Association between serum levels of Klotho and inflammatory cytokines in cardiovascular disease: a case-control study

Introduction

Klotho (KL) is a longevity-related gene discovered in 1997 [1]. Defects in KL expression leads to a syndrome resembling human aging that includes endothelial dysfunction, vascular calcification and progressive atherosclerosis [1]. KL gene encodes a single-pass transmembrane glycoprotein type 1, mainly expressed in renal tubular cells, and a soluble form derived from the action of membrane proteases that cleave the extracellular domain of the membrane-bound protein [2, 3].

Aging is associated with a reduction in the renal expression and serum concentrations of Klotho, which is also observed in diseases characterized by premature vascular aging, such as renal failure, hypertension, and diabetes mellitus [4–6]. Moreover, several clinical studies have observed that lower systemic levels of the soluble form of Klotho are associated with markers of vascular dysfunction [7–10] as well as with the prevalence and severity of cardiovascular disease (CVD) [11–13]. These associations are particularly relevant in renal patients since the largest proportion of systemic Klotho is generated by the kidneys and its levels are reduced along with the renal function decline [14].

Chronic inflammation is an important factor in the development of CVD. Increased serum levels of different acute phase proteins and inflammatory cytokines, as well as infiltration of the vascular tissue by immune cells, are characteristics of a low-grade inflammation scenario which may act as substrate for the development and progression of atherogenic damage. Different inflammatory cytokines, including tumor necrosis factor alpha (TNFα) and interleukin (IL) 10, are involved in the onset and development of the vascular lesion [15–16]. Interestingly, inflammatory cytokines can down-regulate Klotho levels [17–18] whereas, in the opposite direction, Klotho is able to modulate inflammation by inhibiting central signaling pathways and the expression of inflammatory related molecules [19–21]. This raises the intriguing question about the potential interrelationships between Klotho and systemic inflammation in the context of human CVD, an aspect that has not been previously addressed. Therefore, the aim of this work is to analyze, in a group of patients with established CVD and preserved renal function, the serum levels of soluble Klotho and its association with the concentration of inflammatory cytokines related to atherosclerotic vascular damage.

Results

From the 181 potential subjects who were initially considered in the CVD group, 71 were excluded according to exclusion criteria and 110 patients were finally included (Supplementary Figure 1). The demographic, clinical and biochemical data of the entire study population, included the control group, are presented in Table 1. The clinical diagnosis consisted of peripheral artery disease (PAD) (66.4%), transient ischemic attack (TIA) (39.1%), and abdominal aortic aneurysm (AAA) (18.2%). The CVD group was significantly older (P<0.01), with a higher prevalence in hypertension (HT) (P<0.05) but not in diabetes mellitus (DM). This group presented a lower estimated glomerular filtration rate (eGFR) (P<0.05), as well as lower serum concentrations of calcium (P<0.001). The use of antiaggregants (P<0.0001), angiotensin-converting enzyme inhibitors or angiotensin receptor antagonist (P<0.05) and statins (P<0.0001) was significantly higher in the CVD-group.

Serum levels of soluble Klotho and inflammatory cytokines in CVD.

Soluble Klotho concentrations were significantly lower in patients with established CVD (P<0.001) (Table 1 and Figure 1A), without differences according to each specific form of clinical atherosclerotic vascular disease (P=0.47). Regarding the inflammatory cytokines, there were no significant differences concerning serum TNFα levels (P=0.39). However, the serum concentrations of IL10 were significantly higher in the control group (P<0.01). In order to better evaluate the inflammatory status of the subjects, understood as the balance between pro- and anti-inflammatory forces, we calculated the ratios between the serum levels of TNFα, a pro-inflammatory factor, to IL10 concentrations, a cytokine with anti-inflammatory effects. A significantly higher TNFα/IL10 ratio was observed in the CVD group compared with controls (P<0.01) (Figure 1A). No differences in serum levels of these inflammatory cytokines or in the TNFα/IL10 ratio were observed according to the forms of clinical vascular disease. Inflammatory status was also evaluated by the serum concentration of C-reactive protein (CRP) and the neutrophil-to-lymphocyte ratio (NLR). Both parameters were higher in the CVD group, although only the differences for CRP reached statistical significance (P<0.0001), while the mean NLR showed a trend (P=0.19) (Table 1). Most of the patients were receiving antiaggregants, renin-angiotensin system blockers or statins. When evaluating the serum levels of Klotho or the inflammatory cytokines studied between these subjects who received treatment and those who did not, we found no significant differences.

Figure 1. Soluble Klotho and inflammatory cytokines serum levels in CVD and non-CVD groups. (A) Full-study groups, (B) age-matched subgroups. *P<0.05, ** P<0.01, ***P<0.001.

The characteristics of patients in the CVD group stratified by tertiles of soluble Klotho are shown in Table 2. Subjects within the highest tertile of soluble Klotho showed lower serum concentrations of TNFα (P<0.05) and tended to present decreased levels of TNFα/IL10 ratio (P=0.09).

Table 2. Clinical characteristics and biochemical assessments of patients with established CVD stratified by tertiles of serum Klotho levels.

Variable	Tertile 1(<391.1 pg/mL) (n=37)	Tertile 2(391.1-591.8 pg/mL) (n=37)	Tertile 3(>591.8 pg/mL) (n=36)	P-value
Characteristics
Age (years)	67.7±9.3	68.3±8.6	64.2±8.6	NS
Sex (M/F)	31 / 6	27 / 10	30 / 6	NS
Smoking (%)	81.1	75.7	73	NS
Alcoholism (%)	51.4	37.8	51.4	NS
BMI (kg/m2)	29.0±4.1	27.8±4.1	27.4±4.6	NS
Co-morbidities
HT (%)	81.1	81.1	75.0	NS
DM (%)	54.1	54.0	36.1	NS
Pharmacological treatment
Antiaggregants (%)	83.8	86.5	88.9	NS
Beta-blockers (%)	32.4	27.1	25.0	NS
ACEI/ARA2 (%)	51.4	59.5	44.4	NS
CCB (%)	16.2	32.4	19.4	NS
Statins (%)	89.2	89.2	75.0	NS
Laboratory data
eGFR (mL/min/1.73m2)	88.7±11.2	87.4±13.4	93.0±13.4	0.13
Creatinine (mg/dL)	0.84±0.2	0.84±0.2	0.81±0.2	NS
Albumin (g/dL)	4.0±0.6	3.8±0.6	3.9±0.6	NS
Calcium (mg/dL)	9.1±0.5	9.2±0.5	9.2±0.4	NS
Phosphorus (mg/dL)	3.5±0.5	3.5±0.4	3.6±0.5	NS
Uric acid (mg/dL)	5.87±1.6	6.2±1.3	5.4±1.3	NS
Glucose (mg/dL)	110.4±26.1	113.5±32.5	111.2±28.9	NS
Total cholesterol (mg/dL)	157.6±48.1	169.4±40.9	171.8±54.9	NS
HDL (mg/dL)	42.3±12.8	43.9±10.4	44.1±11.8	NS
NLR	3.7±2.1	3.8±2.2	3.6±1.9	NS
CRP (mg/dL)	4.1±2.9	3.5±2.8	3.2±2.6	NS
TNFα (pg/mL)	1.13 (0.88-1.65)	1.02 (0.82-1.35)	1.02 (0.69-1.39)	<0.05
IL10 (pg/mL)	1.05 (0.51-5.81)	2.17 (0.55-4.9)	4.03 (0.58-5.19)	NS
TNFα/IL10	0.73 (0.26-1.81)	0.63 (0.19-1.77)	0.28 (0.15-1.99)	0.09
BMI, body mass index; HT, hypertension; DM, diabetes mellitus; ACEI/ARA2, angiotensin converting enzyme inhibitor/angiotensin receptor antagonist 2; CCB, calcium channels blockers; eGFR, estimated glomerular filtration rate; HDL, high-density lipoprotein; NLR, neutrophil-to-lymphocyte ratio; CRP, C-reactive protein; TNFα, tumor necrosis factor alpha; IL10, interleukin 10; NS, not significant.

To avoid the confounding effect of age on serum Klotho levels, we performed a subanalysis in our cohort by age-matching the study groups (Supplementary Figure 1). The clinical and laboratory data of these subgroups are presented in Supplementary Table 1. Again, serum levels of soluble Klotho were significantly lower in the group of participants with established CVD (P<0.05) (Figure 1B). There were no differences in serum Klotho concentration among patients according to the presence of PAD or TIA. The serum levels of TNFα did not show significant differences between both groups (P=0.31), whereas in the non-CVD group the serum IL10 concentrations were significantly higher (P<0.05) and the values of the TNFα/IL10 ratio were significantly lower (P<0.05) (Figure 1B). No differences in serum levels of these inflammatory cytokines were observed among the patients according to the presence of PAD or TIA. Serum levels of CRP and the NLR showed higher values in the CVD group, differences that almost reached statistical significance (P=0.08 and P=0.10, respectively).

Discussion

In this study, we observed that the decrease in soluble Klotho levels that occurs in CVD (defined as PAD, TIA or AAA) is associated with the systemic inflammatory environment that accompanies this disorder. We determined the serum levels of TNFα and IL10, inflammatory cytokines involved in the onset and development of vascular disease, in addition to other inflammatory parameters such as CRP or NLR. In particular, we observed significantly higher values of CRP and TNFα/IL10 ratio in patients with established CVD, which also presented significantly lower levels of soluble Klotho. The difference between groups for the TNFα/IL10 ratio was due to a reduction in the IL10 serum levels rather than an increase in the TNFα serum concentration. Furthermore, we observed that soluble Klotho levels are directly associated to IL10, while there is an inverse correlation with CRP and TNFα/IL10 ratio. Interestingly, when we matched both groups by age, a factor associated with soluble Klotho levels [22], the correlations between Klotho and the levels of IL10 and the TNFα/IL10 ratio still remained significant. The patients with higher values of serum Klotho showed a trend to have decreased values of the TNFα/IL10 ratio. Moreover, we observed that while age and NLR were independent risk factors for the presence of CVD, higher serum Klotho levels were associated with a lower risk.

The results of this work agree with previous clinical studies that have observed a reduction in soluble Klotho levels in patients with CVD. In 2011, Semba et al. showed in a cohort of 1023 community-dwelling adults that higher plasma Klotho levels were independently associated with a lower likelihood of having CVD (understood as coronary artery disease (CAD), heart failure, stroke or PAD and, similarly to our study, they observed an inverse association between circulating Klotho and CRP levels (r=-0.10, P<0.001) [11]. Besides, we previously reported that patients with significant CAD present lower soluble Klotho levels and, also, that soluble Klotho concentration is independently and negatively associated with the severity of the coronary stenosis [12]. Interestingly, several studies have pointed out a relationship between the decrease in serum soluble Klotho and worse outcomes in vascular functional tests, like reduced capacity of flow-mediated dilation of the brachial artery, or higher values of epicardial fat thickness and carotid artery intima-media thickness (cIMT) in healthy adults [8], or increased ankle-brachial pulse wave velocity, left ventricular mass index, cIMT and carotid atherosclerotic plaque quantity in chronic kidney disease population [5–7]. Furthermore, Pan et al. recently showed, in a prospective study of 252 patients and a 7-year follow-up, that lower baseline levels of circulating Klotho were associated with an increased risk of developing CAD or cerebrovascular accident (CVA) in the long term in diabetic population [13]. Despite the multiple evidences that point to CVD as a state of soluble Klotho deficiency, a definite cause for this decrease is not yet known. A plausible explanation is a decrease in the renal production either by a repression of the KL gene or by a reduction of the expression/activity of the metalloproteases responsible for generating the soluble form. Unfortunately, there are no studies that address these issues. A factor closely related to Klotho expression is renal function, so it is known that in CKD it occurs a reduction in renal and systemic Klotho levels (14). In fact, there are several works that relate a lower production of soluble Klotho in renal patients with a worse vascular function [5–7]. However, in our study we show how this decrease in soluble Klotho during CVD also occurs in individuals with a preserved renal function (all of them with eGFR>60 mL/min/1.73m²; mean eGFR 89.8 ±13mL/min/1.73m²), so there must be others factors beyond the renal status that affect serum Klotho levels.

Inflammation, that usually accompanies CVD, may has some role in the modulation of systemic Klotho. Although all previous studies have provided new insights into the role of soluble Klotho in the development of CVD, none of them have inquired in the potential interactions between Klotho and inflammatory cytokines. We have addressed this question for the first time in this study. The expression of Klotho can be negatively modulated by inflammatory cytokines, such as TNFα or tumor necrosis factor-like weak inducer of apoptosis (TWEAK) [17, 18]. Though in our study we did not observe differences in serum levels of TNFα or an association between soluble Klotho and this pro-inflammatory cytokine, we found an inverse correlation with CRP, another pro-inflammatory marker. However, it is interesting to highlight the potential anti-inflammatory effects of Klotho [19–21]. In our results, we observed a direct relationship between serum Klotho levels and IL10 concentrations, both decreased in patients with CVD. This association could suggest that, partially, the effects on vascular homeostasis in which Klotho participates could be through modulation of anti-inflammatory cytokine levels. Interestingly, we found that this relationship between Klotho and IL10 remained after eliminating the effect of age in our study groups, which is a factor influencing Klotho levels. This might suggest that the role of Klotho in CVD and its relationship with inflammation is independent of age.

Limitations in our study include that, due to logistical reasons, we did not measure levels of other molecules implicated in the inflammatory response with known roles in the framework of CVD, namely IL1, IL6 or transforming growth factor (TGF) β, and therefore we cannot have a broad view of Klotho interactions with different inflammatory mediators. Likewise, we do not have either measurement of molecules of mineral metabolism related to Klotho and with potential impact on CVD, such as fibroblast growth factor 23 or vitamin D, so we cannot rule out their interactions in the Klotho-inflammation relationship. Moreover, a considerable proportion of patients received treatment with statins, a drug with known anti-inflammatory properties, that may interfere with the outcomes of our work. Although we did not observe significant differences for the parameters studied between individuals who received the treatment and those who did not, the possibility of its influence should not be completely ruled out. A further limitation is that we employed a relatively small sample size that could make that some of our results did not reach statistical significance and, consequently, the ability to describe the association between some variables might be limited. Also, this limitation could make that the findings made in these groups not transferable to larger populations. Finally, due to the cross-sectional design of the study we can detect associations but not to infer causality.

Conclusion

Soluble Klotho is decreased in CVD and its decline is associated to the pro-inflammatory background set during disease. Our results are in line with previous findings in the literature that point to soluble Klotho as a protective factor against cardiovascular disorders. Further works that deepen into the interactions between Klotho and systemic inflammatory mediators, with particular interest in IL10, are needed and may offer new strategies for the treatment of vascular pathologies.

Materials and Methods

Supplementary Materials

Supplementary Figure 1

Supplementary Table 1