SEARCH

SEARCH BY CITATION

Keywords:

  • adipose tissue;
  • end-stage renal disease

Abstract

  1. Top of page
  2. Abstract
  3. Introduction
  4. Materials and methods
  5. Results
  6. Discussion
  7. Conflict of interest statement
  8. Acknowledgements
  9. References

Abstract.  Witasp A, Carrero JJ, Heimbürger O, Lindholm B, Hammarqvist F, Stenvinkel P, Nordfors L (Karolinska Institutet, Stockholm, Sweden). Increased expression of pro-inflammatory genes in abdominal subcutaneous fat in advanced chronic kidney disease patients. J Intern Med 2011; 269: 410–419.

Objectives.  Low-grade systemic inflammation, oxidative stress and peripheral insulin resistance are intimately associated and contribute to the increased risk of cardiovascular complications in advanced chronic kidney disease (CKD). Because altered adipose tissue activities have previously been linked to pathophysiological processes in various inflammatory and metabolic diseases we hypothesized that the uraemic milieu in patients with CKD may interact with the adipose tissue, provoking an unfavourable shift in its transcriptional output.

Design.  Twenty-one adipokine mRNAs were quantified in abdominal subcutaneous adipose tissue (SAT) biopsies and serum/plasma concentrations of inflammatory markers and related protein products were measured.

Setting.  The study was conducted at the Karolinska University Hospital, Huddinge, and Karolinska Institutet, Stockholm, Sweden.

Subjects.  Thirty-seven patients with CKD [15 women, median 58 (interquartile range 49–65) years] and nine nonuraemic individuals [four women, age 62 (45–64) years] were recruited prior to initiation of peritoneal dialysis catheter insertion or elective hernia repair/laparoscopic cholecystectomy, respectively.

Results.  Even after correction for body mass index, SAT from patients showed a significant upregulation of inflammatory pathway genes interleukin 6 (3.0-fold, P = 0.0002) and suppressor of cytokine signalling 3 (2.5-fold, P = 0.01), as well as downregulation of leptin (2.0-fold, P = 0.03) and the oxidative stress genes uncoupling protein 2 (1.5-fold, P = 0.03) and cytochrome b-245, alpha polypeptide (1.5-fold, P = 0.005), in relation to controls.

Conclusions.  These gene expression differences suggest that inflammatory and oxidative stress activities may be important features of the intrinsic properties of uraemic adipose tissue, which may have significant effects on the uraemic phenotype.


Abbreviations:
ADIPOQ

adiponectin

ADIPOR2

adiponectin receptor 2

BMI

body mass index

CKD

chronic kidney disease

CRP

C-reactive protein

Ct

threshold cycle

CVD

cardiovascular disease

CYBA

cytochrome b-245, alpha polypeptide

GFR

glomerular filtration rate

IKBKB

inhibitor of kappa light polypeptide gene enhancer in B-cells, kinase beta

IL6

interleukin 6

IL6R

interleukin 6 receptor

LEP

leptin

LEPR

leptin receptor

NAMPT

nicotinamide phosphoribosyltransferase

PD

peritoneal dialysis

PIK3R1

phosphoinositide-3-kinase, regulatory subunit 1 (p85 alpha)

ROS

reactive oxygen species

SAT

subcutaneous adipose tissue

SLC2A4

solute carrier family 2, facilitated glucose transporter member 4

SOCS3

suppressor of cytokine signalling 3

UCP2

uncoupling protein 2

VCAM1

vascular cell adhesion molecule 1

Introduction

  1. Top of page
  2. Abstract
  3. Introduction
  4. Materials and methods
  5. Results
  6. Discussion
  7. Conflict of interest statement
  8. Acknowledgements
  9. References

Patients with advanced chronic kidney disease (CKD) display a spectrum of clinical features including chronic low-grade inflammation, insulin resistance, dyslipidaemia and accelerated atherosclerosis [1–5]. As this range of features resembles that of the metabolic syndrome in the general population, it has been termed the ‘uraemic–metabolic syndrome’ [6]. This uraemic–metabolic syndrome is highly prevalent in patients with CKD [7–9] and contributes to an elevated cardiovascular risk [10]. Obesity, and specifically abdominal obesity, is thought to significantly contribute to the prevalence and severity of the metabolic syndrome in the general population [11] and to predict mortality and cardiovascular complications in patients with CKD [12]. However, there are conflicting data regarding the role of obesity in patients with advanced CKD in whom obesity has also been associated with a survival advantage [13, 14].

It is clear that obesity is not only characterized by impaired insulin sensitivity but also by enhanced inflammatory signalling in adipose tissue, an endocrine organ that is now recognized as a major site for integration of inflammatory and metabolic pathways [15]. The two main fat depots in the body are visceral adipose tissue (VAT) and subcutaneous adipose tissue (SAT) located intra-abdominally and peripherally, respectively. Although it has been suggested that these two compartments are biologically distinct [16, 17], the results of a recent gene expression study in obese individuals suggest that both SAT and VAT display pro-inflammatory and insulin resistance-associated profiles, which possibly contributes to metabolic complications [18]. As abdominal obesity in CKD is associated with systemic inflammation [1, 19, 20], it has been suggested that adipose tissue is an important source of inflammation in this patient group. Consistent with this hypothesis, it has been reported that C-reactive protein (CRP), fat mass and abdominal subcutaneous fat are primary determinants of insulin resistance in haemodialysis patients [21, 22]. However, as only clinical correlations have been demonstrated, the precise contribution of uraemic fat to systemic inflammation and insulin resistance has not been fully elucidated.

In this study, we hypothesized that uraemia causes alterations in the adipose tissue-specific transcription that may link fat mass to CKD complications, similar to observations in the metabolic syndrome. Thus, we performed mRNA quantification analyses in abdominal SAT from 37 consecutive stage 5 CKD (CKD-5) patients scheduled for peritoneal dialysis (PD) catheter insertion and nine nonuraemic individuals. We targeted molecules that are typically expressed in adipose tissue, as well as genes encoding inflammation, oxidative stress-related and insulin signalling pathway molecules.

Materials and methods

  1. Top of page
  2. Abstract
  3. Introduction
  4. Materials and methods
  5. Results
  6. Discussion
  7. Conflict of interest statement
  8. Acknowledgements
  9. References

The experiments were undertaken with the understanding and informed consent of all participants. The study protocol was approved by the Ethics Committee at the Karolinska Institutet.

Patients

Adipose tissue gene expression analyses were performed in a subset of patients with CKD-5 enrolled from a larger cohort that has previously been described in detail [4]. Patients with acute infection, active vasculitis or liver disease at the time of evaluation were excluded, as were patients younger than 18 or older than 70 years. Thirty-seven consecutive patients with CKD-5 [age 58 (49–65) years, 22 (59%) men and glomerular filtration rate (GFR) 6.8 (6.0–8.6) mL min−1] were recruited shortly before initiating PD at the Karolinska University Hospital at Huddinge, Stockholm, during the period 1994–2008, and tissue specimens were collected during PD catheter insertion. The causes of renal failure amongst the 37 patients were diabetic nephropathy (= 11; 30%), polycystic kidney disease (= 6; 16%), chronic glomerulonephritis (= 4; 11%), nephrosclerosis (= 3; 8%) and other or unknown diseases (= 13; 35%). The presence of clinically manifest cardiovascular disease (CVD) was described in 15 patients (41%) and diabetes mellitus type-1 (= 6) or type-2 (= 10) was reported in 16 (42%) patients (baseline data for diabetic and nondiabetic patients are presented in Table 1). Eighteen patients (49%) were receiving statin medication and 21 patients (57%) were treated with beta blockers.

Table 1. Clinical and laboratory data of patients with CKD-5 and controls
 CKD-5 patientsControls
All patientsNondiabetic patientsDiabetic patients
  1. Values are presented as median (25th–75th quartile). Group differences were assessed with nonparametric Wilcoxon rank-sum test.

  2. 1Differences between diabetic and nondiabetic patients.

  3. 2Differences between all patients and controls.

  4. *, **, ***P-values <0.05, <0.01 and <0.001, respectively.

  5. BMI, body mass index; CKD, chronic kidney disease; CVD, cardiovascular disease; GFR, glomerular filtration rate; hsCRP, high sensitivity C-reactive protein; IL6, interleukin 6; VCAM1, vascular cell adhesion molecule 1.

n3721169
Age (years)58 (49–65)54 (37–64)62 (55–68)62 (45–64)
Sex (% males)59526956
Diabetes (%)43
CVD (%)4127731,**
GFR (mL min−1)7 (6–9)6 (5–8)8 (7–9)1,*95 (75–100)2,***
BMI (kg m−2)23.7 (21.8–25.9)23.7 (21.1–25.8)24.1 (22.4–26.4)28.1 (26.9–29.4)2,***
Serum albumin (g L−1)34.0 (32.3–36.0)34.5 (33.0–36.0)34.0 (30.3–37.0)37.0 (33.0–38.5)
Total cholesterol (mmol L−1)4.3 (3.6–5.2)4.3 (3.2–4.8)4.4 (3.8–5.2)4.7 (4.5–5.7)
Triglycerides (mmol L−1)1.6 (1.0–2.3)1.8 (1.0–2.7)1.6 (1.0–2.5)1.5 (1.2–1.7)
Plasma glucose (mmol L−1)5.0 (4.7–6.6)4.9 (4.7–5.2)5.8 (4.5–10.3)5.4 (5.0–6.1)
HbA1c%4.6 (4.3–5.1)4.5 (4.2–4.8)5.2 (4.6–7.4)1,**4.5 (4.3–4.8)
hsCRP (mg L−1)4.2 (1.5–10.0)2.7 (1.6–10.0)6.5 (1.2–11.2)3.5 (2.3–8.5)
IL6 (ng mL−1)6.9 (4.3–10.7)6.7 (4.0–10.0)7.0 (4.7–11.3)2.7 (2.3–4.3)2,***
VCAM1 (ng mL−1)1050 (877–1230)1017 (845–1170)1087 (955–1243)593 (526–640)2,***

Nonuraemic control subjects

The control subjects consisted of patients scheduled for surgical treatment of noninflammatory conditions (cholelithiasis and hernia) at the Karolinska University Hospital. Nine consecutive patients (cholelithiasis = 3 and hernia = 6; 5 [56%] men, age 62 [45–64] years) were recruited prior to the surgical procedure. The exclusion criteria were signs of preoperative systemic inflammation, clinically evident CVD or diabetes mellitus.

Blood and tissue sampling

Approximately 1 g SAT was sampled from the lower abdominal region at the beginning of surgery, immediately snap-frozen in isopentane and stored at −70 °C until required for further analysis. A fasting morning blood sample was drawn preoperatively on the day of surgery from all patients and control subjects for biochemical measurements. Plasma/serum samples were analysed immediately or stored at −80 °C until required for further analyses.

Clinical and laboratory assessments

Concentrations of creatinine, urea, albumin, high-sensitivity CRP, glucose, cholesterol and triglycerides were analysed in plasma or serum, as appropriate, using routine methods at the Department of Clinical Chemistry, Karolinska University Hospital. Serum levels of interleukin 6 (IL6) were quantified using an Immulite Automatic Analyzer (Diagnostic Products Corporation, Los Angeles, CA, USA). Serum-soluble vascular cell adhesion molecule 1 (VCAM1) was analysed using commercially available kits (R&D Systems, Inc., Minneapolis, MN, USA).

Glomerular filtration rate was assessed as the mean of urea- and creatinine-clearance values corrected for body surface area in patients with CKD. In controls, GFR was estimated from the plasma level of cystatin C. Body mass index (BMI) was calculated by dividing the weight in kilograms by the square of the height in metres.

RNA extraction and mRNA measurements

Samples of approximately 80-mg frozen tissue were homogenized and total RNA was isolated using RNeasy Lipid Tissue Mini Kit or Fibrous Tissue Mini Kit (QIAGEN Sciences, Germantown, MD, USA). The RNA concentration and the integrity of each sample were evaluated on a NanoDrop ND-1000 Spectrophotometer (NanoDrop Technologies Inc., Wilmington, DE, USA) and Agilent 2100 Bioanalyzer (Agilent Technologies Inc., Santa Clara, CA, USA), respectively. Random hexamers and the SuperScript™ III First-Strand Synthesis System for RT-PCR (Invitrogen, Carlsbad, CA, USA) were used for reverse transcription of total RNA into cDNA. mRNA quantities of 21 target genes and three endogenous control genes (Table 2) were measured simultaneously in each sample (run in duplicates) by real-time polymerase chain reaction using ABI TaqMan® Low Density Custom Array (Applied Biosystems, Foster City, CA, USA) and TaqMan® Universal Master Mix No AmpErase® UNG on an ABI PRISM 7900 HT Sequence Detection System (Applied Biosystems). Real-time measurements were recorded during 40 amplification cycles, and data were obtained as threshold cycle (Ct) values. Low-abundance target genes were excluded from the study if Ct was greater than 40 in more than 20% of the samples. Data were corrected for sample to sample variation in RNA quality and reverse transcription reaction efficiency by normalizing the target mRNA quantity with regard to both 18S rRNA and beta-2-microglobulin mRNA quantities. The relative gene expression, presented as arbitrary units, was calculated according to a modified delta-Ct method [23]. Glyceraldehyde-3-phosphatase dehydrogenase was excluded from the analyses because of its high variability between samples.

Table 2. Relative gene expression measurements of selected adipokines and other candidates implicated in energy homeostasis, inflammation, insulin/glucose signalling and oxidative metabolism in subcutaneous adipose tissue of nonuraemic controls and patients with CKD-5
GeneCategoryControls (= 9)CKD-5 patients (n = 37)Fold differenceP-value
  1. Relative mRNA quantity values (arbitrary units) are presented as median (25th–75th quartile) for patient and control groups. Differences between groups were calculated as fold differences (mRNA quantities measured in patients relative to mRNA quantities measured in controls) and evaluated by nonparametric Wilcoxon rank-sum test.

  2. CKD, chronic kidney disease.

AdiponectinADIPOQAdipokine1.44 (1.08–2.38)1.55 (1.23–2.14)1.070.8
Adiponectin receptor 1ADIPOR1Adipokine1.53 (1.25–2.15)1.50 (1.22–2.32)−1.020.9
Adiponectin receptor 2ADIPOR2Adipokine1.30 (0.92–1.42)1.36 (1.03–1.71)1.050.3
LeptinLEPAdipokine1.10 (0.78–1.42)0.48 (0.29–0.92)−2.290.01
Leptin receptorLEPRAdipokine1.02 (0.95–1.42)0.81 (0.64–1.09)−1.260.009
Uncoupling protein 2UCP2Energy homeostasis0.87 (0.73–1.00)0.57 (0.45–0.67)−1.530.0002
CD68 moleculeCD68Inflammation0.75 (0.46–1.16)0.52 (0.33–0.99)−1.410.09
Tumour necrosis factorTNFInflammation0.61 (0.33–1.01)0.80 (0.46–1.59)1.310.2
Tumour necrosis factor receptor superfamily, member 1ATNFRSF1AInflammation0.48 (0.21–0.78)0.35 (0.01–0.51)−1.370.1
Inhibitor of kappa light polypeptide gene enhancer in B-cells, kinase betaIKBKBInflammation1.68 (1.43–2.18)1.65 (1.22–1.93)−1.020.6
Nuclear factor of kappa light polypeptide gene enhancer in B-cellsNFKB1Inflammation1.32 (1.10–1.85)1.01 (0.89–1.41)−1.320.06
Interleukin 6IL6Inflammation1.23 (0.70–1.62)4.00 (1.82–6.73)3.250.02
Interleukin 6 receptorIL6RInflammation1.25 (1.04–1.84)1.01 (0.65–1.31)−1.240.04
Suppressor of cytokine signalling 3SOCS3Inflammation1.17 (0.68–1.12)2.20 (1.27–3.36)1.880.03
Signal transducer and activator of transcription 3STAT3Inflammation1.05 (0.93–1.24)0.95 (0.79–1.08)−1.100.1
Nicotinamide phosphoribosyltransferaseNAMPTInflammation2.19 (1.22–4.07)2.47 (1.72–3.29)1.130.6
Insulin receptor substrate 1IRS1Insulin/glucose signalling1.66 (1.03–2.05)1.58 (0.97–2.70)−1.050.8
Phosphoinositide-3-kinase, regulatory subunit 1 (p85 alpha)PIK3R1Insulin/glucose signalling2.20 (1.35–4.03)1.51 (1.23–2.12)−1.470.1
Solute carrier family 2, facilitated glucose transporter member 4SLC2A4Insulin/glucose signalling0.83 (0.59–1.44)1.40 (0.79–2.02)1.690.08
Solute carrier family 2, facilitated glucose transporter member 1SLC2A1Insulin/glucose signalling1.04 (0.74–1.49)1.06 (0.75–1.79)1.010.6
Cytochrome b-245, alpha polypeptideCYBAOxidative stress1.27 (0.97–1.63)0.65 (0.45–1.03)−1.950.001

Statistical analysis

Statistical analyses were performed using jmp in version 7.0.2 and sas version 9.1.3 (SAS Institute Inc., Cary, NC, USA). The descriptive data are expressed as median (25th–75th percentile) values. Nonparametric Wilcoxon rank-sum test or Kruskal–Wallis test was performed to evaluate differences between groups, and Spearman’s rank correlation was used to assess univariate correlations between variables. A general linear model was used to adjust for the presence of CVD and diabetes when comparing gene expression between groups. A P-value less than 0.05 was considered statistically significant. Statistical differences were not corrected for by multiple testing, and results should be considered as exploratory.

Results

  1. Top of page
  2. Abstract
  3. Introduction
  4. Materials and methods
  5. Results
  6. Discussion
  7. Conflict of interest statement
  8. Acknowledgements
  9. References

Baseline status of patients with CKD and controls

Clinical and biochemical data from patients with CKD and control subjects are shown in Table 1. Patients with CKD and controls had similar age and gender distributions but patients with CKD had lower BMI values and higher concentrations of the inflammatory surrogate markers VCAM1 and IL6 than controls. In addition, clinical data for diabetic and nondiabetic patients are presented in Table 1. A history of CVD events was more common amongst diabetic than nondiabetic patients (P = 0.006) and the patients with diabetics showed higher GFR (P = 0.03) and HbA1c (P = 0.0023) values.

Inflammatory pathway genes and relation to inflammatory markers

Estimated relative mRNA quantities are presented in Table 2. Compared to SAT from controls, SAT from patients displayed about threefold and twofold higher mRNA concentrations of IL6 (P = 0.025) and suppressor of cytokine signalling 3 (SOCS3) (P = 0.039), respectively, but slightly reduced IL-6 receptor (IL6R) mRNA concentrations (P = 0.049). SAT mRNA concentrations of IL6 and IL6R did not correlate with each other and were not correlated with circulating plasma concentrations of IL6 or high-sensitivity CRP. However, IL6 and SOCS3 mRNA levels were positively correlated (ρ = 0.54, P = 0.020). There were no differences between patients with CKD and controls in the expression of the remaining inflammatory pathway genes studied.

Oxidative stress and adipokine genes

Cytochrome b-245, alpha polypeptide (CYBA) mRNA levels were decreased twofold (P = 0.0014) in patients compared with controls. Also, significantly lower levels of mRNAs encoding the proteins uncoupling protein 2 (UCP2) (−1.5-fold), leptin (LEP) (−2.3-fold) and leptin receptor (LEPR) (−1.3-fold) were found in patients compared with controls. There was a positive correlation between LEP and LEPR mRNA quantities (ρ = 0.49, P = 0.0027), but only LEP mRNA correlated with circulating concentrations of LEP (ρ = 0.74, P < 0.001). mRNA quantities of the remaining adipokine genes studied were not different between patients and controls.

Insulin signalling and glucose homeostasis pathway genes

None of the studied genes involved in insulin signalling and glucose homeostasis pathways were differentially expressed between patients with CKD and controls.

The impact of BMI on mRNA levels

Because patients with CKD had significantly lower BMI values than the control group, SAT gene expression was normalized for BMI. Under these conditions, the different mRNA profiles observed for IL6, SOCS3, CYBA, LEP and UCP2 remained statistically significant (Fig. 1). In the CKD-5 patient group, BMI was positively correlated with CYBA (ρ = 0.33, P = 0.047) and LEP (ρ = 0.47, P = 0.004) mRNA levels.

image

Figure 1. mRNA levels (relative expression, arbitrary units) of five genes A. IL6, B. SOCS3, C. CYBA, D. UCP2 and E. LEP remained significantly different between patients with CKD-5 (n = 37) and nonuraemic controls (n = 9) after normalization with regard to BMI. mRNA/BMI ratio shown as the median and quartile. *, **, ***P-values <0.05, <0.01 and <0.001, respectively, by Wilcoxon rank-sum test. BMI, body mass index; IL6, interleukin 6; SOCS3, suppressor of cytokine signalling 3; CYBA, cytochrome b-245, alpha polypeptide; UCP2, uncoupling protein 2; LEP, leptin.

Download figure to PowerPoint

Association between baseline characteristics and gene expression

Age was negatively correlated with nicotinamide phosphoribosyltransferase (NAMPT) (ρ = −0.40, P = 0.01) and solute carrier family 2, facilitated glucose transporter member 4 (SLC2A4) (ρ = −0.37, P = 0.02) but positively correlated with CD68 (ρ = 0.50, P = 0.002) and CYBA (ρ = 0.38, P = 0.02) mRNA quantities. Despite no baseline clinical differences amongst the sexes, mRNA levels of six genes were elevated in women compared to men: adiponectin (ADIPOQ; 1.36 vs. 1.93, P = 0.0094), adiponectin receptor 2 (ADIPOR2; 1.09 vs. 1.58, P = 0.0033), inhibitor of kappa light polypeptide gene enhancer in B-cells, kinase beta (IKBKB; 1.57 vs. 1.84, P = 0.0045), solute carrier family 2, facilitated glucose transporter member 1 (SLC2A1; 0.87 vs. 1.44, P = 0.0424), SLC2A4 (0.89 vs. 2.07, P = 0.0030) and phosphoinositide-3-kinase, regulatory subunit 1 p85 alpha (PIK3R1) (1.32 vs. 1.59, P = 0.0257).

To investigate whether the observed significant gene expression differences between groups were because of uraemia per se rather than to diabetes or CVD, we used a general linear model to adjust for the presence of these co-morbidities. After adjustments, UCP2 (P = 0.0002) and CYBA (P = 0.0008) mRNA expression still remained significant between groups, whereas IL6 and LEP mRNA expression differences became almost significant (P = 0.07 for both). The difference in SOCS3 mRNA levels was completely abolished after adjustment for co-morbidities (P = 0.4).

Discussion

  1. Top of page
  2. Abstract
  3. Introduction
  4. Materials and methods
  5. Results
  6. Discussion
  7. Conflict of interest statement
  8. Acknowledgements
  9. References

In this study, we focused on 21 predetermined candidate genes, primarily adipokines and genes with relevance to glucose homeostasis, insulin signalling/sensitivity, oxidative stress and inflammation, which were measured at the mRNA level in abdominal SAT patients with of CKD-5 and controls. Seven of the investigated genes displayed a significantly altered expression profile in uraemic tissues compared with nonuraemic tissues, and the major finding was the significantly elevated mRNA concentrations of the inflammatory pathway signalling genes IL6, IL6R and SOCS3 in patients with CKD-5. Given the strong impact of inflammation and insulin resistance on outcome and CKD complications [24], and the intriguing link between these two conditions [25], our findings maybe improve the understanding of the uraemic phenotype.

In patients with CKD, systemic levels of IL6 are frequently elevated and associated with increased mortality [26–28]. It has been estimated that adipose tissue contributes to about 15–35% of the total circulating IL6 concentration [29]. In accordance, we [19, 20] and others [30–32] have reported an association between biomarkers of inflammation and fat mass in patients with CKD. IL6 production in adipocytes may be stimulated by inflammatory cytokines, such as tumour necrosis factor (TNF) and also by IL6 itself [33], or via stress events including cellular, metabolic or inflammatory stress [34]. It is therefore plausible that the metabolic and/or inflammatory stress, as imposed by the uraemic imbalance, stimulates an increase in IL6 transcription. Accordingly, the patients with CKD-5 in the current study displayed a median serum IL6 concentration about 2.5-fold higher than the control subjects and, in the same direction, an up-regulated IL6 mRNA production in SAT. However, the lack of correlation between circulating IL6 levels and SAT mRNA levels implies that elevated IL6 in the context of uraemia may be caused by retention rather than increased production. Nonetheless, the fact that the increase in IL6 mRNA was accompanied by elevated IL6R and SOCS3 mRNA levels indicates an augmented role for IL6 signalling in the uraemia-exposed adipose tissue. Measurements of local protein concentrations are required to determine whether this potentially increased IL6 effect is because of elevated autocrine or endocrine signalling.

We analysed two oxidative stress-related genes, UCP2 and CYBA, in our targeted gene expression profiling study. The UCP2 gene encodes a protonophore of the mitochondrial inner membrane that is proposed to contribute to decreased mitochondrial reactive oxygen species (ROS) production in cells [35] and to influence the cellular response to glucose [36, 37]. By contrast, the CYBA gene encodes the p22phox subunit of the mitochondrial nicotinamide adenine dinucleotide phosphate (NADPH)-oxidase complex. This complex is a major source of ROS in various cells and hence is implicated in processes leading to enhanced oxidative stress [38, 39]. Our finding of reduced patient mRNA levels of UCP2 and CYBA in the present study may suggest that a distorted balance between these genes could contribute to oxidative stress within the uraemic adipose tissue. This is important because oxidative stress is considered to be a significant risk factor for vascular disease [40], which in turn is implicated in the increased cardiovascular morbidity and mortality in patients with CKD [41].

Increased oxidative stress, through elevated ROS generation, causes insulin resistance by reducing insulin-stimulated glucose transport in 3T3-L1 adipocytes [42]. However, critical components in the downstream insulin signalling pathway did not show altered mRNA levels in patients with CKD-5 compared to controls in the present study. Based on these findings, we suggest two possibilities. First, altered transcription of the investigated genes may not be the mechanism by which insulin sensitivity is impaired in adipose tissue in CKD-5. Second, the toxic uraemic milieu in patients with CKD-5 may influence the expression of other insulin signalling pathway genes. It is also possible that insulin sensitivity is affected via post-transcriptional and post-translational modifications of the insulin signalling pathway mediators. For example, it was recently shown that urea is able to induce ROS generation, which in turn causes insulin resistance by altering phosphorylation of the insulin signalling molecules insulin receptor substrate 1 (IRS1) and v-akt murine thymoma viral oncogene homolog (AKT) [42].

Circulating levels of various adipokines increase concomitantly with a decrease in GFR, which potentially could aggravate their biological effects [43]. It is therefore of interest to understand the cause of the increased circulating adipokines. It has been found that LEP and adiponectin are downregulated at the mRNA level in adipose tissue, despite elevated systemic levels, and therefore, a negative feedback mechanism has been suggested [44, 45]. Accordingly, we found decreased LEP mRNA levels in patients when compared to controls, even after correction for BMI, possibly reflecting the importance of uraemia rather than BMI for LEP transcription. However, as circulating LEP concentrations and LEP mRNA levels were positively correlated in patients, this finding contradicts the suggested negative feedback regulation on LEP gene expression. It may be speculated that the decreased transcription is because of retention of uraemic toxins in patients with CKD-5 rather than by inhibition by LEP per se. To increase our understanding of these complex associations further, we need to better understand whether LEP resistance is a prominent feature of uraemia.

Another possible explanation is that the changes observed in abdominal SAT may not be reflected in those observed in other adipose tissue sites. Indeed, previous studies have reported different gene expression patterns of SAT and VAT [16, 17], including insulin receptor [46] and adipokine expression [47, 48], suggesting site-specific characteristics of different fat depots. Nonetheless, at least in obese individuals, it was recently demonstrated that SAT and VAT display analogous pro-inflammatory and insulin resistance-associated gene expression signatures [18], indicating that these tissues may contribute to similar phenotypes. Moreover, as the body’s total volume of SAT may be threefold to fourfold greater than that of VAT, the total secretory role of SAT and VAT may be comparable [49]. However, as the present study protocol only included analyses of abdominal SAT, this issue remains to be investigated.

Few studies have focused on cytokine and adipokine production at their site of action, which would be of value as adipokines and cytokines tend to act locally via autocrine or paracrine signalling. Gene expression studies of specific organs that are hypothetically linked to a disease, such as in the present study, may reveal valuable clues to the disease pathogenesis. Indeed, abdominal SAT expression of inflammatory cytokines was previously proposed as a potential mechanism linking obesity with its metabolic co-morbidities [50]. In addition, a recent large population-based cohort study of adipose tissue gene expression, covering approximately 85% of protein-coding genes, reported a causal association between inflammatory/immune response genes and obesity-related traits [51]. Adding to these findings, this study shows that there is a possible transcriptional dysregulation of primarily inflammatory and oxidative stress-related molecules in uraemic adipose tissue that may contribute to CKD complications. By contrast, recent data on human adipose tissue suggest that inflammation could in fact be beneficial for the individual, acting as a regulator of the normal physiological integrity, nutrition maintenance and function of adipose tissue [52]. Hence, in the context of uraemia, it may be speculated that adipose tissue inflammation could be raised as a defence mechanism against metabolic aberrations, such as protein–energy wasting and anorexia, which are highly prevalent features of CKD-5 [4]. In this context, it is important to note that, in CKD, diabetes and CVD are common co-morbidities [1], which are associated with pro-inflammatory events [53]. However, adjusting for these co-morbidities did not strongly influence the observed associations and, thus, it may be suggested that the uraemia per se primarily accounts for the observed gene expression differences, rather than background co-morbidities. On the other hand, it can be argued that co-morbidities such as diabetes and CVD are closely associated with uraemia, acting both as a cause and a consequence of CKD. It should also be noted that sensitivity analyses in small sample sets, such as the present one, should be interpreted with caution because of the risk of overadjustment.

There are some limitations that need to be considered in the present study. These include the relatively small number of patients with CKD-5 and control subjects and the fact that the patient and control groups were not matched for BMI. It should be considered that the higher BMI values of the controls may reflect an increased fat mass, and potentially a larger number of infiltrated immune cells, such as macrophages, in the adipose tissue, which may contribute to an augmented inflammatory [54] and oxidative stress activity [55]. However, although it was recently shown that women with advanced CKD have an increased SAT mRNA expression of the immune cell marker CD68 and infiltration of CD68-immunopositive cells compared to nonuraemic women [31], we did not observe any difference in CD68 expression between patients and controls. In agreement, TNF, being mainly expressed by macrophages [54], was not differently expressed in controls compared to patients in the current study. Clearly, more studies are needed to elucidate the role of macrophages in uraemic adipose tissue and their potential contribution of inflammatory or oxidative stress-related molecules. Further study limitations to take into account include the possibility that an equivalent relationship between mRNA and protein product may not always be expected and the fact that we only performed systemic, not local, measurements of cytokine concentrations. Finally, results in the present study are based upon observational analyses with no consensus regarding dependencies amongst the data and we have therefore chosen not to adjust for multiple testing [56]. Ideally, future confirmatory studies in independent samples of uraemic SAT should be performed to verify the hypotheses and significant findings.

In conclusion, patients with CKD-5 display an altered expression profile of inflammatory and oxidative stress genes in abdominal SAT compared with controls. This implies that adipose tissue in CKD may, in combination with impaired renal clearance, contribute to adipokine dysregulation, systemic inflammation and oxidative stress. Adding to the existing data demonstrating a close association between systemic inflammation, oxidative stress and poor outcome in CKD [24], our findings suggest that adipose tissue-specific events may contribute to the pathological changes observed in the uraemic state, which aggravate CKD complications and negatively influence outcome. Future studies are needed to address the interacting effects of the uraemic state and the intrinsic properties of uraemic fat.

Conflict of interest statement

  1. Top of page
  2. Abstract
  3. Introduction
  4. Materials and methods
  5. Results
  6. Discussion
  7. Conflict of interest statement
  8. Acknowledgements
  9. References

A.W., J.C., O.H., F.H. and L.N. have nothing to declare. B.L. is employed by Baxter Healthcare Inc. P. S. is a member of the scientific advisory board of Gambro AB.

Acknowledgements

  1. Top of page
  2. Abstract
  3. Introduction
  4. Materials and methods
  5. Results
  6. Discussion
  7. Conflict of interest statement
  8. Acknowledgements
  9. References

We express our gratitude to the patients and personnel at Karolinska University Hospital who participated in the study. Special thanks to Annika Moberg and Christina Bäckmark for surgery planning, to Annika Nilsson, Ann-Christin Emmoth and Ulrika Jensen for assistance in patient recruitment and excellent nursing and to John Sandberg for patient recruitment. This study was supported by the GENECURE (http://www.genecure.eu) project (FP6 EU-grant LSHM-CT-2006-037697), the Swedish Research Council (No. 521-2007-3336), Karolinska Institutet Centre for Gender Medicine, Loo and Hans Ostermans’ and Westman’s Foundations, the Åke Wiberg Foundation, the Heart and Lung Foundation, the Swedish Kidney Foundation and Baxter Healthcare. Anna Witasp’s doctoral grant was supported by the Karolinska Institutet faculty for funding of postgraduate students (KID).

References

  1. Top of page
  2. Abstract
  3. Introduction
  4. Materials and methods
  5. Results
  6. Discussion
  7. Conflict of interest statement
  8. Acknowledgements
  9. References