As previously defined, CMD was considered a percentage change in coronary blood flow ≤50% in response to intracoronary acetylcholine or CFR ≤2. The threshold for abnormal percentage change in coronary blood flow ⁵ has been commonly used by clinicians and investigators to assess coronary endothelial dysfunction based on the observed relationships with cardiovascular morbidity, mortality, and risk association. ⁵ , ⁶ , ⁸ , ¹⁴

Clinical Data and Biochemical Testing

Clinical data were obtained by questionnaire at time of study enrollment and electronic medical record. The atherosclerotic cardiovascular disease (ASCVD) risk estimator referred to 10‐year cardiovascular risk for participants aged 60 to 79 years and lifetime cardiovascular risk for participants aged 20 to 59 years. Coronary angiogram imaging was used to calculate a Gensini score by a blinded operator to assess degree of atherosclerosis in noCAD. ¹⁵ Routine biochemical testing was performed within a 2‐week period before routine coronary angiogram. Arterial blood samples were obtained from either arterial access sheath used during coronary angiogram or the aorta. Fresh arterial whole blood was used for EPC analysis immediately after collection. Remaining arterial blood samples were partitioned, and plasma was stored in EDTA at −80 °C per study protocol. Stored plasma from the aortas was used for human soluble klotho protein (catalog number 8050; MicroVue) and human fibroblast growth factor‐23 (FGF‐23) C‐terminus (catalog number 60‐6100; MicroVue) ELISA testing in duplicate, per manufacturer's protocol. Immunoassays were performed in duplicate. All analyses were conducted by personnel blinded to patient characteristics and group allocation.

Flow Cytometry

The circulating endothelial progenitor cells were assessed by flow cytometry, as previously described. ¹³ , ¹⁶ Briefly, fresh arterial blood obtained during coronary angiogram was analyzed. Mononuclear cell isolation was performed using a Ficoll density gradient. Cell fluorescence (FACSCalibur, Becton Dickson) was measured immediately following immunofluorescent staining by CD34‐PerCP Cy 5.5 (Becton Dickson) and KDR‐APC (R&D Systems) by a laboratory technician blinded to patient study data using CellQuest software (Becton Dickson). Forward and side scatter cell characteristics were used to distinguish events within the lymphocyte gate from 150 000 total events. Data expressed as counts per 100 000 events were obtained from the lymphocyte gate. Software was used to determine CD34 and KDR isotype controls below background threshold of 0.3%. EPCs that coexpressed CD34 and other receptors are referred to as CD34⁺KDR⁺, CD34⁺CD133⁺, and CD34⁺CD133⁺KDR.

Statistical Analysis

Baseline characteristics among participants with and without CMD were examined with a 2‐tailed Student t‐test when variables were normally distributed (expressed as mean±SD) or with a Mann‐Whitney U test when variable distribution was skewed (expressed as median [interquartile range]). A Shapiro‐Wilk test assessed population distribution characteristics. Categorical variables were examined using a χ² test. Univariable logistic regressions were performed to test possible associations between all studied variables and CMD. Statistically significant variables, age, and sex were included in a multivariable logistic regression model to test a potential association between klotho and CMD. A second logistic regression model additionally included known traditional cardiovascular risk factors, including hypertension, diabetes, and low‐density lipoprotein cholesterol. All multivariable logistic models were adjusted for age and sex. A subgroup analysis of participants with an ASCVD 10‐year (or lifetime risk for appropriate patients) <5% was performed to assess the relationship between CMD and klotho in the same manner. To investigate patient characteristics that may be associated with higher or lower klotho levels, continuous klotho data were dichotomized and grouped according to (1) klotho levels greater than or equal to the cohort plasma klotho concentration median or (2) klotho levels less than the cohort median. Associations between higher klotho levels and EPCs were performed using EPC data to reflect a normal distribution of progenitor cells. Two‐tailed P<0.05 was considered statistically significant. Statistical analyses were performed using STATA (StataCorp, 2019, Stata Statistical Software: Release 16, College Station, TX: StataCorp LLC) and illustrated by GraphPad Prism, version 9.5.0 for Windows (GraphPad Software, San Diego, CA, www.graphpad.com).

RESULTS

Baseline Characteristics

Table 1 lists baseline demographic and clinical characteristics among case participants with CMD (n=62) and references without CMD (n=36). Additional characteristics are listed in Table S1.

Table 1.

Baseline Characteristics Among Patients With and Without CMD

Participant characteristics	No CMD (n=36)	CMD (n=62)	P value
Age, y	51±11	55±10	0.074
Female sex, n (%)	29 (81)	45 (73)	0.376
BMI, kg/m2	29 (24 to 33)	29 (25 to 35)	0.261
Diabetes, n (%)	1 (3)	5 (8)	0.293
Hypertension, n (%)	21 (58)	32 (52)	0.520
Hyperlipidemia, n (%)	21 (58)	40 (65)	0.543
Tobacco use, n (%)			0.527
Never	20 (60)	38 (66)
Previous	11 (33)	19 (33)
Current	2 (6)	1 (2)
ASCVD risk estimator, %	1.17 (0.67 to 2.13)	1.96 (0.97 to 4.59)	0.451
Low risk (<5%), n (%)	24 (67)	34 (55)
Borderline risk (5%–7.4%), n (%)	1 (3)	3 (5)
Intermediate risk (7.5%–19.9%), n (%)	3 (8)	4 (7)
High risk (≥20%), n (%)	8 (22)	21 (34)
Glycosylated hemoglobin, %	5.3 (5.1 to 5.5)	5.2 (5.5 to 5.7)	0.142
eGFR, mL/min per 1.73 m2	89±16	83±17	0.086
Creatinine, mg/dL	0.82±0.18	0.87±0.19	0.153
BUN, mg/dL	14 (12 to 16)	16 (12 to 18)	0.020
ΔCAD, %	–4 (−9 to 4)	−16 (−32 to −4)	< 0.001
ΔCBF, %	121 (86 to 196)	7 (−29 to 25)	< 0.001
CFR	3 (2.7 to 3.4)	2.7 (2.54 to 3.21)	0.008
Klotho, ng/mL	1.75±2.38	0.88±1.50	0.034
FGF‐23 (C‐terminus), RU/mL	71±77	65±124	0.801
CD34+, cell counts/100 000	740 (320 to 1020)	835 (535 to 1145)	0.247
CD34+KDR+, cell counts/100 000	10 (0 to 30)	10 (0 to 40)	0.364
CD34+CD133+, cell counts/100 000	115 (70 to 220)	135 (60 to 240)	0.601
CD34+CD133+KDR+, cell counts/100 000	3.95 (0 to 10.70)	8.10 (1.16 to 14.48)	0.172

Data are given as mean±SD or median (interquartile range) unless otherwise indicated. ASCVD indicates atherosclerotic cardiovascular disease; BMI, body mass index; BUN, blood urea nitrogen; ΔCAD, change in coronary artery diameter; ΔCBF, change in coronary blood flow; CFR, coronary flow reserve; CMD, coronary microvascular dysfunction; eGFR, estimated glomerular filtration rate; and Klotho, α‐Klotho.

Most participants with CMD had endothelium‐dependent microvascular dysfunction (n=58), and 4 had predominant endothelium‐independent microvascular dysfunction. Participants with CMD were aged 55±10 years and 73% women (versus references aged 51±11 years and 81% women; P=0.074 and P=0.376, respectively). Clinical risk factors for cardiovascular disease, including body mass index, hypertension, diabetes, hyperlipidemia, and tobacco use history, were similar between both case and reference groups. Neither participants among case nor reference groups had significant renal disease; however, participants with CMD had statistically significant (but not clinically relevant) higher circulating blood urea nitrogen (BUN) (16 [12–18] versus 14 [12–16] mg/dL; P=0.020), with both group medians within the laboratory normal reference range (6–24 mg/dL). The mean estimated glomerular filtration rate among those with CMD was 0.87±17 mL/min, and it was 89±16 mL/min among those without CMD. There was a trend toward higher high‐sensitivity CRP (C‐reactive protein) (2.5 [1.0–4.5] versus 1.2 [0.5–3] mg/L) in the group with CMD. There were no differences in the reported use of pharmacologic agents, including angiotensin receptor inhibitors, lipid‐lowering therapy, and hypoglycemic medications on CMD. Additionally, there were no significant correlations between Gensini score and log levels of circulating EPCs, including those that were CD34⁺ (β‐coefficient, 0.770 [95% CI, –17 to 27]), CD34⁺KDR⁺ (β‐coefficient, 1.17 [95% CI, –0.41 to 2.75]; P=0.143), CD34⁺133⁺ KDR⁺ (β‐coefficient, 0.26 [95% CI, –0.38 to 0.91]; P=0.418), and CD34⁺133⁺ (β‐coefficient, 1.87 [95% CI, –6.10 to 9.83]; P=0.642).

The median circulating klotho level of the entire cohort was 0.28 (0.14–1.16) ng/mL. Circulating arterial plasma FGF‐23 levels were not significantly different between those with and without cEVA (65±124 versus 71±77 pg/mL; P=0.801). Phosphorous levels were similar between both case and reference groups (3.4 [3.1–3.6] versus 3.4 [3.2–3.8] mg/dL, respectively; P=0.522). The predicted ASCVD 10‐year risk (or lifetime risk when appropriate) for adverse cardiovascular event was similar between participants with and without CMD (median, 1.17% versus 1.96% risk for cardiovascular event over 10 years, respectively; P=0.069). Most participants in both groups were categorized as “low risk” (55% in CMD versus 67% without for ASCVD risk <5%). There were no differences in the reported use of pharmacologic agents, including angiotensin receptor inhibitors, lipid‐lowering therapy, and hypoglycemic medications on circulating klotho levels above and below the cohort mean. There was no association between klotho levels and Gensini score in a linear regression model (B‐coefficient, – 0.027 [95% CI, –0.11 to 0.06]; P=0.531).

EPCs expressing markers associated with cardiovascular protection (ie, CD34⁺, CD34⁺KDR⁺, CD34⁺CD133⁺, and CD34⁺CD133⁺KDR⁺) were not significantly different between groups. Participants with CMD had significantly less circulating klotho (0.88±1.50 versus 1.75±2.38 ng/mL; P=0.034).

Participants with CMD had concomitant epicardial endothelium‐dependent dysfunction in response to acetylcholine (−16 [−32 to −4] versus −4 [−9 to 4]; P<0.001) (Figure 1). The CFR in CMD was 2.7 (2.4–3.1), which is considered above the threshold used to suggest potential underlying endothelium‐independent microvascular dysfunction (>2.5), but was significantly lower than the CFR among reference (3 [2.7–3.4]; P=0.008).

Figure 1. Association between the functional characteristics of early coronary vascular aging and klotho (α‐klotho).

Early coronary vascular aging was characterized by coronary microvascular dysfunction (CMD). CMD was defined by percentage change in coronary blood flow (%ΔCBF) ≤50 in response to intracoronary acetylcholine or coronary flow reserve (CFR) ≤2. %ΔCAD indicates percentage change in coronary artery diameter.

Associations Between Klotho and CMD

Univariable logistic regressions of variables listed in Table 1 were performed to examine potential associations with CMD. There were significant associations between CMD and both klotho and BUN (Table 2). There was a 20% greater likelihood for participants with CMD to have lower circulating klotho (odds ratio [OR], 0.80 [95% CI, 0.636–0.996]; P=0.047). This increased likelihood persisted as an independent association between CMD and lower circulating klotho in a multivariable logistic regression model controlled for age, sex, and BUN (OR, 0.78 [95% CI, 0.618–0.986]; P=0.038). When estimated glomerular filtration rate rather than BUN was used in the multivariable logistic regression model controlled for age and sex, the results remained unchanged and there remained a significant likelihood that participants with CMD have low klotho (OR, 0.043 [95% CI, 0.607–0.980]; P=0.034). When comorbidities known to be associated with cardiovascular disease (ie, hypertension, diabetes, and low‐density lipoprotein cholesterol) and low circulating klotho were added to the multivariable regression model, there remained nearly a 20% independent likelihood that participants with CMD would have lower klotho than those without CMD (Table 3; Table S2). When body mass index was included in a multivariable logistic regression model controlled for age, sex, BUN, hypertension, diabetes, and low‐density lipoprotein cholesterol, klotho remained independently associated with CMD (OR, 0.776 [95% CI, 0.607–0.993]; P=0.044).

Table 2.

Univariable Logistic Regression Association Between CMD and Klotho

Variable	CMD
	OR	95% CI	P value
Klotho	0.80	0.636–0.996	0.047
FGF‐23 (C‐terminus)	1.00	0.996–1.00	0.978
BUN, mg/dL	1.15	1.024–1.300	0.019

BUN indicates blood urea nitrogen; CMD, coronary microvascular dysfunction; FGF‐23, fibroblast growth factor‐23; Klotho, α‐Klotho; and OR, odds ratio.

Table 3.

Multivariable Logistic Regression Association Between CMD and Klotho

CMD
Variables	OR	95% CI	P value
Klotho controlled for age, sex, and BUN*	0.78	0.618–0.986	0.038
Klotho further controlled for traditional cardiovascular comorbidities*, †	0.78	0.615–0.996	0.046

BUN indicates blood urea nitrogen; CMD, coronary microvascular dysfunction; Klotho, α‐Klotho; and OR, odds ratio.

^*Independent association.

^† Multivariable logistic regression model including age, sex, BUN, hypertension, diabetes, and low‐density lipoprotein cholesterol.

It is known that some patients with CMD may not have traditional cardiovascular risk factors. This poses a potential challenge when evaluating affected patients. To test the association between klotho and CMD in patients with low traditional cardiovascular risk, a subgroup of participants with ASCVD estimates of <5% risk for future adverse cardiovascular event, or “low‐cardiovascular risk,” was examined (n=58). Participants with low ASCVD risk were aged 52±8 years, and all 58 were women. Within this subgroup, there were 34 with CMD and 24 references without CMD (Table S3). Participants with low‐cardiovascular risk and CMD had less klotho than those without (0.29 [0.14–0.57] versus 0.82 [0.15–2.63] ng/ml; P=0.039). There remained a 20% greater likelihood that klotho would be lower in CMD (OR, 0.80 [95% CI, 0.636–0.996]; P=0.047). This likelihood was no longer significant in a multivariable logistic regression model controlled for age, sex, and the previously examined covariate risk‐factors for cardiovascular disease (ie, hypertension, diabetes, and low‐density lipoprotein cholesterol) (Table S4).

Associations Between Klotho and EPCs

EPCs were characterized and categorized by surface markers previously associated with cardioprotection and fewer future cardiovascular events, including CD34⁺, CD34⁺KDR⁺, and CD34⁺CD133⁺KDR⁺. ¹⁶ EPCs were then compared between patients with higher versus lower circulating klotho. Cases were defined by klotho levels greater than or equal to the study cohort median (≥0.28 ng/mL), and references were defined by klotho levels below the cohort median (Table S5). Cases with higher klotho levels tended to be older (55±10 versus 52±11 years; P=0.209) and had a similar sex distribution as references with lower klotho levels (73% versus 78% women; P=0.523). Although both groups had similar occurrences of traditional cardiovascular comorbidities, cases had a slightly higher body mass index (30 [27–36] versus 27 [24–31] kg/m²; P=0.014). Cases with higher klotho had significantly lower FGF‐23 (38 [7–70] versus 54 [43–101] pg/ml; P=0.009). Plasma potassium and phosphorous levels were similar between both groups.

Higher klotho levels were associated with a greater proportion of circulating EPCs expressing CD34⁺ (850 [650–1160] versus 610 [340–1070]; P=0.022), CD34⁺KDR⁺ (10 [10–30] versus 0 [0–30]; P=0.035), and CD34⁺CD133⁺KDR⁺ EPCs (8 [4–12] versus 1 [0–14]; P=0.031) (Figure 2; Table S5). There was a greater likelihood that a participant with higher klotho levels would have more EPCs expressing CD34⁺ (OR, 2.47 [95% CI, 1.26–4.87]; P=0.004), CD34⁺KDR⁺ (OR, 1.36 [95% CI, 1.06–1.74]; P=0.011), and CD34⁺CD133⁺KDR⁺ (OR, 1.46 [95% CI, 1.06–2.07]; P=0.018). These increased likelihoods maintained independent significance in a multivariable model controlled for age and sex (Table 4).

Figure 2. Association between klotho (α‐klotho) and endothelial progenitor cells.

Klotho groups were stratified by klotho levels greater than (or equal to) or below the cohort klotho level median (0.28 ng/mL). The figure illustrates the comparison between endothelial progenitor cell counts per 100 000 and klotho levels as a categorical variable. P values were derived from a Mann–Whitney U test. P<0.05 was considered statistically significant.

Table 4.

Association Between Higher Circulating Klotho and EPCs

Association between higher circulating klotho levels* and EPCs
	OR	95% CI	P value
Univariable
CD34+	2.47	1.26–4.87	0.009
CD34+KDR+	1.36	1.06–1.74	0.014
CD34+CD133+KDR+	1.48	1.06–2.07	0.021
Multivariable†
CD34+ ‡	2.44	1.26–4.73	0.008
CD34+KDR+ ‡	1.34	1.04–1.72	0.022
CD34+CD133+KDR+ ‡	1.46	1.04–2.07	0.028

EPC indicates endothelial progenitor cell; Klotho, α‐Klotho; and OR, odds ratio.

^*Higher circulating klotho levels defined by klotho levels greater than or equal to the cohort median.

^† Multivariable regression adjusted for age and sex.

^‡ Independent association.

DISCUSSION

Summary of Findings

The current study suggested that patients with CMD were more likely to have lower klotho than those without CMD. Among the subgroup of patients to have low traditional cardiovascular risk (ie, ASCVD risk estimate <5%), CMD continued to be associated with lower klotho levels. Finally, lower klotho levels were associated with a lower proportion of circulating vasculoprotective EPCs. These findings provide novel insight into correlation between vascular health and klotho, and the potential for klotho as a risk marker for CMD.

Klotho in Cardiovascular Disease

Klotho deficiency was first linked to cardiovascular disease in an experiment observing Klotho knockout mice that developed an accelerated phenotype of human aging, including skin atrophy, osteoporosis, arteriosclerosis, endothelial dysfunction, aortic valve calcification, and premature death. Interestingly, these pathophysiological characteristics associated with accelerated aging were alleviated in vivo with Klotho delivery. ¹⁷ Although normal physiological aging in humans is associated with an expected decline in circulating klotho over time, lower levels of circulating klotho have been observed in disease states associated with vascular aging, such as coronary atherosclerosis, hypertension, diabetes, and end‐stage renal disease. ¹¹ , ¹² Early studies closely associated the levels of circulating klotho to the renal cofactor FGF‐23 and propose klotho levels may be dependent on FGF‐23. Importantly, more recent evidence supports potential FGF‐23–independent functions of klotho. ¹⁸ The potential protective mechanisms of klotho against pathophysiological characteristics of aging have been attributed to optimization of EPC cardioprotective populations ¹⁹ and attenuation of vascular stressors. ²⁰ Furthermore, low klotho has recently been linked to increased senescent cell burden. Senolytic agents, which selectively eliminate senescent cells, ²¹ caused urinary klotho to increase in subjects with idiopathic pulmonary fibrosis, a cellular senescence‐driven disease. ²² In turn, senescent cells appear to contribute to impaired vascular contractility and vascular calcification with both aging and atherosclerosis in animal models, further indicating links among klotho, aging mechanisms, and vascular dysfunction. ²²

Patients With CMD Have Less Klotho

The current study adds biomarker evidence to support CMD as a feature of coronary vascular aging. Our novel findings associating CMD with lower circulating klotho are further supported by clinical studies in patients with advanced coronary artery disease. Navarro‐González et al observed an inverse correlation between soluble klotho levels and degree of obstructive CAD that increased with CAD severity. ¹¹ The present study suggested that the low klotho phenotype is present early in the pathophysiology of CAD and precedes plaque development. Moreover, this study is the first to observe lower circulating klotho in humans with endothelium‐dependent coronary microvascular disease, extending the observations from in vivo studies of murine models showing an association between klotho deficiency and impaired endothelial‐dependent vasodilation and accelerated coronary atherosclerosis. ¹⁷ Interestingly, impaired endothelial‐dependent vasodilation in klotho‐deficient mouse models was alleviated by Klotho gene delivery. ²³ Our findings support that measurement of klotho may be a noninvasive technique to enhance risk assessment in patients with CMD and suggest klotho may have therapeutic implications for CAD in humans. In light of this possibility, it remains possible that circulating klotho levels may be a downstream consequence of CMD or noCAD. The direct mechanisms that link lower klotho levels to CMD in humans require further investigation.

Although the direct mechanistic link between klotho and coronary artery health in humans remains speculative, several important observations should be noted. An in vitro study by Richter et al confirmed the presence of membrane‐bound klotho in human coronary artery endothelial cells and suggested that the presence of klotho was permissive and necessary for a balanced coronary endothelial response to reactive oxygen species by increasing NO release and accelerating reactive oxygen species degradation. ²⁴ More recent human in vitro studies have mapped possible molecular pathways associated with the antioxidative, antiapoptotic, and anti‐inflammatory effects of klotho on vascular endothelial cells ²⁵ and suggested human EPCs appropriately produced angiogenic factors when klotho was present, but changed their receptor profile to become senescent when klotho was absent. ²⁶ These mechanistic observations support our findings that associate the presence of higher klotho with vasculoprotective EPCs.

Positive Association Between Klotho and Vasculoprotective EPCs

The findings of this study support a positive association between higher circulating klotho and EPCs that express cell markers linked to vascular repair and cardioprotective potential. EPCs are cells of presumed bone marrow origin that play an important role in the regenerative potential of endothelium. Although EPCs have also been identified within the vascular wall itself and the direct mechanism by which EPCs participate in human cardiovascular disease continues to be explored, our previous findings and those by others have shown that patients with higher levels of EPCs expressing the markers CD34⁺ and KDR⁺ are at lower risk for cardiovascular events at long‐term follow‐up. ²⁷ Indeed, progenitor cells are subject to changes with aging, as all other in vivo cells in humans. Both the number and function of EPCs decline with age, and evidence supports a correlation between declining EPCs and vascular aging. ²⁸ The potential role of EPCs in mitigating the effects of aging was further suggested by a study that associated a Klotho gene variant in humans with greater circulating CD34⁺ cells, longevity, and decreased likelihood of cardiovascular disease. ¹⁹ The direct mechanistic link between EPCs and klotho was explored by a recent in vitro study that proposed a critical role of klotho in human EPC cell surface marker expression. Human EPCs were influenced by FGF‐23 at a klotho‐independent receptor (fibroblast growth factor receptor 1) to downregulate the expression of cell homing receptors C chemokine receptor type 4 (CXCR4) and influence the production of injury‐response signals (vascular endothelial growth factor‐A and interleukin‐6). Interestingly, EPCs were unable to downregulate homing receptors or produce injury response signals when klotho transcription was silenced. ²⁹ Although evidence continues to link the EPC response to vascular injury and klotho, further studies are needed to understand the complex role that klotho may directly or indirectly play. The current study adds clinical insight to prior observations and supports an association between klotho and the maintenance of a satisfactory endothelial progenitor population in patients without cardiovascular disease.

Role of Klotho in Risk Assessment

The current study enhances our understanding of risk features associated with CMD in early CAD and may complement how we risk stratify a population of patients with CMD who are younger, predominantly women, and overall lack traditional cardiovascular risk factors. Participants with low ASCVD risk and CMD had lower klotho levels than those without. Interestingly, all participants with low ASCVD risk were women. It is known that female patients with chest pain and nonobstructive coronary artery disease have a high prevalence of CMD that does not correlate with conventional cardiovascular risk factors. ⁴ Thus, emerging biomarkers, such as circulating klotho, may highlight potential cardiovascular risk in select populations of patients without traditional cardiovascular comorbidities.

Through the National Institutes of Health Translational Geroscience Network (R33 AG61456), its Facility for Geroscience Analysis is currently assaying klotho across multiple clinical trials of interventions targeting fundamental aging processes (including senolytics, metformin, rapamycin, nicotinamide riboside (NAD) precursors, anti‐inflammatories, and others) for several indications. This provides an opportunity to test links among klotho, other biomarkers of aging processes (“gerodiagnostics”), cardiovascular function, and CAD events across clinical trials and diverse populations. ³⁰ Through the Facility for Geroscience Analysis, efforts are underway to develop composite scores of gerodiagnostic markers to track fundamental aging processes, links to disorders, such as cardiovascular dysfunction, and responses to interventions, such as senolytics. The first such composite score of blood gerodiagnostics was reported in 2019, and many others are under development. The current study supports considering adding klotho to such composite scores. ³¹

Limitations

The current study draws conclusions from a rigorous retrospective analysis. Patient selection was conducted by examining all eligible participants from a retrospective database rather than from matched case and reference groups within a larger population. Although clinical investigations of unique, relatively homogeneous groups of patients may strengthen group‐specific findings and limit potential clinical factors that may confound results, future studies in larger, diverse groups enrolled in prospective clinical trials are ultimately needed to further investigate associations between CMD and klotho in diverse patient groups. Additional large prospective studies are needed to better understand the role of klotho as a biomarker to stratify risk patients with chest pain and early features of coronary disease. The direct mechanisms that link klotho and humans require further study to better understand whether klotho levels are linked to the pathophysiology of CMD or may instead represent a downstream consequence. Klotho is known to exist as a transmembrane protein within tissues, as a cleaved circulating protein, or as a secreted protein, with each derivative potentially possessing independent functions. ¹⁰ The presented findings focused on circulating klotho; further studies are needed to assess the relationship between tissue klotho versus circulating klotho in patients with early coronary disease.

CONCLUSIONS

In summary, we present novel findings to support an association between CMD and lower circulating klotho in a population of patients with CMD and noCAD who may be considered to have low cardiovascular risk by traditional measures. Patients with more circulating klotho were significantly less likely to have underlying CMD and more likely to have EPCs associated with vascular repair potential and cardioprotection. Klotho may be a marker for early coronary artery disease in patients who may be considered at low‐cardiovascular risk by traditional measures. Further studies are needed to investigate therapeutic implications in humans.

Sources of Funding

This work was supported by the Mayo Clinic Foundation. Dr Akhiyat is supported by the National Institutes of Health (grant T32 HL007111). Drs Tchkonia and Kirkland are supported by the National Institutes of Health (grants R37AG013925, R33AG061456, R01AG68048, R01AG 64 165, and P01AG062413); the Connor Fund; the Robert J. and Theresa W. Ryan Fund; and the Noaber Foundation.

Disclosures

Drs Tchkonia and Kirkland have a financial interest related to this research, including patents and pending patents covering senolytic drugs and their uses that are held by Mayo Clinic. This research has been reviewed by the Mayo Clinic Conflict of Interest Review Board and was conducted in compliance with Mayo Clinic conflict of interest policies. The remaining authors have no disclosures to report.

Supporting information

Tables S1–S5