Changes in Fat Mass Correlate With Changes in Soluble scd163, a Marker of Mature Macrophages, in Patients With CKD

Transcription

1 Changes in Fat Mass Correlate With Changes in Soluble scd163, a Marker of Mature Macrophages, in Patients With CKD Jonas Axelsson, MD, PhD, Holger Jon Møller, MSc, PhD, Anna Witasp, MSc, Abdul Rashid Qureshi, MD, PhD, Juan Jesus Carrero, MSc, PhD, Olof Heimbürger, MD, PhD, Peter Bárány, MD, PhD, Anders Alvestrand, MD, PhD, Bengt Lindholm, MD, PhD, Søren K. Moestrup, MD, PhD, and Peter Stenvinkel, MD, PhD Background: Recently, adipose tissue was shown to contain macrophages capable of contributing to systemic inflammation and cardiovascular disease (CVD). Here, we investigate this putative relationship in patients with chronic kidney disease (CKD) by using the novel macrophage marker soluble (s)cd163. Methods: One hundred twenty patients with CKD stage 5 (mean glomerular filtration rate [GFR], 7 1 ml/min [.12.2 ml/s; mean age, 53 1 years; 65% men), 38 patients with CKD stages 3 to 4 (mean GFR, 33 3 ml/min [.55.5 ml/s]; mean age, 67 2 years; 68% men), and 28 healthy controls (mean GFR, 89 3 ml/min [ ml/s]; mean age, 63 2 years; 69% men) were characterized post hoc with a follow-up of up to 5 years (mean, 47 1 months). scd163 levels, body composition (dual-energy x-ray absorptiometry), clinical parameters, and levels of circulating inflammatory markers (enzyme-linked immunosorbent assay) were assessed at baseline and, in a subset population, after 1 year of dialysis therapy (hemodialysis, n 19; peritoneal dialysis, n 3). Results: scd163 level increased in patients with both severe (median, 4.3 mg/l; range, 1.3 to 23.4 mg/l) and moderate (median, 3.6 mg/l; range, 1.6 to 9.8 mg/l) CKD compared with controls (median, 2.6 mg/l; range,.8 to 7.6 mg/l; P <.1). Furthermore, scd163 levels correlated with both truncal (.17; P <.5) and total (.17; P <.5) fat mass, as well as with all measured markers of inflammation and endothelial adhesion molecules. After 1 year, patients who increased body fat mass (average, 11% 5% versus 5% 5%; P <.5) also showed a significant increase in scd163 levels (median, 2.2 versus.97 mg/l; P <.1). Finally, patients with scd163 levels greater than 4. mg/l had a statistically significantly worse outcome than patients with scd163 levels less than this value, even after adjustment for age, sex, and diabetes mellitus (chi-square 19.98; P <.1). However, this effect was lost after adjustment for either inflammation or CVD. Conclusion: We show that increasing fat mass is associated with increasing levels of scd163, a circulating marker of macrophages also associated with inflammatory biomarkers. We thus hypothesize that adipose tissue macrophages may have a role in the proinflammatory state observed in patients with renal disease. Finally, we propose the term uremic-metabolic syndrome to describe this state of increased adipose tissue signaling in patients with uremia, a phenomenon that may share some characteristics with the metabolic syndrome of obesity. Am J Kidney Dis 48: by the National Kidney Foundation, Inc. INDEX WORDS: Adipose tissue; scd163; inflammation; body composition; chronic kidney disease (CKD); endstage renal disease (ESRD). IRRESPECTIVE OF PRIMARY disease, increased circulating levels of several proinflammatory cytokines are common in patients with chronic kidney disease (CKD) 1 and are associated with a markedly increased risk for cardiovascular disease (CVD) and death. 2,3 We previously showed that increased fat mass is associated with increased circulating interleukin 6 (IL-6) and C-reactive protein (CRP) levels in this patient population. 4 This suggests that adipose tissue may be a clinically important source of inflammation in patients with renal disease. In nonrenal patients, it was recently reported that macrophages infiltrating adipose tissue are an important source of circulating proinflammatory cytokines. 5,6 Furthermore, this infiltration was From the Divisions of Renal Medicine and Baxter Novum and Department of Molecular Medicine, Karolinska Institutet, Stockholm, Sweden; and Departments of Clinical Biochemistry and Medical Biochemistry, Aarhus University Hospital, Aarhus, Denmark. Received May 12, 26; accepted in revised form August 22, 26. Originally published online as doi:1.153/j.ajkd on November 6, 26. Support: This study was supported in part by unconditional grants from the Söderberg Foundation (P.S.), the Swedish Medical Research Council (P.S.), the Swedish Kidney Foundation (J.A.), and the Danish Medical Research Council ( , to H.J.M.). J.J.C. is supported by a Fellowship from the ERA-EDTA. Potential conflicts of interest: B.L. is an employee of Baxter Healthcare Inc. Address reprint requests to Jonas Axelsson, MD, PhD, Divisions of Renal Medicine and Baxter Novum, Department of Clinical Science, Intervention and Technology, K56 Karolinska University Hospital, Stockholm, Sweden. jonas.axelsson@ki.se 26 by the National Kidney Foundation, Inc /6/486-5$32./ doi:1.153/j.ajkd American Journal of Kidney Diseases, Vol 48, No 6 (December), 26: pp

2 FAT MASS IS ASSOCIATED WITH MACROPHAGE NUMBERS IN CKD 917 hypothesized to be the result of local ischemia, which may occur more frequently in very obese patients or during rapid changes in fat metabolism. 7 Unlike in the general population, there appears to be no positive association between body mass index (BMI) and increased cardiovascular mortality in patients with CKD. Rather, in this population, even a morbidly high BMI ( 3 kg/m 2 ) appears to confer a survival advantage over patients with a normal or low BMI, 8,9 and in a recent longitudinal study, Kalantar-Zadeh et al 1 showed that body fat (measured by means of anthropometry) itself confers a survival advantage. To date, mechanistic studies of possible etiologic pathways are lacking. However, because CKD is associated with marked abnormalities in both tissue perfusion 11 and immune function, 12 we hypothesized that the putative relationship between fat mass and macrophages would be relevant in these patients. Moreover, this association could contribute to the balance of proinflammatory and anti-inflammatory signals that may promote or impede the development of CVD in this patient group. In addition, because initiation of dialysis therapy is associated with marked changes in body fat stores, 13,14 we hypothesized that any observed associations would be magnified after the start of dialysis therapy. We thus investigated a cohort of patients with moderate renal dysfunction (CKD stages 3 to 4) and end-stage renal failure (CKD stage 5) starting renal replacement therapy in a longitudinal study relating body composition to the monocyte/ macrophage-specific plasma marker soluble (s)cd163, a novel marker shed from the cell surface upon inflammatory stimuli and therefore a marker of activated macrophage numbers. 15 METHODS Patients One hundred fifty-eight patients with CKD with a mean age of 53 1 years were included, all analyzed post hoc in an ongoing prospective study that followed up patients for up to 5 years (mean, 47 1 months). Of these, 12 were consecutive patients with CKD stage 5 (glomerular filtration rate [GFR] 15 ml/min [.25 ml/s], 65% men; mean age, 53 1 years) enrolled at initiation of renal replacement therapy in the renal program of the Karolinska University Hospital, Huddinge, Sweden, between 1994 and 25 (included as participants of an ongoing prospective study, parts of which we previously described 16 ). We also enrolled 38 prevalent patients (age, 67 2 years; 68% men) with CKD stages 3 to 4 (GFR, 15 to 6 ml/min [.25 to 1. ml/s]) recruited by advertisement from the hospital renal outpatient clinic. Study exclusion criteria were age younger than 18 years or older than 7 years (for patients with CKD stage 5) or 8 years (for patients with CKD stages 3 to 4), clinical signs of acute infection, active vasculitis or liver disease at the time of evaluation, or unwillingness to participate in the study. Causes of CKD were chronic glomerulonephritis in 44 patients, diabetic nephropathy in 37 patients, polycystic kidney disease in 24 patients, and other or unknown in 53 patients. Patients on hemodialysis therapy received 3 to 4 hours of therapy 3 times per week using a high-flux synthetic membrane, whereas peritoneal dialysis patients used glucose- and/or polyglucose-based solutions with 3 or 4 daily exchanges. Eighteen patients (11%) had type 1 diabetes, defined as insulin dependent from the outset, whereas 18 patients (11%) had type 2 diabetes, defined as initially non insulin-dependent diabetes mellitus. In accordance with current therapy recommendations, most patients initially administered oral antiglycemic agents or following restricted diets had been switched to insulin therapy at the time of inclusion in the study. The majority of patients were administered antihypertensive medications (angiotensinconverting enzyme inhibitors and/or angiotensin II receptor antagonists, n 85; -blockers, n 92; calcium channel blockers, n 7) and other drugs commonly used in patients with CKD, such as phosphate and potassium binders, diuretics, erythropoiesis-stimulating agents, iron substitution, and vitamin B, C, and D supplements. Only 14 patients were administered lipid-lowering medication (3- hydroxy-3-methylglutaryl coenzyme A reductase inhibitors). Patients were divided into groups according to degree of renal failure by using the staging recommended by the Kidney Disease Outcomes Quality Initiative guidelines. 17 The Ethics Committee of the Karolinska Institute approved the study at the Karolinska University Hospital, Huddinge, and informed consent was obtained from each patient. Control Subjects A population-based randomly selected group of 28 control subjects (2 men; 71%) with a median age of 62 years (range, 37 to 79 years) were used for comparative analyses of biochemical and metabolic parameters. Control subjects were investigated according to a similar protocol as the patient group. The random selection of subjects in the Stockholm region was performed by Statistics Sweden. No other exclusion criteria than unwillingness to participate in the study were applied in the selection of the control group. Thus, 1 control subject was considered malnourished by means of subjective global assessment (SGA). Measurement Methods After an overnight fast, venous blood samples were drawn and stored at 8 C for biochemical analyses. GFR was estimated as the mean of creatinine and urea clearance calculated from 24-hour urinary samples collected from patients with CKD stage 5 or as iohexol clearance in patients with CKD stages 3 to 4, as well as in controls. Plasma analysis for insulin was performed on an Immulite system

3 918 (DPC Corp, Los Angeles, CA) by using commercially available assays (from DPC Corp), as were measurements of soluble intercellular adhesion molecule 1 (sicam-1) and soluble vascular cell adhesion molecule 1 (svcam-1; R&D Systems Europe Ltd, Abingdon, UK). Serum cholesterol and triacylglycerol levels were analyzed by means of standard enzymatic procedures (Roche Diagnostics GmbH, Mannheim, Germany). High-density lipoprotein (HDL) cholesterol level was determined after precipitation of apolipoprotein (apo) B containing lipoproteins by using phosphotungstic acid. Plasma resistin, leptin, and adiponectin levels were measured using a commercially available high-sensitivity photometric enzyme-linked immunosorbent assay (Linco Research, St Charles, MS), as was serum visfatin (Phoenix Pharmaceuticals Inc, Bellmont, CA). Plates were read by using an enzyme-linked immunosorbent assay VERSAmax reader (Molecular Devices Corp, Sunnyvale, CA), and data were analyzed using SoftmaxPRO software (Molecular Devices Corp). Levels of apo A1 (apo A), apo B, and high-sensitivity (hs) CRP were determined by using an immunonephelometric procedure (Behring AG, Marburg, Germany), whereas the remaining biochemical analyses were done using routine methods at the Department of Clinical Chemistry at Karolinska University Hospital at Huddinge. Nutritional status was recorded at the time of inclusion, concurrent with the drawing of blood samples, assessed by using SGA. 18,19 BMI was calculated as weight (kg)/height (m 2 ). In patients without diabetes, insulin resistance was calculated by using quantitative insulin-sensitivity check index (QUICKI: 1/[log (fasting plasma insulin [ U/mL])] log(fasting plasma glucose [mmol/l]), 2 as well as by using homeostasis model assessment for insulin resistance (fasting serum insulin [ U/mL] fasting plasma glucose [mmol/l]/ 22.5), 21 both recently validated in patients with end-stage renal disease. 22 Insulin resistance was not assessed in subjects with diabetes. Lean body mass and truncal fat mass were estimated by means of dual-energy x-ray absorptiometry using the DPX-L device (Lunar Corp, Madison, WI). With this technique, fat and lean body mass distributions are estimated directly without making assumptions about the 2-compartment model. 23 Because of the inability of dualenergy x-ray absorptiometry to specifically isolate visceral fat, we grouped fat readings as truncal (ie, all fat in the trunk, both visceral and subcutaneous) and nontruncal. Also, it must be kept in mind that although state of hydration does not affect the estimate of fat mass with dual-energy x-ray absorptiometry, it affects that of lean body mass. Determination of scd163 Concentration scd163 was measured by using an enzyme-linked immunosorbent assay as previously described. 24 Briefly, rabbit anti-scd163, 4 mg/l, was coated onto microtiter wells. After washing, 1 L of sample (diluted 1:5 in phosphatebuffered saline with albumin, ph 7.2) was added and incubated for 1 hour. Wells then were washed, and 1 L of monoclonal anti-scd163 (GHI/61; diluted 1:5) was added and incubated for 1 hour. After a second wash, 1 L of peroxidase-labeled antibody (goat antimouse immunoglobulins; Dako P447; diluted 1:4; Dako A/S, Glostrup, Denmark) was added and incubated for 1 hour. Wells were washed, and 1 L of a hydrogen peroxide/1,2-phenylenediamine dihydrochloride substrate solution was added. After 15 minutes, 5 L of 1 mol/l of sulphuric acid (H 2 SO 4 ) was added, and plates were read at 492/62 nm. Control samples and standards of purified scd163 were coanalyzed in each run. Statistical Analysis Results are expressed as mean SEM (normally distributed variables) or median and range (non-normal distribution) unless otherwise indicated, with P less than.5 indicating significance. Normality of distribution for each variable was assessed by using Shapiro-Wilk test. 25 Comparisons between groups were made using Wilcoxon (2 groups) or Kruskal-Wallis ( 2 groups) test. Comparisons between groups for nominal variables were made using chi-square test. Because many values were not normally distributed, correlations between variables were calculated according to Spearman rank. Partial nonparametric correlations using Spearman rank were used to adjust for scd163 level. Patient survival was calculated by using nonadjusted and adjusted Kaplan-Meier survival curves with the cohort divided into 2 groups according to a receiver operating characteristic (ROC) analysis of scd163 level as a predictor of all-cause mortality. Adjusted hazard ratios for survival were calculated using the Cox Proportional Hazards model, after first checking for proportionality. Because of the high mortality rate, analysis was limited to 6 months, at which time more than 5% of patients had died. All analyses were performed using statistical software SAS, version 8.2 (SAS Institute, Cary, NC). RESULTS AXELSSON ET AL Clinical Characteristics and Differences Between Groups Clinical characteristics of the 3 groups are listed in Table 1. In the entire material, scd163 level correlated with age (.18; P.5), but there were no correlations between scd163 level and sex or GFR in the entire material or when analyzed separately in the 3 groups. ROC Analysis Using ROC analysis (Fig 1) of scd163 levels in relation to all-cause mortality, the cutoff value was 4. mg/l with an area under the curve of.6.5 (95% confidence interval [CI],.51 to.69), sensitivity was 64%, and specificity was.54%. Correlations With Body Composition and Blood Lipid Levels Relevant correlations between scd163 levels and body composition and blood lipid levels in patients with CKD are listed in Table 2. scd163 levels correlated with BMI (.19; P.5) and total (.16; P.5) and truncal (

4 FAT MASS IS ASSOCIATED WITH MACROPHAGE NUMBERS IN CKD 919 Table 1. Basal Clinical Characteristics and Nutritional Markers in Patients and Controls Controls (n 28) CKD Stages 3-4 (n 38) CKD Stage 5 (n 12) Significance* (P) Age (y) Male sex (%) NS Diabetes mellitus type 1/2 (%) / 1/ /14.1 CVD (%) Inflamed (%; CRP 5 mg/l) Malnourished (%; SGA 1) GFR (ml/min) Blood hemoglobin (g/dl) Markers of glucose homeostasis in subjects without diabetes and blood lipids Fasting plasma glucose (mg/dl) NS Fasting plasma insulin ( IU/mL) Homeostasis model assessment index QUICKI Blood hemoglobin A 1c (%) Serum triglycerides (mg/dl) Serum cholesterol (mg/dl) 193 ( ) 217 (89-29) 213 (14-34) NS Serum HDL cholesterol (mg/dl) 54.1 ( ) 46.4 ( ) 46.4 ( ).1 Inflammation biomarkers and cell adhesion molecules Serum albumin (g/dl) Serum IL-6 (pg/ml) 2. (.4-1.) 3.2 (.6-9.7) 5.3 ( ).1 Serum hs-crp (mg/l) 1.2 (.2-9.1) 1.8 (.2-49.) 4.4 (.2-89.).1 Serum TNF- (pg/ml) 3.8 ( ) 8.8 ( ) 9.5 ( ).1 Serum svcam-1 (ng/ml) 747 (42-1,62) 951 (457-1,719) 1296 (492-2,933).1 Serum sicam-1 (ng/ml) 232 (19-324) 271 ( ) 247 (99-551) NS Adipokines and adipose tissue cytokines Serum leptin (ng/ml) 8.5 ( ) 17.2 ( ) 1.8 ( ).1 Serum adiponectin ( g/ml) 12.2 ( ) 28.9 ( ) 2.2 ( ).1 Serum resistin (ng/ml) 8.3 ( ) 18.7 ( ) 32.4 ( ).1 Serum fibrinogen (g/l) Serum visfatin (ng/ml) 27. ( ) 32.4 ( ) 31.2 ( ) NS scd163 (mg/l) 2.6 (.8-7.6) 3.6 ( ) 4.3 ( ).1 Body composition BMI (kg/m 2 ) NS Lean body mass (kg) Total fat mass (kg) NS Truncal fat mass (kg) NOTE. To convert GFR in ml/min to ml/s, multiply by.1667; hemoglobin and albumin in g/dl to g/l, multiply by 1; glucose in mg/dl to mmol/l, multiply by.5551; serum triglycerides in mg/dl to mmol/l, multiply by.1129; serum cholesterol and HDL cholesterol in mg/dl to mmol/l, multiply by Abbreviation: NS, not significant. *Kruskal-Wallis test. N 26/4/17. Only in patients without diabetes (n 38/27/93)..18; P.5) fat mass (Fig 2). There were no correlations between scd163 and blood lipid levels. Correlations With Inflammation, Adhesion Molecules, and Adipokines Table 2 also lists correlations of scd163 levels with levels of markers of inflammation, adhesion molecules, and circulating adipokines in patients with CKD. Notably, scd163 levels correlated with levels of all markers of inflammation, as well as with ICAM-1 (.27; P.1) and VCAM-1 levels (.34; P.1). Figure 3 shows significant differences in adhesion molecule levels between individuals with scd163 levels less than or greater than 4. mg/l. Furthermore, scd163 levels correlated with serum levels of fibrinogen

5 92 Sensitivity 1,,75,5,25,,,25,5,75 1, 1-Specificity Fig 1. ROC curve of scd163 levels in relation to patient all-cause mortality used for grouping. (.32; P.1), leptin (.17; P.1), and resistin (.23; P.1), but not visfatin or adiponectin levels. Correlations With Glucose Homeostasis Whereas there were no differences in mean age or scd163 level between patients with and without diabetes, patients with diabetes had a significantly greater total ( versus kg; P.5) and truncal ( versus kg; P.5) fat mass. Patients with diabetes and CKD also had greater median circulating levels of tumor necrosis factor (TNF- ; 9.1 pg/ml; range, 3.1 to 26.5 pg/ml versus 8.1 pg/ml; range, 1. to 49.2 pg/ml; P.5) and tended to have greater median IL-6 (4.5 pg/ml; range, 1.2 to 15.6 pg/ml versus 3.5 pg/ml; range,.4 to 31.2 pg/ml; P.6) and median visfatin levels (3.1 ng/ml; range, 9.5 to 71.2 ng/ml versus 27.4 ng/ml; range, 11. to 73.4 ng/ml; P.9). There were no significant differences between patients with and without diabetes in any other measured cytokines or adipokines, including scd163, and scd163 level did not correlate with insulin resistance expressed as QUICKI in patients without diabetes (Table 3). Longitudinal Data Forty-nine patients initiated therapy during the study period and had both baseline and 1-year data. Although there were no baseline differences in age, body composition, or scd163 levels, hemodialysis patients remained stable in body and truncal fat mass during the 12-month follow-up period, whereas peritoneal dialysis patients increased AXELSSON ET AL significantly in both body (2.6 kg; CI, 1.5 to 3.6; P.1) and truncal (1.2 kg; CI, 1.1 to 1.4; P.1) fat mass. Furthermore, in matched paired analysis, this change in fat mass was accompanied by an increase in circulating scd163 levels (3.1 mg/l; CI, 1.4 to 4.8; P.1) not observed in hemodialysis patients (Fig 4). However, no significant correlation between change in scd163 levels and either change in total or truncal fat mass was observed in any group. Unadjusted and Adjusted Survival Figure 5 shows nonadjusted and adjusted survival curves for patients divided according to ROC results. Briefly, scd163 levels also predicted outcome after adjustment for age, sex, and diabetes mellitus (chi-square 19.98; P.1). However, if clinical CVD or inflammation was included in the model, the significant difference between groups was lost. Table 2. Spearman Rank Correlations Between Baseline Levels of Serum scd163 and Relevant Markers of Body Composition and Inflammation in 158 Patients With CKD Clinical markers Age (y).21* GFR (ml/min).7 QUICKI.9 Serum triglycerides (mg/dl).11 Serum HDL cholesterol (mg/dl).16 Blood hemoglobin (g/dl).5 Inflammatory biomarkers and adhesion molecules Serum hs-crp (mg/l).28 Serum IL-6 (pg/ml).31 Serum TNF- (pg/ml).24* Serum svcam-1 (ng/ml).34 Serum ICAM-1 (ng/ml).27 Body composition and adipokines BMI (kg/m 2 ).19 Lean body mass (kg).3 Total fat mass (kg).16 Truncal fat mass (kg).18 Serum leptin (ng/ml).17* Serum resistin (ng/ml).23* Serum visfatin (ng/ml).17 Serum fibrinogen (g/l).32* NOTE. Bold type indicates significant correlation. *P.1. N 26/38/17. Only in patients without diabetes (n 38/27/93). P.1. P.5.

6 FAT MASS IS ASSOCIATED WITH MACROPHAGE NUMBERS IN CKD 921 A Body fat mass (kg) B Truncal fat mass (kg) rho=.17 p< Ln CD163 (mg/ L) rho=.17 p< Ln CD163 (mg/ L) Fig 2. Significant associations between serum levels of scd163 (mg/l) and (A) total-body fat mass and (B) truncal fat mass, measured by using dual-energy x-ray absorptiometry in 12 patients with CKD stage 5 ( males, Πfemales) and 38 patients with CKD stages 3 to 4 ( ) patients, as well as in 28 population controls ( ). Multivariate Analysis Table 3 lists effects of nonparametric adjustment for scd163 level on Spearman rank correlations between truncal fat mass and selected adipokine levels. Briefly, after adjustment for scd163 level, the previously significant relationship between truncal fat mass and levels of the proinflammatory cytokines IL-6 and hs-crp disappears. DISCUSSION In the present study, we investigate correlations between body composition and circulating levels of a number of adipokines and cytokines with scd163, a specific marker of monocytes/ macrophages, 26 both cross-sectionally and longitudinally. Although it increasingly is recognized that even mild renal function impairment is an independent cardiovascular risk factor, 2 the underlying causative mechanisms remain unknown. However, CKD shares multiple phenotypes with metabolic syndrome of obesity, 27 a constellation of risk factors also associated with a high risk for CVD. Thus, patients with CKD have chronic low-grade inflammation, 1 marked peripheral insulin resistance with maintained hepatic insulin sensitivity, 28 dyslipidemia, 27 and accelerated atherosclerosis. 29 As in patients with metabolic syndrome, adipose tissue signaling is changed markedly in uremic patients. 4,27 We thus hypothesized that alterations in body composition caused by both uremia per se and renal replacement therapy could influence some or all components of the proposed uremic-metabolic syndrome. We previously showed that truncal (visceral) fat mass is associated with circulating levels of A svcam-1 (ng/ml) B ICAM-1 (ng/ml) scd163 <4 mg/l scd163 <4 mg/l p<.1 p<.1 scd163 >4 mg/l scd163 >4 mg/l Fig 3. Significant differences in circulating (A) svcam-1 and (B) sicam-1 levels in 186 individuals included in the study divided according to scd163 level less than or greater than 4. mg/l (cutoff point by ROC curve function).

7 922 AXELSSON ET AL Table 3. Partial Spearman Rank Correlations: Impact of Correction for scd163 Level on Associations Between Truncal Fat Mass and Selected Variables in 12 Patients With CKD Stage 5 Uncorrected P Corrected for scd163 P Association between truncal fat mass and selected clinical characteristics Age (y) QUICKI* Serum HDL (mg/dl).11 NS.9 NS S-GFR (ml/min).5 NS.6 NS Association between truncal fat mass and selected cytokines and adipokines Serum hs-crp (mg/l) NS Serum IL-6 (pg/ml) NS Serum TNF- (pg/ml).1 NS.4 NS Serum leptin (ng/ml) Serum adiponectin ( g/ml).11 NS.1 NS Serum resistin (ng/ml).15 NS.18 NS Serum visfatin (ng/ml).9 NS.16 NS Serum fibrinogen (g/l).12 NS.6 NS Abbreviation: NS, not significant. *Only in patients without diabetes. such cytokines as IL-6 in patients with CKD stage 5. 4 However, the mechanisms involved remain unclear. It may be that adipocytes can directly secrete proinflammatory cytokines along with the diverse family of adipokines. 3 However, current evidence favors an alternative hypothesis, 5,31 suggesting that activated macrophages resident in fat tissue under certain conditions release such proinflammatory cytokines as IL-6 and TNF- that, in turn, activate a systemic inflammatory response. 32 The present study supports the latter theory by showing that circulating scd163 levels correlate with fat mass, circulating levels of proinflammatory cytokines, and increased circulating levels of endothelial adhesion molecules (Table 2). Furthermore, after adjustment for scd163 level, the previously significant relationship between fat mass and levels A PD PD FBM, kg p<.1 CD-1 6 3, mg/ L p<.1 Baseline 12 mont hs Baseline 12 mont hs B 6 HD p=ns 25 HD p=ns FBM, kg Baseline 12 months CD-163, mg/ L Baseline 12 months Fig 4. Longitudinal changes during 12 months in body fat mass and scd163 levels in (A) 3 peritoneal dialysis (PD) and (B) 19 hemodialysis (HD) patients starting renal replacement therapy.

8 FAT MASS IS ASSOCIATED WITH MACROPHAGE NUMBERS IN CKD 923 Fig 5. (A) Unadjusted and (B) adjusted (for age, sex, and presence of diabetes mellitus) survival curves according to CD163 level greater than or less than 4. mg/l in 158 patients with CKD. of inflammatory biomarkers (such as IL-6 or CRP) disappears, whereas relationships between fat and leptin level, as well as between fat and insulin resistance, remain significant (Table 3). Our findings are present in the age-matched control group, patients with mild to moderate renal dysfunction, and patients with end-stage renal disease, for whom it is most dominant. Because it was hypothesized that local ischemia is an important promoter of macrophage activation and migration, 7 we speculate that the markedly decreased vascular reactivity 33 observed in patients with CKD may act together with plasma volume changes induced during hemodialysis to stimulate adipose tissue mediated migration of macrophages in the end-stage renal disease patient group. Further confirming the putative association between fat mass and macrophages and having clinical implications, we report that longitudinal changes in fat mass after initiation of dialysis therapy are associated with changes in scd163 levels and systemic inflammation (Fig 3). In contrast to hemodialysis, initiation of peritoneal dialysis often is accompanied by an increase in body fat stores, probably linked to energy absorption from the glucose-rich dialysis fluid (estimated at 4 to 13 kcal/kg/d). 34,35 This led us to investigate body composition changes during the first 12 months of therapy separately in peritoneal dialysis and hemodialysis patients, comparing these with changes in scd163 levels and inflammation (Fig 3). As expected, whereas hemodialysis patients lost body fat mass, peritoneal dialysis patients gained an average of 2.6 kg (CI, 1.4 to 3.8) of fat. Furthermore, this increase was associated with increased scd163 levels, thus strengthening the circumstantial evidence for a direct relationship between fat mass and resident macrophage activity and indicating the need for additional studies to elucidate whether fat mass accumulation in peritoneal dialysis patients is associated with increased inflammatory activity. In the present study, we also found a previously unreported association between increased levels of circulating scd163 and circulating adhesion molecules (suggesting cardiovascular damage; Fig 3), as well as increased risk for mortality in renal patients (Fig 5) with increased scd163 levels. One possible explanation for these phenomena is the earlier finding by Jung et al 36 that macrophages are the dominant source of resistin found in atherosclerotic lesions, where resistin is an important mediator of endothelial dysfunction and cardiovascular damage. In addition to the positive association with adhesion molecule levels, we also found scd163 level to correlate positively with circulating levels of both resistin and fibrinogen. Thus, when survival is adjusted for either the presence of clinical CVD or circulating levels of one of the inflammatory cytokines, the predictive value of scd163 level on

9 924 survival was lost. This finding appears contrary to published reports 8,37 of a positive association between BMI (taken to reflect increased fat mass) and survival in patients with CKD (sometimes termed reverse epidemiology). However, there are several likely explanations for this, most notably that adipose tissue signaling comprises multiple adipokines and cytokines with both pathogenic and protective properties. Also, BMI may not accurately reflect fat mass, especially in patients with end-stage renal disease, for whom fluctuating total-body water and markedly decreased muscle mass are important confounders. 38 Several weaknesses of the present study should be kept in mind when interpreting the data. First, this is a post hoc observational study, limiting the amount of data available for analysis. Thus, we have follow-up data for only a subset of patients, and these do not include biochemical measurements other than CD163. Also, because of the study design, only correlative associations can be presented, making it impossible to comment about the cause(s) of observed covariation. In conclusion, the present study shows that plasma levels of scd163 (a marker of monocytes/ macrophages) correlate with fat mass and systemic inflammation in both healthy individuals and patients with renal function impairment. Furthermore, an increase in fat mass after initiation of dialysis therapy appears to correlate with a concomitant increase in serum scd163 levels. This increase also is associated with increased risk for mortality, even after adjustment for confounders. Finally, because the significant relationship between truncal fat mass and IL-6 level is lost after adjustment for scd163 level, we hypothesize that a uremic-metabolic syndrome linked to aberrant adipose tissue signaling in patients with uremia may explain some of the metabolic and cardiovascular complications of CKD. ACKNOWLEDGMENT The authors express their appreciation for the excellent work done by their support staff: Kirsten Bank Petersen (scd163 measurements), Anna-Lena Blom (data handling), Ann Dreiman-Lif, RN (clinical study coordinator), Annika Nilsson, RN (clinical studies), and Anki Emmoth (clinical studies). AXELSSON ET AL REFERENCES 1. Stenvinkel P: Inflammation in end-stage renal disease A fire that burns within. Contrib Nephrol 149: , Go AS, Chertow GM, Fan D, et al: Chronic kidney disease and the risks of death, cardiovascular events, and hospitalization. N Engl J Med 351: , Stenvinkel P, Ketteler M, Johnson RJ, et al: IL-1, IL-6, and TNF-alpha: Central factors in the altered cytokine network of uremia The good, the bad, and the ugly. Kidney Int 67: , Axelsson J, Qureshi AR, Suliman ME, et al: Truncal fat mass as a contributor to inflammation in end-stage renal disease. Am J Clin Nutr 8: , Clement K, Viguerie N, Poitou C, et al: Weight loss regulates inflammation-related genes in white adipose tissue of obese subjects. FASEB J 18: , Wellen KE, Hotamisligil GS: Inflammation, stress, and diabetes. J Clin Invest 115: , Weisberg SP, McCann D, Desai M, et al: Obesity is associated with macrophage accumulation in adipose tissue. J Clin Invest 112: , Kalantar-Zadeh K, Kopple JD, Block G, et al: Association among SF36 quality of life measures and nutrition, hospitalization, and mortality in hemodialysis. J Am Soc Nephrol 12: , Beddhu S, Pappas LM, Ramkumar N, et al: Effects of body size and body composition on survival in hemodialysis patients. J Am Soc Nephrol 14: , Kalantar-Zadeh K, Kuwae N, Wu DY, et al: Associations of body fat and its changes over time with quality of life and prospective mortality in hemodialysis patients. Am J Clin Nutr 83:22-21, Chaignon M, Chen WT, Tarazi RC, et al: Blood pressure response to hemodialysis. Hypertension 3: , Girndt M, Ulrich C, Kaul H, et al: Uremia-associated immune defect: The IL-1-CRP axis. Kidney Int Suppl 84:S76- S79, Ishimura E, Okuno S, Kim M, et al: Increasing body fat mass in the first year of hemodialysis. J Am Soc Nephrol 12: , Fernstrom A, Hylander B, Moritz A, et al: Increase of intra-abdominal fat in patients treated with continuous ambulatory peritoneal dialysis. Perit Dial Int 18: , Droste A, Sorg C, Hogger P: Shedding of CD163, a novel regulatory mechanism for a member of the scavenger receptor cysteine-rich family. Biochem Biophys Res Commun 256:11-113, Stenvinkel P, Heimbürger O, Paultre F, et al: Strong association between malnutrition, inflammation, and atherosclerosis in chronic renal failure. Kidney Int 55: , National Kidney Foundation: K/DOQI Clinical Practice Guidelines for Chronic Kidney Disease: Evaluation, Classification, and Stratification. Am J Kidney Dis 39:S1- S266, 22 (suppl 1) 18. Qureshi AR, Alvestrand A, Danielsson A, et al: Factors predicting malnutrition in hemodialysis patients: A cross-sectional study. Kidney Int 53: , 1998

10 FAT MASS IS ASSOCIATED WITH MACROPHAGE NUMBERS IN CKD Detsky AS, McLaughlin JR, Baker JP, et al: What is subjective global assessment of nutritional status? JPEN J Parenter Enteral Nutr 11:8-13, Katz A, Nambi SS, Mather K, et al: Quantitative insulin sensitivity check index: A simple, accurate method for assessing insulin sensitivity in humans. J Clin Endocrinol Metab 85: , Matthews DR, Hosker JP, Rudenski AS, et al: Homeostasis model assessment: Insulin resistance and beta-cell function from fasting plasma glucose and insulin concentrations in man. Diabetologia 28: , Wimmer NJ, Cucchiara AJ, Pappachen B, Townsend RR: Correlation of Steady State Insulin Sensitivity (SI-SS) Indices with the Frequently Sampled Intravenous Glucose Tolerance Test (SI-FSIGTT) in Chronic Kidney Disease (CKD). Poster presented at American Society of Nephrology 38th Annual Meeting, F-PO811, Philadelphia, PA, November 8-3, Kerr PG, Strauss BJG, Atkins RC: Assessment of the nutritional state of dialysis patients. Blood Purif 14: , Moller HJ, Hald K, Moestrup SK: Characterization of an enzyme-linked immunosorbent assay for soluble CD163. Scand J Clin Lab Invest 62: , Shapiro SS, Wilk MB: An analysis of variance test for normality (complete samples). Biometrika 52: , Moestrup SK, Moller HJ: CD163: A regulated hemoglobin scavenger receptor with a role in the anti-inflammatory response. Ann Med 36: , Beddhu S, Kimmel PL, Ramkumar N, et al: Associations of metabolic syndrome with inflammation in CKD: Results from the Third National Health and Nutrition Examination Survey (NHANES III). Am J Kidney Dis 46: , DeFronzo RA, Alvestrand A, Smith D, et al: Insulin resistance in uremia. J Clin Invest 67: , Wang MC, Tsai WC, Chen JY, et al: Stepwise increase in arterial stiffness corresponding with the stages of chronic kidney disease. Am J Kidney Dis 45:494-51, Mohamed-Ali V, Flower L, Sethi J, et al: -Adrenergic regulation of IL-6 release from adipose tissue: In vivo and in vitro studies. J Clin Endocrinol Metab 86: , Fain JN, Madan AK, Hiler ML, et al: Comparison of the release of adipokines by adipose tissue, adipose tissue matrix, and adipocytes from visceral and subcutaneous abdominal adipose tissues of obese humans. Endocrinology 145: , Wellen KE, Hotamisligil GS: Obesity-induced inflammatory changes in adipose tissue. J Clin Invest 112: , Kaysen GA, Eiserich JP: The role of oxidative stress Altered lipoprotein structure and function and microinflammation on cardiovascular risk in patients with minor renal dysfunction. J Am Soc Nephrol 15: , Nordfors L, Heimburger O, Lonnqvist F, et al: Fat tissue accumulation during peritoneal dialysis is associated with a polymorphism in uncoupling protein 2. Kidney Int 57: , Bergstrom J, Furst P, Alvestrand A, et al: Protein and energy intake, nitrogen balance and nitrogen losses in patients treated with continuous ambulatory peritoneal dialysis. Kidney Int 44: , Jung HS, Park KH, Cho YM, et al: Resistin is secreted from macrophages in atheromas and promotes atherosclerosis. Cardiovasc Res 69:76-85, Kalantar-Zadeh K, Abbott KC, Salahudeen AK, et al: Survival advantages of obesity in dialysis patients. Am J Clin Nutr 81: , Rigalleau V, Lasseur C, Chauveau P, et al: Body composition in diabetic subjects with chronic kidney disease: Interest of bio-impedance analysis, and anthropometry. Ann Nutr Metab 48:49-413, 24