Arterial and Cellular Inflammation in Patients with CKD

Abstract

CKD associates with a 1.5- to 3.5-fold increased risk for cardiovascular disease. Both diseases are characterized by increased inflammation, and in patients with CKD, elevated C-reactive protein level predicts cardiovascular risk. In addition to systemic inflammation, local arterial inflammation, driven by monocyte-derived macrophages, predicts future cardiovascular events in the general population. We hypothesized that subjects with CKD have increased arterial and cellular inflammation, reflected by ¹⁸F-fluorodeoxyglucose (¹⁸F-FDG) positron emission tomography computed tomography (PET/CT) of the arterial wall and a migratory phenotype of monocytes. We assessed ¹⁸F-FDG uptake in the arterial wall in 14 patients with CKD (mean±SD age: 59±5 years, mean±SD eGFR: 37±12 ml/min per 1.73 m²) but without cardiovascular diseases, diabetes, or inflammatory conditions and in 14 control subjects (mean age: 60±11 years, mean eGFR: 86±16 ml/min per 1.73 m²). Compared with controls, patients with CKD showed increased arterial inflammation, quantified as target-to-background ratio (TBR) in the aorta (TBR_max: CKD, 3.14±0.70 versus control, 2.12±0.27; P=0.001) and the carotid arteries (TBR_max: CKD, 2.45±0.65 versus control, 1.66±0.27; P<0.001). Characterization of circulating monocytes using flow cytometry revealed increased chemokine receptor expression and enhanced transendothelial migration capacity in patients with CKD compared with controls. In conclusion, this increased arterial wall inflammation, observed in patients with CKD but without overt atherosclerotic disease and with few traditional risk factors, may contribute to the increased cardiovascular risk associated with CKD. The concomitant elevation of monocyte activity may provide novel therapeutic targets for attenuating this inflammation and thereby preventing CKD-associated cardiovascular disease.

CKD is a major risk factor for developing cardiovascular (CV) events,¹ with an overall estimated 20%–30% increase in CV risk for each 30% reduction in eGFR.² In fact, subjects with early stages of CKD are more likely to die from heart disease than to actually ever reach ESRD.³ As CKD is highly prevalent (estimated 8%–16% prevalence worldwide),⁴ these numbers underscore the importance of adequate CV prevention in patients with CKD.

Traditional cardiovascular disease (CVD) risk factors underestimate CVD risk by as much as 50% in the CKD population,⁵ suggesting other disease-causing mechanisms in CKD-associated CVD. Over the past several decades, inflammation has gained increasing interest as a contributing factor to the increased CVD risk in patients with CKD. Serum biomarkers of inflammation, such as C-reactive protein (CRP) and IL-6, are associated with increased CVD risk, not only in the general population, but also in patients with CKD.^6,7 Furthermore increased CRP levels were known to correlate with eGFR decline.⁸ The prevailing mechanisms behind this relationship are still unclear. In addition to systemic inflammation, local inflammatory activity of the arterial wall can be assessed measuring ¹⁸F-fluorodeoxyglucose (¹⁸F-FDG) uptake with positron emission tomography computed tomography (PET/CT).^{9,10 18}F-FDG uptake, representing increased metabolic activity, has been previously linked to plaque macrophage content,¹¹ and is an independent predictor for future CV events.¹² Data on arterial inflammation in patients with CKD are not available to date.

The enhanced inflammatory activity in the arterial wall in CVD patients is thought to be primarily driven by continuous influx of circulating monocytes,¹³ eventually leading to subendothelial inflammatory macrophages. In CVD, data supports that systemic stimuli can polarize circulating monocytes toward more inflammatory phenotypes, rendering them more prone to enter the arterial wall and promote local inflammation.^14,15 Interestingly, decreased kidney function is associated with a shift toward a more adhesive phenotype in circulating monocytes,¹⁶ which was shown to predict CV events in the CKD population.¹⁷

In this study, we set out to evaluate arterial wall inflammation in patients with CKD using ¹⁸F-FDG PET/CT. We also addressed activation and function of monocytes ex vivo, using flow cytometry and an endothelial migration assay.

Results

We included 14 patients with CKD, (59±5 years, 50% men) and selected 14 age- and sex-matched controls (age 60±11 years, 50% men). Clinical characteristics are outlined in Table 1. Patients with CKD had an eGFR of 37±12 ml/min per 1.73 m² versus 86±16 ml/min per 1.73 m² (P<0.001) with median plasma creatinine levels of 183 (interquartile range [IQR], 123–197) umol/L versus 77 (IQR, 68–84) umol/L in controls (P<0.001). Creatinine clearance was 48 (IQR, 29–65) ml/min and 24-hour protein excretion was low (0.17 [IQR, 0.1–0.3] g/24 h). Plasma urea levels were elevated compared with reference ranges (11.2 [IQR, 8–17] mmol/L; reference range, 2.5–6.4).

Table 1.

Clinical characteristics of patients with CKD and healthy controls

Characteristic	Control (n=14)	CKD (n=14)	P Value
Sex, men/women	7/7	7/7	1.00
Age, yr	60.4±11	58.6±5	0.58
BMI, kg/m2	22.6±2	25.2±4	0.04
Smoking, yes/no	0/14	3/11	0.07
SBP, mmHg	131±10	135±18	0.51
DPB, mmHg	77±7	80±8	0.29
Creatinine, umol/L	77[68–84]	183 [123–197]	<0.001
eGFR (CKD-EPI), ml/min per 1.73 m2	86±16	37±12	<0.001
Total cholesterol, mmol/L	5.6±0.8	5.7±1.3	0.92
LDL cholesterol, mmol/L	3.5±0.88	3.5±1.0	0.92
HDL cholesterol, mmol/L	1.8±0.34	1.4±0.31	0.02
Triglycerides, mmol/L	0.69 [0.50–0.98]	1.34 [0.98–1.81]	0.01
Coronary calcium score	0[0–0]	0[0–1.5]	0.83
CRP, mg/dl	1.1 [0.7–1.5]	2.2 [0.7–3.8]	0.16
WBC, 10E9/L	5.6±1.5	5.3±1.6	0.61
Lymphocytes	1.9±0.5	1.5±0.5	0.93
Neutrophils	3.2±1.4	3.2±1.4	0.06
Monocytes	0.4±0.07	0.4±0.1	0.60

Values are n, mean±SD, or median [interquartile range] for skewed data. SBP, systolic BP; DBP, diastolic BP; CKD-EPI, Chronic Kidney Disease Epidemiology Collaboration; WBC, white blood cell count.

CV risk factors, including body mass index (BMI), BP, and plasma cholesterol levels, were comparable between patients and control subjects, although there were slightly more smokers in the CKD group. Plasma triglycerides were slightly increased in CKD subjects (1.34 [IQR, 0.98–1.81] mmol/L versus 0.69 [0.50–0.98] mmol/L in controls; P=0.01) and plasma HDL cholesterol was decreased (1.4±0.31 mmol/L versus 1.8±0.34 mmol/L in controls; P=0.02). Further characterization of the patients with CKD is outlined in Table 2.

Table 2.

Detailed CKD-related parameters

Characteristic	All (n=14)	Reference Rangesa
Diagnosis
ADPKD	10
Hypertensive	3
Monokidney	1
Plasma sodium, mmol/L	141±1.7	135–145
Plasma potassium, mmol/L	4.2±0.43	3.5–5.0
Plasma phosphate, mmol/L	1.0±0.11	0.9–1.5
Plasma PTH, pmol/L	7.5 [5.4–8.3]	0.6–6.7
Plasma urea, mmol/L	11.2 [8–17]	2.5–6.4
Plasma uric acid, mmol/L	0.42±0.06	Men: 0.20–0.42 Women: 0.12–0.34
Plasma albumin, g/L	43±2.4	35–55
Plasma venous HCO3−, mmol/L	24.7±2.6	21–27
Plasma calcium, mmol/L	2.5±0.11	2.10–2.55
Medication use		N/A
Statins	2
Antihypertensive	12
ACI or ARB	9
Diuretic	3
β Blocker/calcium antagonist	3

Values are n, mean±SD, or median [interquartile range] for skewed data. ADPKD, adult dominant polycystic kidney disease; PTH, parathyroid hormone; HCO3⁻, bicarbonate; N/A, not applicable; ACI, angiotensin-converting-enzyme inhibitor; ARB, Angiotensin II receptor antagonists.

^aReference ranges are derived from reference values clinical chemistry (referentiewaarden klinische chemie, www.farmacotherapeutischkompas.nl).

Increased Arterial Wall Inflammation in CKD Patients

We performed ¹⁸F-FDG PET/CT imaging to assess aortic and carotid arterial wall inflammation. ¹⁸F-FDG uptake in patients with CKD, quantified as target-to-background ratio (TBR), was increased in the arterial wall of both the aorta and the carotid arteries (aorta_max TBR: CKD, 3.14±0.70 versus control, 2.12±0.27, adjusted P=0.001; carotid_max TBR: CKD, 2.45±0.65 versus control, 1.66±0.27, adjusted P<0.001; P values adjusted for CV risk factors which differed at baseline: BMI, current smoking status, and statin use) (Figure 1, C and E). Uptake in the most diseased segment (MDS) was also elevated in both arterial beds in patients with CKD (aorta_MDS: CKD, 3.21±0.71 versus control, 2.20±0.26, adjusted P=0.001; carotid_MDS: CKD, 2.55±0.68 versus control, 0.79±0.26, adjusted P=0.001) (Figure 1, D and F). Coronary calcium scores were also derived from the computed tomography scans, but were low for both groups (CKD, 0 [0–0] versus control, 0 [0–1.5]; P=0.83), and thus showed no relation to ¹⁸F-FDG uptake.

Figure 1.

Increased arterial wall inflammation in patients with CKD. ¹⁸F-FDG uptake in the aortic arch (top) and the carotid arteries (bottom) was quantified as the TBR in controls (n=14, represented in A) and patients with CKD (n=14, represented in B). In patients with CKD uptake in the whole vessel (TBR_max) (C, E) as well as the MDS (D, F). Data are mean±SD; *P<0.05; **P<0.01. C, carotid; JV, jugular vein.

To explore which factors could potentially explain the TBR differences, we performed univariate and multivariate (with BMI, current smoking status, and statin used as covariates, none of which showed a correlation with TBR; Supplemental Tables 1 and 2) linear regression with aortic TBR_max as the outcome variable. We found eGFR and plasma urea levels significantly correlated with ¹⁸F-FDG uptake, with a concomitant trend for plasma creatinine levels and systolic BP. Plasma HDL cholesterol and triglyceride levels, which differed from controls at baseline, showed no correlation with arterial wall inflammation, and neither did LDL cholesterol or CRP levels (Table 3). None of the other baseline parameters showed correlation with TBR (Supplemental Table 3). As hypertension was found to be a potential confounder in CKD-associated arterial wall inflammation, we subsequently selected a cohort of hypertensive, non-CKD subjects (n=8) from a contemporary PET/CT study performed at our center.¹⁸ This revealed that although the selected hypertensive patients had a worse CV risk profile compared with patients with CKD, based on traditional risk assessment (men only, higher BMI, higher diastolic BP, and mean arterial pressure), aortic TBR was lower compared with CKD subjects, whereas carotid TBR was nonsignificantly increased in CKD subjects compared with hypertensive subjects (baseline and TBR values outlined in Supplemental Tables 4 and 5). Finally, we assessed the contribution of systolic BP to the arterial inflammation in CKD by adding systolic BP to the covariance analysis used to assess differences in TBR. If systolic BP was added to the adjusted model, the TBR difference between patients with CKD and healthy controls remained highly significant (P<0.001 for aorta_max; P=0.001 for aorta_MDS; P<0.001 for carotid_max; and P=0.001 for carotid_MDS). However, when adding eGFR to the model, significance was lost (P=0.25 for aorta_max; P=0.23 for aorta_MDS; P=0.17 for carotid_max; and P=0.28 for carotid_MDS).

Table 3.

Univariate and multivariate linear regression analysis with aortic TBR_max as the dependent variable

Characteristic	Univariate Analyses		Multivariate Analysesa
	β (95% CI)	P Value	β (95% CI)	P Value
Urea	0.51 (−0.01 to 0.18)	0.08	0.82 (0.01 to 0.26)	0.04
Creatinine	0.36 (−0.01 to 0.01)	0.23	0.85 (0.00 to 0.02)	0.05
eGFR (CKD-EPI)	−0.40 (−0.06 to 0.01)	0.18	−0.08 (−0.09 to 0.00)	0.05
SBP	0.61 (0.01 to 0.05)	0.04	0.86 (−0.01 to 0.07)	0.06
LDL cholesterol	−0.01 (−0.45 to 0.44)	1.00	−0.09 (−0.69 to 0.57)	0.84
HDL cholesterol	−0.10 (−1.84 to 1.35)	0.75	−0.13 (−2.25 to 1.63)	0.73
TG	−0.16 (−1.34 to 0.82)	0.61	−0.22 (−1.73 to 1.01)	0.56
CRP	−0.10 (−0.14 to 0.10)	0.74	−0.01 (−0.21 to 0.21)	0.98

Data are standardized coefficient (β) with 95% confidence intervals (95% CI). CKD-EPI, Chronic Kidney Disease Epidemiology Collaboration; SBP, systolic BP; TG, triglycerides.

^aBMI, current smoking status, and statin use were used as covariates.

Cellular Activation in Patients with CKD

Finally, we assessed phenotype and function of freshly isolated monocytes. To this end, we included 14 controls, matched for age and sex to the CKD cohort, and the clinical characteristics were comparable to the control cohort used for the PET/CT (Supplemental Table 6). Monocytes were classified according to HLA-DR, CD14, and CD16 expression¹⁹: The distribution across the classic (Mon1: CD14⁺⁺CD16⁻), intermediate (Mon2: CD14⁺⁺CD16⁺), and nonclassic (Mon3: CD14⁽⁺⁾CD16⁺) subsets did not differ between patients with CKD and controls (Figure 2A). However, CCR7 expression was increased on all monocyte subsets in CKD subjects compared with controls, with a similar pattern for CCR2 (Figure 2, B and C), whereas no significant effect was found for CCR5 expression (Figure 2D). In a transendothelial migration assay, we found a significant increase in monocytes that crossed the endothelial barrier (transmigrated cells per millimeter²: CKD, 88±27 versus control, 57±15; P=0.01).

Figure 2.

Monocytes of CKD subjects show increased expression of chemokine receptors and enhanced migratory capacity. Flow cytometry on whole blood was performed to study expression of monocyte surface markers. Monocytes were divided into classic (CD14⁺⁺CD16⁻), intermediate (CD14⁺⁺CD16⁺), or nonclassic (CD14⁽⁺⁾CD16⁺) monocytes. (A) Surface expression of monocyte CCR7, CCR2, and CCR5 represented as Δ median fluorescence intensity (MFI), in CKD subjects (n=14) versus controls (n=14). (B, C, D) Transendothelial migratory capacity was quantified as the number of transmigrated cells per millimeter². (E) For each subject, transmigrated cells are calculated of independent counts of five frames of view. Data represent mean±SEM. *P<0.05; **P<0.01; ***P<0.001.

Discussion

Here we show that patients with CKD are characterized by an increased ¹⁸F-FDG uptake of the arterial wall on PET/CT, compatible with an elevated inflammatory state and increased CVD risk. Furthermore, monocytes from patients with CKD display a proadhesive phenotype on flow cytometry, as well as enhanced migratory capacity ex vivo. In multifactorial analysis, eGFR and urea plasma levels were independent predictors of arterial wall inflammation. These data imply that CKD-associated inflammatory activation may contribute to the elevated CVD risk in CKD-patients.

¹⁸F-FDG-PET/CT imaging is a validated technique to quantify arterial wall inflammation and used as an independent predictor of CVD risk.^12,20,21 Increased ¹⁸F-FDG arterial wall uptake was found in active culprit lesions,^22,23 primarily in macrophage rich areas,²² and correlated to macrophage content¹¹ as well as inflammatory gene expression.²⁴ Hence, ¹⁸F-FDG PET/CT imaging is increasingly used to monitor therapeutic efficacy of antiatherosclerotic and anti-inflammatory strategies.^25–29 Here we show that CKD subjects without overt atherosclerotic disease (represented by the low coronary calcium scores in our study) or comorbidity, have increased inflammatory activity in the arterial wall in comparison to age- and sex-matched controls. The correlation with eGFR and plasma urea levels, and to a lesser extent with creatinine and systolic BP, suggests a ‘dose-dependent’ effect of (the consequences of) renal insufficiency on arterial wall inflammation. Loss of kidney function reflects the accumulation of uremic retention solutes: a diverse group of compounds that adversely affect different organs and cells.³⁰ These uremic toxins have been proposed to play an important role in development and progression of CKD-associated CVD through diverse mechanisms, including activation of leukocytes and enhancing monocyte-endothelial interactions.³¹ The elevated systolic BP observed in our CKD cohort can contribute to endothelial activation³² but cannot exclusively explain the arterial inflammation, evidenced by the lower levels in hypertensive subjects without CKD, thereby further augmenting this process.

Monocyte trafficking across the arterial wall is an important contributor to arterial wall inflammation.¹³ In contrast to previous data³³ we did not find differences in monocyte subsets in our study, likely due to the limited number of patients. We did observe increased monocyte CCR7 and (to an extent) CCR2 expression. CCR7 and its ligands, Chemokine (CC-motif) ligand 19 (CCL19) and 21 (CCL21) have been associated with increased monocyte migration and are associated with increased CVD risk in humans.³⁴ CCR2, the receptor for monocyte chemo-attractant protein-1 (MCP-1), also enhances chemotaxis³⁵ and is important in the process of plaque formation in mice.³⁶ In our study, monocytes of patients with CKD also showed enhanced migratory capacity compared with healthy controls, suggesting that expression of these chemokine receptors in humans may also contribute to the inflammatory status of the arterial wall. Thus, our data supports a role for activation of circulating monocytes contributing to the enhanced CVD risk in patients with CKD.

On the basis of the significant reduction in CV events after lipid-lowering treatment in patients with CKD,³⁷ the Kidney Disease Improving Global Outcomes (KDIGO) guidelines recently proposed that for those >50 years, an eGFR <60 ml/min per 1.73 m² is an indication for primary prevention using statins, independent of baseline cholesterol levels and other risk factors,³⁸ and Schneider et al. showed that implementation of this guideline requires a substantial increase in prescription rates.³⁹ Our study provides a potential explanation for the benefits of statin therapy in patients with CKD³⁷ in the absence of elevated cholesterol, as several studies have shown direct effects of statin therapy on arterial inflammation.^29,40,41 However, arterial wall inflammation in our study correlated with measures of kidney decline, and not with lipids. Post hoc analyses of the Treating to New Targets (TNT) and Stroke Prevention by Aggressive Reduction in Cholesterol Levels (SPARCL) trials also suggest beneficial effects of statins on eGFR,^42,43 and this proposes an additional mechanism by which statins may reduce CV risk in patients with CKD. Combined, these data imply that the reduction in arterial inflammation after statin therapy may contribute to lowering of CVD risk in patients with CKD. Future studies are necessary to investigate if statin therapy fully suffices to lower CKD driven inflammation and subsequent CV risk, or if more specific therapies are warranted that target the direct inflammatory effects of decreased kidney function.

Potential Limitations

Several limitations of this study merit closer consideration. First, because of the limited sample size and cross-sectional design, we cannot draw definite conclusions on causality of our findings. Traditional risk factors are also increased in patients with CKD,⁴⁴ and distinguishing which particular factors dominate CVD risk is challenging. However, in this study, baseline characteristics were comparable between CKD subjects and controls and, where applicable, outcomes were corrected for any (non-CKD–related) parameter that differed at baseline. However, as has been widely recognized, virtually all forms of kidney disease coincide with hypertension.⁴⁵ Therefore, full dissection of the impact of kidney disease versus BP increase is challenging. Second, due to ethical restraints considering radiation exposure, and the need of fresh monocytes for the ex vivo studies, different control groups were used for the imaging and monocyte studies. Both control groups were comparable regarding baseline characteristics, minimizing the chance that this would significantly affect the assessed parameters. Finally, although ¹⁸F-FDG uptake has been shown to predict subsequent events in the general population,¹² whether this also holds true for patients with CKD requires long-term outcome studies.

Summary and Conclusions

This is the first study to show increased arterial wall inflammation in patients with CKD, without known atherosclerotic or inflammatory disease, and with few traditional risk factors and comorbidity. Arterial wall inflammation correlates directly to measures of kidney function, implying a potentially causal relationship between these observations. As arterial inflammation is directly associated with CVD risk, these data support the need for early CVD prevention, as was recently incorporated in the new KDIGO guidelines on statin use. Alternatively, specific anti-inflammatory therapies may be required to further reduce the large CVD burden in patients with CKD, as LDL cholesterol is not the dominant risk factor in these patients.

Concise Methods

Study Population

We performed a cross-sectional cohort study in subjects >50 years old with a diagnosis of CKD stages 3 or 4, defined as eGFR <60 but >15 ml/min per 1.73 m². Exclusion criteria comprised any chronic inflammatory disease (including GN), gout, diabetes, history of a CV event, uncontrolled hypertension, proteinuria >1 g/L, overt lipid disorders, and use of anti-inflammatory drugs. ¹⁸F-FDG PET/CT scans were performed as published previously.^9,12 Images were analyzed with dedicated software (OsiriX, Geneva, Switzerland; http://www.osirix-viewer.com). Standardized uptake values were averaged for each artery, and divided by the average venous background activity to obtain the TBR.¹⁰ Also, computed tomography scans were used to calculate coronary calcium scores as described previously.¹² Because of ethical constraints relating to radiation exposure, for the imaging studies, healthy controls were selected from a contemporaneous study using identical imaging protocols and performed on the same scanner. For the ex vivo monocyte studies, healthy controls matched on age and sex were included on the same study days as the patients with CKD. Monocyte phenotype was determined ex vivo using flow cytometry, classifying monocytes on the basis of HLA-DR, CD14, and CD16 expression, and subsequently assessing surface marker chemokine (CCR2, CCR7, and CCR5) expression. Migratory capacity was assessed using a transendothelial migration assay, as described previously.¹³ CD14 bead-isolated monocytes were added to a confluent layer of human aortic endothelial cells (HAECs) for 30 minutes and then fixed with formaldehyde. Multiple images were recorded and adhered (bright morphology), and transmigrated monocytes (dark morphology) were quantified. The study protocol was approved by the Institutional Review Board of the Academic Medical Center in Amsterdam, The Netherlands. Written informed consent was obtained from each participant. Full details on the methods are available in the Supplemental Material.

Statistical Analyses

All data were analyzed using Prism version 5.0 (GraphPad Software, La Jolla, CA) and SPSS version 21.0 (SPSS Inc., Chicago, IL). Data are presented as mean±SD for normally distributed data, or medians with interquartile range for non-normally distributed data, unless stated otherwise. Depending on distribution, all comparisons of subgroups were performed using t test or Mann–Whitney U test. Differences in TBR were assessed with covariance analysis, correcting for possible confounders which were different at baseline. Univariate and multivariate linear regression analysis was used to assess the influence of clinical parameters on arterial wall inflammation.

Disclosures

E.S.S. has received lecturing fees from Merck (Kenilworth, NJ), Novartis (Basel, Switzerland), Ionis Pharmaceuticals (Carlsbad, CA), Amgen (Thousand Oaks, CA), and Sanofi-Aventis (Paris, France), none of which are related to the contents of this manuscript. All other authors have no disclosures.