Associations of Soluble CD14 and Endotoxin with Mortality,
Cardiovascular Disease, and Progression of Kidney disease
among Patients with Chronic Kidney Disease
Ruben Poesen,1 MD ; Ali Ramezani,2 PhD; Kathleen Claes,1 MD PhD; Patrick Augustijns,³
Pharm PhD; Dirk Kuypers,1 MD PhD; Ian R Barrows,4 MD; Jagadeesan Muralidharan,2 MD;
Pieter Evenepoel,1 MD PhD; Björn Meijers,1 MD PhD; Dominic S Raj,2 MD
1Department
of Microbiology and Immunology, Division of Nephrology, University Hospitals
Leuven, B-3000 Leuven, Belgium
2Division
of Renal Diseases and Hypertension, George Washington University, Washington,
DC 20037, USA
³Department of Pharmaceutical and Pharmacological Sciences, Drug Delivery and
Disposition, University of Leuven, B-3000 Leuven, Belgium
4George
Washington University School of Medicine, Washington DC 20037, USA
Tables: 4 – Figures: 4
Supplementary tables: 4 – Supplementary Figures: 2
Running title: Soluble CD14 in CKD
Key words: Soluble CD14, – endotoxin, inflammation, microbiome, chronic kidney disease
Address for correspondence: Dominic Raj, MD
Division of Renal Diseases and Hypertension
The George Washington University School of Medicine
2150 Pennsylvania Avenue NW
Washington, DC 20037, USA
Tel. 202 741 2283
Fax. 202 741 2285
E-mail: draj@mfa.gwu.edu
Address for reprints:
Björn Meijers, MD, PhD
Division of Internal Medicine, Department of Nephrology
University Hospitals Leuven
Herestraat 49, B-3000 Leuven, Belgium
Tel. +32 16 344580
Fax. +32 16 344599
E-mail: bjorn.meijers@uzleuven.be
1
Abstract
Background and objectives CD14 plays a key role in the innate immunity as patternrecognition receptor of endotoxin. Higher levels of soluble CD14 (sCD14) are associated with
overall mortality in hemodialysis patients. The influence of kidney function on plasma sCD14
levels, as well its relationship with adverse outcomes in CKD patients not yet on dialysis is
unknown. This study examines the associations between plasma levels of sCD14 and
endotoxin with adverse outcomes in CKD patients.
Design, setting, participants, & measurements We measured plasma levels of sCD14 and
endotoxin in 495 Leuven Mild-to-Moderate CKD Study participants. Mild to moderate CKD
was defined as presence of kidney damage or eGFR < 60 ml/min/1.73m² for ≥ 3 months,
with exclusion of patients on renal replacement therapy. Study participants were enrolled
between November 2005 and September 2006
Results Plasma sCD14 was negatively associated with eGFR (ρ -0.34, P < 0.001). During a
median follow-up of 54 [IQR 23 – 58] months, 53 patients died. Plasma sCD14 was predictive
of mortality, even after adjustment for renal function, Framingham risk factors, markers of
mineral bone metabolism as well as nutritional and inflammatory parameters (HR per SD
higherof 1.90 [1.32, 2.74], P < 0.001). After adjustment for the same risk factors, plasma
sCD14 was also a predictor of cardiovascular disease (HR 1.30 [1.00, 1.69], P 0.05). Although
plasma sCD14 was associated with progression of CKD, defined as reaching end-stage renal
disease or doubling of serum creatinine in models adjusted for CKD specific risk factors (HR
1.24, [1.01, 1.52] P 0.04), significance was lost when adjusted for proteinuria (HR 1.19 [0.96,
1.48] P 0.11). There was neither correlation between plasma endotoxin and sCD14 (ρ -0.06,
P 0.20), nor was endotoxin independently associated with adverse outcome during followup.
2
Conclusions Plasma sCD14 is elevated in patients with decreased kidney function and
associates with mortality and cardiovascular disease in CKD patients not yet on dialysis.
3
Introduction
Chronic kidney disease (CKD) is an important and independent risk factor for cardiovascular
disease (CVD) and death.(1) In an ongoing search for determinants underlying the increased
incidence of adverse outcomes in CKD, subclinical endotoxemia may be an attractive factor
to explore. The human gut is host to 100 trillion commensal organisms, which contributes to
an enteric reservoir of about 1 g of endotoxin.(2) Impaired gut barrier function in CKD could
permit translocation of gut-derived endotoxin in to the systemic circulation.(3-5) The
Bruneck Study showed that elevated plasma level of endotoxin is associated with CVD in the
general population.(6) Endotoxemia has also been shown to be related to inflammation and
atherosclerosis in peritoneal dialysis patients.(7) However, the impact of subclinical
endotoxemia in patients with CKD has not been fully elucidated.
Actions of endotoxin or lipopolysaccharide (LPS) are mediated by interaction with the
toll-like receptor 4/MD-2 complex and its co-receptor CD14 on monocytic cells, initiating an
innate immune response and pro-inflammatory signaling.(8;9) (Supplemental Figure 1) CD14
is either membrane-bound with a glycosylphosphatidylinositol anchor or present as a soluble
molecule (sCD14) after secretion or enzymatically cleavage.(10) sCD14 can participate in cell
activation indirectly by promoting transfer of LPS to membrane-bound CD14(11) or directly
by transferring LPS to toll-like receptor 4/ MD-2 complex on cells that do not express
membrane-bound CD14.(12) Level of sCD14 increases in response to a LPS challenge.(13)
We previously reported that plasma sCD14, but not endotoxin is a predictor of
mortality in maintenance hemodialysis patients.(14;15)
However, endotoxemia in
hemodialysis patients may be due to translocation from dialysate into the blood
4
compartment.(16) Thus, it is important to understand the determinants of circulating
endotoxin and plasma sCD14 in patients with CKD. In this study we tested the hypothesis
that elevated plasma sCD14 and endotoxin are associated with CVD, mortality and loss of
kidney function in CKD patients not yet on dialysis.
Materials and Methods
Study Population
This was an ancillary study of the LeuvenMild-to-Moderate CKD cohort, a prospective cohort
to investigate the role of uremic retention solutes in patients with CKD not yet on dialysis
(clinicaltrials.gov NCT00441623).(17) Prevalent CKD patients, followed at the nephrology
outpatient clinic of the University Hospitals Leuven were eligible for inclusion with
enrollment between November 2005 and September 2006. Mild to moderate CKD was
defined as presence of kidney damage (i.e., pathologic abnormalities or abnormalities in
urine or imaging tests) or eGFR by CKD-EPI equation < 60 ml/min/1.73m² for ≥ 3 months,
with exclusion of patients on renal replacement therapy. The study was performed
according to the Declaration of Helsinki and approved by the ethics committee of the
University Hospitals Leuven. Informed consent was obtained from all patients.
Laboratory Measurements
At inclusion, blood was taken for measurement of hemoglobin (g/dl), albumin (g/l), Creactive protein (mg/l), cholesterol (mg/dl), calcium (mg/dl), phosphate (mg/dl), biointact
parathyroid hormone (PTH) (ng/l), creatinine (mg/dl), urea (mg/dl), sCD14 (µg/ml) and
endotoxin (EU/ml). Hemoglobin, albumin, C-reactive protein, cholesterol, calcium,
5
phosphate, biointact PTH, creatinine and urea were all measured using standard laboratory
techniques. The eGFR was calculated using the CKD-EPI equation.(18) Plasma sCD14
concentration was determined using ELISA (sCD14 Quantikine ELISA Kit, R&D Systems, MN,
USA; minimum detectable level of 125 pg/ml). The inter- and intra-assay coefficient of
variations are <7.5 and <6.5%, respectively. Plasma levels of endotoxin were quantified by
kinetic chromogenic LAL assay (Kinetic-QCL, Lonza, MD, USA; sensitivity range of 0.005-50
EU/ml). The coefficient of variation was 3-9%. For sCD14 and endotoxin, all samples were
tested in duplicate and mean values were reported.
Endpoint Evaluation
Patients were prospectively followed at 3- to 6-month intervals until December 31,
2010 with recording ofpredefined endpoints : i.e., overall mortality, first CV event and
progression of CKD. Cause of death was classified as either CV, infectious, malignancy, or
other. CV deaths included fatal myocardial infarction, sudden death, and death due to
congestive heart failure. Out-of-hospital deaths were coded after consultation of the general
practitioner. First CV event was a composite of cardiac death, non-lethal cardiac event,
ischemic stroke, or peripheral vascular disease, whichever occurred first. Non-lethal cardiac
events included myocardial infarction, diagnosed based on elevated levels of cardiac
enzymes and/or typical electrocardiography changes, myocardial ischemia with typical
electrocardiography changes without elevated cardiac enzymes, coronary intervention
(thrombolysis, percutaneous coronary intervention, or coronary artery bypass grafting), and
ventricular arrhythmia. Ischemic stroke was defined as a neurologic deficit lasting more than
24 hours. Peripheral vascular disease included new-onset ischemic lower limb pain, with
abnormal ankle brachial pressure index or radiologic evidence of peripheral vascular disease,
6
new-onset ischemic necrotic lesions, or surgical arterial intervention. Progression of CKD was
defined as reaching end-stage renal disease or doubling of serum creatinine during followup.
Statistical Analysis
Data are expressed as mean (standard deviation [SD]) for normally distributed
variables or median (interquartile range [IQR]) for non-normally distributed variables.
Differences between baseline variables according to tertiles of plasma sCD14 were tested
using parametric ANOVA, Kruskal-Wallis or chi-squared test as appropriate. Spearman’s rank
correlation coefficients were used to calculate correlations between plasma sCD14 and
other variables.
Independent determinants of plasma sCD14 were identified using
multivariate linear regression analysis. For this purpose, prespecified demographic (i.e., age,
gender, presence of diabetes mellitus, smoking status, body mass index) and biochemical
(i.e., hemoglobin, C-reactive protein, albumin, eGFR, 24 hour proteinuria, endotoxin)
parameters were subjected to a first backward elimination procedure on P < 0.2 anda final
backward elimination step on P < 0.05. Cumulative incidence of the endpoint was estimated
with the Kaplan-Meier method using the log rank test to compare differences between
tertiles of plasma sCD14. Time to first event analysis was performed using Cox proportional
hazards analysis. For multivariate analysis, a double backward elimination approach was
used with inclusion of all variables at P < 0.20 for secondary backward elimination at P <
0.05. To test the proportionality assumption, each model was tested against log (time) with
the use of time-dependent covariates whenthe proportionality assumption was violated.
With respect to overall mortality, for both Kaplan-Meier and Cox proportional hazard
analysis, data were censored at start of renal replacement therapy, loss to follow-up, or at
7
the end of the study observation period. For analysis of first CV event, censoring was
performed at start of renal replacement therapy, death other than CV, loss to follow-up, or
at the study end, while for progression of CKD, data were censored at death, loss to followup, or at the end of study period. For all statistical analysis, P-values less than 0.05 were
considered significant. Statistical analysis was performed using SAS (version 9.3, the SAS
institute, Cary, NC, USA).
Results
Study Population
Between November of 2005 and September of 2006, 495 prevalent patients with CKD
KDOQI stages 1–5 followed at the nephrology outpatient clinic of the University Hospitals
Leuven, Belgium, were included in the Leuven Mild-to-Moderate CKD Study (clinicaltrials.gov
NCT00441623).(17) Measurements of sCD14 and endotoxin were available in a total of 495
patients (Supplemental Figure 2). Baseline characteristics of the study population are
presented in Table 1. Glomerular disease was the most prevalent underlying renal disease
(27.9 %), followed by vascular disease (10.1 %), diabetic nephropathy (8.7 %), autosomal
dominant polycystic disease (8.7 %) and tubulointerstitial disease (4.0 %).
Determinants of Plasma sCD14
The median level of sCD14 was 3.72 µg/ml (interquartile range (IQR) 3.15 – 4.40).
There was a significant and graded inverse relationship between plasma sCD14 and eGFR
(Spearman’s rank correlation ρ -0.34, P < 0.001) (Figure 1A). Higher level of plasma sCD14
was also significantly correlated with older age (ρ 0.13, P 0.003), female gender (ρ 0.12, P
0.008) and prior CVD (ρ 0.12, P 0.006). sCD14 was also associated with lower levels of
8
hemoglobin (ρ -0.31, P < 0.001), low-density lipoprotein (ρ -0.09, P 0.04) and serum
albumin (ρ -0.10, P 0.03). C-reactive protein (CRP) (ρ 0.20 P < 0.001), phosphate (ρ 0.22, P <
0.001), PTH (ρ 0.18, P < 0.001) and 24h proteinuria (ρ 0.12, P 0.02) were positively associated
with sCD14 (Supplemental Table 1). There was no correlation between plasma sCD14 and
endotoxin (ρ -0.06, P 0.20). Plasma endotoxin was significantly correlated with eGFR (ρ 0.21,
P < 0.001) with higher levels present in patients with eGFR above 60 ml/min/1.73m² (Figure
1B).
In multivariate regression analysis, independent determinants of plasma sCD14 were
body mass index (β -0.005, P 0.008), hemoglobin (β -0.02, P < 0.001), CRP (β 0.04, P < 0.001)
and eGFR (β -0.09, P < 0.001) (Model R² 0.18) (Supplemental Table 2). Plasma endotoxin was
not associated with plasma sCD14 (P 0.24).
sCD14 and Mortality
During a median follow-up of 54 (23 – 58) months, 53 patients died (Supplemental
Table 3) with more deaths observed among patients with sCD14 in higher tertiles (Tertile 1
to 3: 12, 17 and 24 events, respectively, log rank P 0.02) (Figure 2). In univariate Cox
proportional hazard analysis, plasma sCD14 was significantly associated with mortality
(Hazard ratio (HR) per SD higher of 1.81 [1.36, 2.41], P < 0.001) (Table 2). This association
remained significant in various multivariate models with adjustment for kidney function,
Framingham risk factors, markers of mineral bone metabolism, CRP and serum albumin (HR
1.90 [1.32, 2.74], P < 0.001). Although we also noted a significant, albeit inverse, relationship
between plasma levels of endotoxin and risk of death in univariate Cox proportional hazard
analysis (HR 0.61 [0.47, 0.79], P < 0.001), plasma endotoxin was not predictive of mortality
9
after adjustment for the same risk factors as in the previous analysis (HR 0.96 [0.70– 1.32], P
0.80).
sCD14 and Cardiovascular Disease
Next, we explored the relationship between plasma sCD14 and CVD. A total of 78 CV
events were observed during follow-up (52 [17 – 58] months) (Supplemental Table 4). The
number of events was higher among patients in higher sCD14 tertiles (Tertile 1 to 3: 18, 29
and 31 events, respectively, log rank P 0.01) (Figure 3). In unadjusted analysis, plasma sCD14
was significantly associated with CV events during follow-up (HR 1.43 [1.13, 1.81], P 0.003)
(Table 3). This association remained highly significant after adjustment for Framingham risk
factors (HR 1.44 [1.12, 1.85], P 0.004). Significance was lost when adjusted for eGFR (HR 1.22
[0.95, 1.57], P 0.11). However, there was significant interaction between eGFR and plasma
sCD14 in this model (P 0.03). In the full multivariate model, plasma sCD14 was a borderline
significant predictor of CV events during follow-up (HR 1.30 [1.00, 1.69], P 0.05). There was
a significant inverse relationship between endotoxin levels and CVD in unadjusted analysis
(HR 0.69 [0.56, 0.86], P < 0.001), but not in the fully adjusted model (HR 0.992 [0.78 – 1.26],
P 0.95).
sCD14 and Progression of CKD
We also examined the potential associations between plasma sCD14 and progression of CKD,
defined as reaching end-stage renal disease or doubling of serum creatinine. During a
median follow-up of 53 (22 – 58) months, 132 patients were classified as having progressive
CKD. We observed more CKD progressors among patients in higher tertiles of sCD14 (Tertile
1 to 3: 23, 47, 62 events, respectively, log rank P < 0.001)) (Figure 4). Plasma sCD14 was a
10
significant predictor of CKD progression in the unadjusted analysis (HR 1.66 [1.39, 1.99], P <
0.001) (Table 4). When adjusted for eGFR, this association was no longer significant (HR 1.20
[0.99, 1.47], P 0.06). In the full model, including CKD-specific and Framingham risk factors, as
well as CRP and serum albumin, plasma sCD14 remained a significant predictor of CKD
progression (HR 1.24 [1.01, 1.52], P 0.04). However, when proteinuria was added in the
model the significance was lost (HR 1.19 [0.96, 1.48], P 0.11). Plasma endotoxin was not
significantly associated with progression of CKD in both unadjusted (HR 0.84 [0.71, 1.00], P
0.06) and full multivariate (HR 1.08 [0.88, 1.33], P 0.44) models.
Discussion
In this study we examined the role of plasma sCD14 and endotoxin in patients with
CKD not yet on dialysis. The key findings are: (i) There is a graded relationship between
plasma sCD14 and renal function with higher levels of plasma sCD14 present in patients with
more advanced CKD; (ii) Plasma sCD14 is a strong and independent predictor of mortality;
(iii) Elevated level of sCD14 is associated with CV events during follow-up; and (iv) There is
no significant association between plasma endotoxin and adverse outcomes in patients with
CKD.
Endotoxin refers to the biologically active lipopolysaccharide (LPS) complex
associated with the outer membrane of Gram-negative bacteria. Endotoxin provokes an
array of host responses by binding to the CD14 receptor.(19) Endotoxin concentrations as
low as 1 pg/mL could induce cellular activation and expression of CD14.(20;21) Recent
evidence
indicates
that
sCD14
may
be
11
derived
by
enzymatic
cleavage
of
glycosylphosphatidylinositol-anchored cell membrane-bound CD14 or secreted by liver in
response to inflammation or infection.(22) Interestingly, sCD14 can both potentiate and
down-regulate responses to LPS by transfer of LPS to lipoproteins for subsequent
removal.(23) Thus, higher levels of sCD14 can reduce the amount of monocyte-bound LPS
and attenuate the inflammatory response.
In multivariate regression analysis, kidney function, body mass index, hemoglobin
and C-reactive protein were all associated with plasma sCD14. Renal function was the
strongest determinant of plasma sCD14 with higher levels of sCD14 observed in patients
with lower eGFR. Reduced renal excretion might also be responsible for accumulation of
plasma sCD14 in CKD.(24) On the other hand, elevated sCD14 could be a response to
subclinical endotoxemia. In this study, we did not observe a significant association between
plasma sCD14 and endotoxin levels. Other investigators have also reported lack of
association between endotoxin and sCD14.(25) Interestingly, we noted higher levels of
endotoxin in patients with eGFR above 60 ml/min/1.73m². Previous studies have reported
elevated or unchanged endotoxin levels in patients with more advanced CKD.(26-30) This
apparent discrepancy may be explained by several factors. First, it must be noted that halflife of endotoxin is very short, with approximately 90 % of LPS being removed within 5
minutes after entrance in the systemic circulation.(31) Thus, it may be questioned whether
measurement of systemic LPS is representative of the actual exposure to endotoxin. In this
regard, it may also be more logical to consider the host response to endotoxemia, of which
plasma sCD14 may be a good surrogate. Second, the limulus amebocyte lysate (LAL) test
assay for endotoxin may be inaccurate. It has been reported that various plasma
components may interfere with detection of LPS by the LAL test.(32;33) Finally, it must be
12
noted that CD14 is a multifunctional receptor molecule that responds to endotoxin from
gram-negative bacteria as well as multiple other ligands,(34;35) possibly confounding the
relationship between plasma endotoxin and sCD14, in which sCD14 may be better
considered a general marker of monocyte activation.(36)
We found that higher plasma level of sCD14 is predictive of overall mortality, even
after adjustment for renal function, CV risk factors, markers of mineral bone metabolism, as
well as inflammatory and nutritional parameters. Investigators have reported that higher
levels of plasma sCD14 are associated with increased mortality in patients
with HIV
infection(25) and those with gram negative sepsis.(37) We also explored the relationship
between sCD14 and CV events, again demonstrating the prognostic role of sCD14 with
respect to CVD in CKD. In patients without CKD, elevated sCD14 level is associated with
aortic stiffness, carotid plaque formation and unstable angina.(38;39) In cardiovascular
health study, sCD14 levels were associated with both subclinical vascular disease and with
risk of future clinical CVD.(40) sCD14 is capable of mediating LPS-activation of mCD14negative cells such as endothelial and smooth muscle cells, which is important in
atherogenesis.(41)
The relationship between plasma sCD14 and CKD progression has not been explored.
We noted that sCD14 was associated with progression of CKD after adjustment for CKDspecific and CV risk factors, as well as inflammatory and nutritional parameters. However,
when adjusted for proteinuria, the significance of the association was lost. Assuming that
sCD14 is a marker of endotoxin burden, this finding may expand the spectrum of renal
disease due to endotoxemia from sepsis-related acute kidney injury to CKD. (42) Even low
13
concentrations of endotoxin are capable of inducing tubular cell injury as sCD14 largely
enhances sensitivity to endotoxin.(24) Urine sCD14 may also be increased in patients with
proteinuria,(24) which either may be a consequence or a cause of proximal tubulopathy.
Furthermore, it has been shown that CD14 is upregulated in tubular cells after kidney injury,
probably mediating apoptosis.(43) In a small study in patients with autosomal dominant
polycystic disease, expression of CD14 in the kidney, urine sCD14 and plasma sCD14 were
also related to progression of polycystic disease.(44) Further research is necessary to identify
the role of plasma sCD14 with respect to CKD progression.
It remains to be elucidated why plasma sCD14 associates with adverse outcomes in
patients with renal dysfunction. Although plasma levels of sCD14 have been linked to
exposure to endotoxin,(13;45) we did not observe a correlation between endotoxin and
sCD14 levels. Furthermore, plasma sCD14, but not endotoxin levels were predictive of
adverse outcomes in our cohort, a finding also noted in previous studies.(14;46) This does
not, however preclude a pathophysiological role of endotoxemia in CKD as explained above.
Whatever the cause may be, higher plasma sCD14 are indicative of a persistent activation of
monocytic cells, which may contribute to the micro-inflammation in CKD and associated
adverse outcomes.(47)
There are limitations to our study. First, the observational study design precludes
causal inferences. Second, unavailability of urine sCD14 measurements does not allow
gaining more insights in the renal handling of sCD14. Third, our study population mainly
consists of patients of Caucasian origin. Care must be taken when extrapolating our data to
14
other patient populations. Finally, our analyses are based on single baseline biochemical
measurements that are potentially subjected to certain variability over time.
In conclusion, levels of plasma sCD14 are elevated in patients with more advanced
CKD. In addition, elevated plasma sCD14 is associated with mortality and CVD. The
association between sCD14 and CKD progression was significantly attenuated by addition of
proteinuria in the model. Plasma endotoxin level was not associated with any clinical
outcome in our study. The potential utility of sCD14 as a prognostic and therapeutic target in
CKD requires further investigation.
15
Acknowledgments
RP is the recipient of a Ph.D. fellowship of the Research Foundation - Flanders (FWO) (grant
11E9813N). Part of the research has been funded by the Research Foundation - Flanders
(FWO) (grant G077514N). DR is supported by the National Institutes of Health Grants
1R01DK073665-01A1, 1U01DK099924-01 and 1U01DK099914-01.
Disclosures
None.
16
References
1. Go,A.S., Chertow,G.M., Fan,D., McCulloch,C.E., and Hsu,C.Y., Chronic kidney disease and the
risks of death, cardiovascular events, and hospitalization. N.Engl.J.Med 351, 1296-1305, 2004.
2. Berg,R.D., The indigenous gastrointestinal microflora. Trends Microbiol. 4, 430-435, 1996.
3. Vaziri,N.D., Wong,J., Pahl,M., Piceno,Y.M., Yuan,J., Desantis,T.Z., Ni,Z., Nguyen,T.H., and
Andersen,G.L., Chronic kidney disease alters intestinal microbial flora. Kidney Int. 2012.
4. Vaziri,N.D., Yuan,J., Rahimi,A., Ni,Z., Said,H., and Subramanian,V.S., Disintegration of colonic
epithelial tight junction in uremia: a likely cause of CKD-associated inflammation.
Nephrol.Dial.Transplant. 27, 2686-2693, 2012.
5. Magnusson,M., Magnusson,K.E., Sundqvist,T., and Denneberg,T., Impaired intestinal barrier
function measured by differently sized polyethylene glycols in patients with chronic renal
failure. Gut 32, 754-759, 1991.
6. Wiedermann,C.J., Kiechl,S., Dunzendorfer,S., Schratzberger,P., Egger,G., Oberhollenzer,F., and
Willeit,J., Association of endotoxemia with carotid atherosclerosis and cardiovascular disease:
prospective results from the Bruneck Study. J Am Coll.Cardiol. 34, 1975-1981, 1999.
7. Szeto,C.C., Kwan,B.C., Chow,K.M., Lai,K.B., Chung,K.Y., Leung,C.B., and Li,P.K., Endotoxemia is
related to systemic inflammation and atherosclerosis in peritoneal dialysis patients. Clin J Am
Soc Nephrol 3, 431-436, 2008.
8. Miyake,K., Innate recognition of lipopolysaccharide by Toll-like receptor 4-MD-2. Trends
Microbiol. 12, 186-192, 2004.
9. Wright,S.D., Ramos,R.A., Tobias,P.S., Ulevitch,R.J., and Mathison,J.C., CD14, a receptor for
complexes of lipopolysaccharide (LPS) and LPS binding protein. Science 249, 1431-1433, 1990.
10. Ulevitch,R.J. and Tobias,P.S., Receptor-dependent mechanisms of cell stimulation by bacterial
endotoxin. Annu.Rev.Immunol. 13, 437-457, 1995.
11. Hailman,E., Vasselon,T., Kelley,M., Busse,L.A., Hu,M.C., Lichenstein,H.S., Detmers,P.A., and
Wright,S.D., Stimulation of macrophages and neutrophils by complexes of lipopolysaccharide
and soluble CD14. J Immunol. 156, 4384-4390, 1996.
12. Pugin,J., Schurer-Maly,C.C., Leturcq,D., Moriarty,A., Ulevitch,R.J., and Tobias,P.S.,
Lipopolysaccharide activation of human endothelial and epithelial cells is mediated by
lipopolysaccharide-binding protein and soluble CD14. Proc.Natl.Acad.Sci.U.S.A 90, 2744-2748,
1993.
13. Schutt,C., Schilling,T., Grunwald,U., Schonfeld,W., and Kruger,C., Endotoxin-neutralizing
capacity of soluble CD14. Res.Immunol. 143, 71-78, 1992.
14. Raj,D.S., Carrero,J.J., Shah,V.O., Qureshi,A.R., Barany,P., Heimburger,O., Lindholm,B.,
Ferguson,J., Moseley,P.L., and Stenvinkel,P., Soluble CD14 levels, interleukin 6, and mortality
among prevalent hemodialysis patients. Am.J.Kidney Dis. 54, 1072-1080, 2009.
15. Raj,D.S., Shah,V.O., Rambod,M., Kovesdy,C.P., and Kalantar-Zadeh,K., Association of soluble
endotoxin receptor CD14 and mortality among patients undergoing hemodialysis. Am.J.Kidney
Dis. 54, 1062-1071, 2009.
16. Coulliette,A.D. and Arduino,M.J., Hemodialysis and water quality. Semin.Dial. 26, 427-438,
2013.
17. Meijers,B.K., Claes,K., Bammens,B., de,L.H., Viaene,L., Verbeke,K., Kuypers,D.,
Vanrenterghem,Y., and Evenepoel,P., p-Cresol and cardiovascular risk in mild-to-moderate
kidney disease. Clin.J.Am.Soc.Nephrol. 5, 1182-1189, 2010.
18. Levey,A.S., Stevens,L.A., Schmid,C.H., Zhang,Y.L., Castro,A.F., III, Feldman,H.I., Kusek,J.W.,
Eggers,P., Van,L.F., Greene,T., and Coresh,J., A new equation to estimate glomerular filtration
rate. Ann.Intern.Med. 150, 604-612, 2009.
19. Pugin,J., Heumann,I.D., Tomasz,A., Kravchenko,V.V., Akamatsu,Y., Nishijima,M., Glauser,M.P.,
Tobias,P.S., and Ulevitch,R.J., CD14 is a pattern recognition receptor. Immunity. 1, 509-516,
1994.
17
20. Erridge,C., Attina,T., Spickett,C.M., and Webb,D.J., A high-fat meal induces low-grade
endotoxemia: evidence of a novel mechanism of postprandial inflammation. Am J Clin.Nutr. 86,
1286-1292, 2007.
21. Nockher,W.A. and Scherberich,J.E., Monocyte cell-surface CD14 expression and soluble CD14
antigen in hemodialysis: evidence for chronic exposure to LPS. Kidney Int 48, 1469-1476, 1995.
22. Bas,S., Gauthier,B.R., Spenato,U., Stingelin,S., and Gabay,C., CD14 is an acute-phase protein. J
Immunol. 172, 4470-4479, 2004.
23. Kitchens,R.L., Thompson,P.A., Viriyakosol,S., O'Keefe,G.E., and Munford,R.S., Plasma CD14
decreases monocyte responses to LPS by transferring cell-bound LPS to plasma lipoproteins. J
Clin Invest 108, 485-493, 2001.
24. Bussolati,B., David,S., Cambi,V., Tobias,P.S., and Camussi,G., Urinary soluble CD14 mediates
human proximal tubular epithelial cell injury induced by LPS. Int.J.Mol.Med 10, 441-449, 2002.
25. Sandler,N.G., Wand,H., Roque,A., Law,M., Nason,M.C., Nixon,D.E., Pedersen,C.,
Ruxrungtham,K., Lewin,S.R., Emery,S., Neaton,J.D., Brenchley,J.M., Deeks,S.G., Sereti,I., and
Douek,D.C., Plasma levels of soluble CD14 independently predict mortality in HIV infection. J
Infect.Dis. 203, 780-790, 2011.
26. Nisbeth,U., Hallgren,R., Eriksson,O., and Danielson,B.G., Endotoxemia in chronic renal failure.
Nephron 45, 93-97, 1987.
27. Goncalves,S., Pecoits-Filho,R., Perreto,S., Barberato,S.H., Stinghen,A.E., Lima,E.G.,
Fuerbringer,R., Sauthier,S.M., and Riella,M.C., Associations between renal function, volume
status and endotoxaemia in chronic kidney disease patients. Nephrol.Dial.Transplant. 21, 27882794, 2006.
28. Szeto,C.C., Kwan,B.C., Chow,K.M., Lai,K.B., Chung,K.Y., Leung,C.B., and Li,P.K., Endotoxemia is
related to systemic inflammation and atherosclerosis in peritoneal dialysis patients.
Clin.J.Am.Soc.Nephrol. 3, 431-436, 2008.
29. McIntyre,C.W., Harrison,L.E., Eldehni,M.T., Jefferies,H.J., Szeto,C.C., John,S.G., Sigrist,M.K.,
Burton,J.O., Hothi,D., Korsheed,S., Owen,P.J., Lai,K.B., and Li,P.K., Circulating endotoxemia: a
novel factor in systemic inflammation and cardiovascular disease in chronic kidney disease.
Clin.J.Am.Soc.Nephrol. 6, 133-141, 2011.
30. Taniguchi,T., Katsushima,S., Lee,K., Hidaka,A., Konishi,J., Ideguchi,H., and Kawaguchi,Y.,
Endotoxemia in patients on hemodialysis. Nephron 56, 44-49, 1990.
31. HERRING,W.B., HERION,J.C., WALKER,R.I., and PALMER,J.G., Distribution and clearance of
circulating endotoxin. J.Clin Invest 42, 79-87, 1963.
32. Cohen,J., The detection and interpretation of endotoxaemia. Intensive Care Med 26 Suppl 1,
S51-S56, 2000.
33. Mattsby-Baltzer,I., Lindgren,K., Lindholm,B., and Edebo,L., Endotoxin shedding by
enterobacteria: free and cell-bound endotoxin differ in Limulus activity. Infect.Immun. 59, 689695, 1991.
34. Pugin,J., Heumann,I.D., Tomasz,A., Kravchenko,V.V., Akamatsu,Y., Nishijima,M., Glauser,M.P.,
Tobias,P.S., and Ulevitch,R.J., CD14 is a pattern recognition receptor. Immunity. 1, 509-516,
1994.
35. Devitt,A., Moffatt,O.D., Raykundalia,C., Capra,J.D., Simmons,D.L., and Gregory,C.D., Human
CD14 mediates recognition and phagocytosis of apoptotic cells. Nature 392, 505-509, 1998.
36. Nockher,W.A., Wigand,R., Schoeppe,W., and Scherberich,J.E., Elevated levels of soluble CD14
in serum of patients with systemic lupus erythematosus. Clin Exp Immunol. 96, 15-19, 1994.
37. Landmann,R., Zimmerli,W., Sansano,S., Link,S., Hahn,A., Glauser,M.P., and Calandra,T.,
Increased circulating soluble CD14 is associated with high mortality in gram-negative septic
shock. J Infect.Dis. 171, 639-644, 1995.
38. Zalai,C.V., Kolodziejczyk,M.D., Pilarski,L., Christov,A., Nation,P.N., Lundstrom-Hobman,M.,
Tymchak,W., Dzavik,V., Humen,D.P., Kostuk,W.J., Jablonsky,G., Pflugfelder,P.W., Brown,J.E.,
and Lucas,A., Increased circulating monocyte activation in patients with unstable coronary
syndromes. J Am Coll.Cardiol 38, 1340-1347, 2001.
18
39. Amar,J., Ruidavets,J.B., Bal Dit,S.C., Bongard,V., Boccalon,H., Chamontin,B., Drouet,L., and
Ferrieres,J., Soluble CD14 and aortic stiffness in a population-based study. J Hypertens. 21,
1869-1877, 2003.
40. Reiner,A.P., Lange,E.M., Jenny,N.S., Chaves,P.H., Ellis,J., Li,J., Walston,J., Lange,L.A.,
Cushman,M., and Tracy,R.P., Soluble CD14: genomewide association analysis and relationship
to cardiovascular risk and mortality in older adults. Arterioscler.Thromb.Vasc.Biol 33, 158-164,
2013.
41. Pugin,J., Schurer-Maly,C.C., Leturcq,D., Moriarty,A., Ulevitch,R.J., and Tobias,P.S.,
Lipopolysaccharide activation of human endothelial and epithelial cells is mediated by
lipopolysaccharide-binding protein and soluble CD14. Proc.Natl.Acad.Sci.U.S.A 90, 2744-2748,
1993.
42. Schrier,R.W. and Wang,W., Acute renal failure and sepsis. N.Engl.J.Med 351, 159-169, 2004.
43. Morrissey,J., Guo,G., McCracken,R., Tolley,T., and Klahr,S., Induction of CD14 in tubular
epithelial cells during kidney disease. J.Am.Soc.Nephrol 11, 1681-1690, 2000.
44. Zhou,J., Ouyang,X., Cui,X., Schoeb,T.R., Smythies,L.E., Johnson,M.R., Guay-Woodford,L.M.,
Chapman,A.B., and Mrug,M., Renal CD14 expression correlates with the progression of cystic
kidney disease. Kidney Int. 78, 550-560, 2010.
45. Nockher,W.A. and Scherberich,J.E., Monocyte cell-surface CD14 expression and soluble CD14
antigen in hemodialysis: evidence for chronic exposure to LPS. Kidney Int. 48, 1469-1476, 1995.
46. Sandler,N.G., Wand,H., Roque,A., Law,M., Nason,M.C., Nixon,D.E., Pedersen,C.,
Ruxrungtham,K., Lewin,S.R., Emery,S., Neaton,J.D., Brenchley,J.M., Deeks,S.G., Sereti,I., and
Douek,D.C., Plasma levels of soluble CD14 independently predict mortality in HIV infection.
J.Infect.Dis. 203, 780-790, 2011.
47. Kato,S., Chmielewski,M., Honda,H., Pecoits-Filho,R., Matsuo,S., Yuzawa,Y., Tranaeus,A.,
Stenvinkel,P., and Lindholm,B., Aspects of immune dysfunction in end-stage renal disease. Clin
J.Am.Soc.Nephrol 3, 1526-1533, 2008.
19
Figure Legends
Figure 1: Plasma soluble CD14, endotoxin and renal function. Distribution of plasma soluble
CD14 (A) and endotoxin (B) levels as a function of eGFR. Between-group comparisons are
Bonferroni-corrected.
Figure 2: Kaplan-Meier showing death by tertiles of sCD14 concentration. Tertile 1 to 3: 12,
17 and 24 deaths, respectively. Log rank test P 0.02.
Figure 3: Kaplan-Meier curve of time to first cardiovascular event by tertiles of sCD14
concentration. Tertile 1 to 3: 18, 29 and 31 cardiovascular events, respectively. Log rank test
P 0.01.
Figure 4: Kaplan-Meier renal survival curve by tertiles of sCD14 concentration. Tertile 1 to 3:
23, 47 and 62 renal progressors, respectively. Log rank test P 0.01.
Supplemental Figure 1: CD14 is a Pattern Recognition Receptor (PRR) that binds to
lipopolysaccharide (LPS). It is expressed on most innate immune response cells and exists
either in an anchored membrane form (mCD14) or in a circulating soluble form (sCD14). The
latter is a 43-53 kD glycoprotein that derives from either protease-mediated membrane
CD14 shedding or possibly liver synthesis. sCD14 can activate cells participate in cell
activation by facilitating transfer of LPS to mCD14 or by transferring LPS to toll-like receptor
4/ MD-2 complex on cells that do not express membrane-bound CD14.
LBP, lipopolysaccharide binding protein; NF-κB, Nuclear factor-κB -4; HDL, high density
lipoprotein.
20
Supplemental Figure 2: Flow chart demonstrating patient screening and inclusion.
21
Tables
Table 1 – Baseline characteristics of study population
Variable
Age (yr)
Gender: male (%)
Prior CVD: yes (%)
Diabetes: yes (%)
Current smoker: yes (%)
Body mass index (kg/m²)
Systolic blood pressure (mmHg)
Diastolic blood pressure (mmHg)
Hemoglobin (g/dl)
Albumin (g/dl)
C-reactive protein (mg/l)
Cholesterol (mg/dl)
LDL (mg/dl)
HDL (mg/dl)
Calcium (mg/dl)
Phosphate (mg/dl)
Parathormone (pg/ml)
Creatinine (mg/dl)
eGFR (ml/min per 1.73 m²)
24 hour proteinuria (g)
Therapy with ACEI/ARB: yes (%)
Therapy with statins: yes (%)
Therapy with 25-OH-vitamin D: yes (%)
Overall:
1.92 – 7.29 µg/ml
(n = 495)
64 (50 – 75)
271 (54.7)
137 (27.7)
89 (18.0)
90 (18.2)
25.69 (22.99 – 29.06)
135 (120 – 150)
80 (70 – 85)
13.4 (1.8)
4.48 (4.24 – 4.68)
2 (1 – 6)
178 (152 – 205)
85 (67 – 112)
57 (48 – 72)
9.6 (9.2 – 9.9)
3.3 (2.9 – 3.8)
23.8 (12.7 – 52.0)
1.79 (1.27 – 2.47)
34 (23 – 55)
0.28 (0.11 – 0.88)
349 (70.5)
238 (48.1)
92 (18.6)
Tertile 1:
1.92 – 3.32 µg/ml
(n = 165)
sCD14
Tertile 2:
3.32 – 4.11 µg/ml
(n = 165)
Tertile 3:
4.12 – 7.29 µg/ml
(n = 165)
58 (46 – 72)
105 (63.6)
33 (20.0)
25 (15.2)
30 (18.2)
25.96 (23.41 – 30.02)
130 (120 – 150)
80 (70 – 88)
14.0 (1.8)
4.56 (4.24 – 4.74)
1 (1 – 3)
179 (156 – 207)
89 (68 – 115)
55 (48 – 70)
9.6 (9.3 – 9.9)
3.2 (2.8 – 3.6)
17.5 (9.3 – 37.4)
1.51 (1.09 – 1.97)
46 (31 – 72)
0.19 (0.10 – 0.59)
118 (71.5)
75 (45.5)
23 (13.9)
70 (52 – 77)
84 (50.9)
51 (30.9)
34 (20.6)
27 (16.4)
25.51 (22.49 – 29.45)
140 (125 – 160)
80 (70 – 85)
13.3 (1.8)
4.45 (4.16 – 4.67)
2 (1 – 5)
178 (156 - 201)
88 (69 – 114)
58 (49 – 71)
9.6 (9.2 – 9.9)
3.4 (3.0 – 3.8)
27.1 (16.2 – 58.9)
1.83 (1.34 – 2.56)
32 (22 – 50)
0.29 (0.11 – 1.08)
116 (70.3)
82 (49.7)
26 (15.8)
64 (53 – 74)
82 (49.7)
53 (32.1)
30 (18.2)
33 (20.0)
24.91 (22.86 – 28.34)
135 (120 – 150)
80 (70 – 80)
12.8 (1.5)
4.47 (4.29 – 4.64)
3 (1 – 7)
176 (148 – 205)
81 (61 – 104)
58 (47 – 74)
9.5 (9.2 – 9.9)
3.5 (3.1 – 4.0)
30.2 (14.3 – 57.9)
2.00 (1.46 – 2.86)
28 (18 – 41)
0.43 (0.12 – 1.01)
115 (69.7)
81 (49.1)
43 (26.1)
22
P
< 0.001
0.01
0.01
0.47
0.67
0.12
0.02
0.26
< 0.001
0.007
< 0.001
0.45
0.07
0.67
0.85
< 0.001
< 0.001
< 0.001
< 0.001
0.04
0.72
0.51
0.005
Therapy with phosphate binder: yes (%)
sCD14 (µg/ml)
Endotoxin (EU/ml)
136 (27.5)
3.72 (3.15 – 4.40)
0.71 (0.38 – 1.3)
31 (18.1)
2.93 (2.65 – 3.15)
0.82 (0.45 – 1.49)
50 (30.3)
3.72 (3.52 – 3.89)
0.64 (0.34 – 1.23)
55 (33.3)
4.69 (4.40 – 5.22)
0.71 (0.38 – 1.3)
0.003
< 0.001
0.05
Data are expressed as mean (SD) or median (IQR), as appropriate. Differences between tertiles were tested using parametric ANOVA, KruskalWallis of chi-squared test, as appropriate. CVD, cardiovascular disease; LDL, low-density lipoprotein; HDL, high-density lipoprotein; eGFR,
estimated glomerular filtration rate; ACEI, angiotensin-converting enzyme inhibitor; ARB, angiotensin receptor blocker; 25-OH-vitamin D, 25hydroxy-vitamine D; sCD14, soluble CD14.
23
Table 2 – Multivariate analysis showing association of plasma sCD14 and mortality
Variable
Hazard ratio (95% CI)
P
per SD higher
1. Unadjusted: sCD14 (Ln)
1.81 (1.36, 2.41)
< 0.001
2. Adjusted for renal function: eGFR (Ln)
1.65 (1.21, 2.26)
0.002
3. Adjusted for Framingham risk factors: age, gender,
systolic blood pressure, current smoker, diabetes mellitus,
cholesterol
1.86 (1.37, 2.55)
< 0.001
4. Full model: Adjusted for creatinine (Ln), age, gender,
systolic blood pressure, current smoker, diabetes mellitus,
cholesterol, calcium, phosphate, parathyroid hormone
(Ln), C-reactive protein (Ln), albumin
1.90 (1.32, 2.74)
< 0.001
24
Table 3 – Cox proportional hazard analysis of time to first cardiovascular event
Variable
Hazard ratio (95% CI) per
P
SD higher
1. Unadjusted: sCD14 (Ln)
1.43 (1.13, 1.81)
0.003
2. Adjusted for renal function: eGFR (Ln)
1.22 (0.95, 1.57)
0.11
3. Adjusted for Framingham risk factors: age, gender,
systolic blood pressure, current smoker, diabetes
mellitus, cholesterol *
1.44 (1.12, 1.85)
0.004
4. Full model: Adjusted for creatinine (Ln), age,
1.30 (1.00, 1.69)
gender, systolic blood pressure, current smoker,
diabetes mellitus, cholesterol, calcium, phosphate,
parathyroid hormone (Ln), C-reactive protein (Ln),
albumin *
*Age and cholesterol entered as time-dependent covariate.
0.05
25
Table 4 – Multivariable Cox proportional hazard model showing association between sCD14
and progression of CKD†
Variable
Hazard ratio (95% CI)
P
per SD higher
1. Unadjusted: sCD14 (Ln)
1.66 (1.39, 1.99)
< 0.001
2. Adjusted for renal function: eGFR (Ln)
1.20 (0.99, 1.47)
0.06
3. Adjusted for Framingham risk factors: age, gender,
systolic blood pressure, current smoker, diabetes mellitus,
cholesterol
1.65 (1.37, 1.98)
< 0.001
4. Full model: Adjusted for creatinine (Ln), hemoglobin,
bicarbonate, calcium, phosphate, parathyroid hormone
(Ln), urea (Ln), therapy with ACE/ARB, age, gender,
systolic blood pressure, current smoker, diabetes mellitus,
cholesterol,
C-reactive
protein (Ln), albumin
+ 24h proteinuria
(n = 412)
1.24 (1.01, 1.52)
0.04
1.19 (0.96 – 1.48)
0.11
†Reaching end-stage renal disease or doubling of serum creatinine during follow-up.
26
Supplemental Table 1 – Spearman’s rank correlation between sCD14 plasma concentration
and baseline characteristics
Variable
ρ
P
Age
Gender (female vs. male)
Prior CVD
Diabetes mellitus
Current smoker
Body mass index
Systolic blood pressure
Diastolic blood pressure
Hemoglobin
Albumin
C-reactive protein
Cholesterol
LDL
HDL
Calcium
Phosphate
Parathormone
Creatinine
eGFR
24h proteinuria
Therapy with ACEI/ARB
Therapy with statin
Therapy with 25-OH-vitamin D
Therapy with phosphate binder
Endotoxin (EU/ml)
0.13
0.12
0.12
0.05
0.03
- 0.09
0.03
- 0.07
- 0.31
- 0.10
0.20
- 0.05
- 0.09
0.005
- 0.02
0.22
0.18
0.30
- 0.34
0.12
- 0.01
0.05
0.12
0.14
- 0.06
0.003
0.008
0.006
0.25
0.50
0.05
0.51
0.14
< 0.001
0.03
< 0.001
0.27
0.04
0.91
0.65
< 0.001
< 0.001
< 0.001
< 0.001
0.02
0.78
0.25
0.009
0.002
0.20
CVD, cardiovascular disease; LDL, low-density lipoprotein; HDL, high-density lipoprotein;
eGFR, estimated glomerular filtration rate; ACEI, angiotensin-converting enzyme inhibitor;
ARB, angiotensin receptor blocker; ; 25-OH-vitamin D, 25-hydroxy-vitamine D; sCD14,
soluble CD14.
27
Supplemental Table 2 – Multivariate regression analysis: Factors associated with sCD14
plasma concentration (Ln)
Variable
Body mass index
Hemoglobin
C-reactive protein (Ln)
eGFR (Ln)
Unit
kg/m²
g/dL
mg/L
ml/min/1.73m²
β
-0.005
-0.02
0.04
-0.09
Model R²
28
P
0.008
< 0.001
< 0.001
< 0.001
0.18
Supplemental Table 3 – Cause of death
Cause (n = 53)
Cardiovascular
Malignancy
Infectious
Other
N (%)
17 (32.1%)
15 (28.3%)
3 (5.7%)
18 (34.0%)
29
Supplemental Table 4 – Cardiovascular events
Events (n = 78)
Non-fatal
Cardiac
New onset angina, conservative
New onset angina, invasive
Acute myocardial infarction
Ventricular arrythmia
Ischemic cerebrovascular accident
Peripheral arterial disease
Fatal
N (%)
66 (84.6%)
30 (38.5%)
11 (14.1%)
6 (7.7%)
10 (12.8%)
3 (3.8%)
5 (6.4%)
31 (39.7%)
12 (15.4%)
30