SEARCH

SEARCH BY CITATION

Abstract

  1. Top of page
  2. Abstract
  3. Introduction
  4. Methods and Procedures
  5. Results
  6. Discussion
  7. Acknowledgments
  8. Disclosure
  9. REFERENCES

Cystatin C, an endogenous inhibitor of cathepsin proteases has emerged as a biomarker of cardiovascular risk and reduced renal function. Epidemiological studies indicate that serum cystatin C increased in human obesity. Here, we evaluated the contribution of adipose tissue to this elevation, based on our previous observation that cystatin C is produced by in vitro differentiated human adipocytes. We measured serum cystatin C in 237 nonobese (age: 51 ± 0.8 years; BMI: 22.8 ± 0.11 kg/m2) and 248 obese subjects (age: 50 ± 0.8 years; BMI: 34.7 ± 0.29 kg/m2). Creatinine-based estimated glomerular filtration rate (eGFR) was calculated to account for renal status. Cystatin C gene expression and secretion were determined on surgical adipose tissue biopsies in a distinct group of subjects. Serum cystatin C is elevated in obese subjects of both genders, independently of reduced eGFR. Cystatin C mRNA is expressed in subcutaneous and omental adipose tissue, at twice higher levels in nonadipose than in adipose cells. Gene expression and cystatin C release by adipose tissue explants increase two- to threefold in obesity. These data confirm elevation of serum cystatin C in human obesity and strongly argue for a contribution of increased production of cystatin C by enlarged adipose tissue. Because cystatin C has the potential to affect adipose tissue and vascular homeostasis through local and/or systemic inhibition of cathepsins, this study adds a new factor to the list of adipose tissue secreted bioactive molecules implicated in obesity and obesity-linked complications.


Introduction

  1. Top of page
  2. Abstract
  3. Introduction
  4. Methods and Procedures
  5. Results
  6. Discussion
  7. Acknowledgments
  8. Disclosure
  9. REFERENCES

Epidemiological evidence indicates that human obesity strongly associates with cardiovascular diseases (1). An emerging theory suggests the contribution of factors secreted by adipose tissue to obesity-linked vascular alterations. We have proposed such a role for the proatherogenic cysteine protease cathepsin S, which is produced by the adipose tissue and demonstrates increased circulating concentration in obese subjects (2). Cathepsins' enzymatic activity is regulated by endogenous inhibitors, the most abundant being cystatin C (3). In humans, cystatin C protein is detected in normal arteries, but present at low levels in the atherosclerotic plaque where cathepsin S is abundant (4). In animal models, the deletion of cystatin C gene in atherogenesis prone apolipoprotein E-deficient mice, promotes the atherosclerotic process (5,6). Thus, reduced cystatin C and/or imbalance between cystatin C and cathepsin S could be considered a risk factor for vessel wall damage. Contrary to this hypothesis, epidemiological studies consistently report a positive relationship between elevated circulating cystatin C and cardiovascular outcomes (7,8,9,10,11,12). This counterintuitive association suggests that elevation of cystatin C represents a compensatory mechanism to reduce cathepsin S activity.

Unrelated to its functional role as an inhibitor of cathepsins, cystatin C has emerged as a surrogate marker of the glomerular filtration rate (GFR) at least as good as serum creatinine (13,14). Unlike serum creatinine, serum cystatin C is independent of muscle mass (15). However, a few studies have reported that increased adiposity or obesity associates with higher cystatin C circulating concentrations. Positive associations of serum cystatin C with BMI were found in a geriatric population (16), in the general population (9), and, with body weight, in a large white population (17). Higher waist circumference, an index of visceral adiposity, and increased per cent of body fat also associate with higher serum cystatin C in apparently healthy subjects (18,19). Recently, a graded association between higher BMI and elevated serum cystatin C was reported in American adults (20). In these epidemiological studies, however, the contribution of increased adiposity vs. reduced GFR in obesity-linked elevation of serum cystatin C was unclear. Previous observations from our group have shown that cystatin C is expressed and released by human preadipocytes differentiated in vitro (21). This raises the hypothesis that adipose tissue might directly contribute to enhance cystatin C circulating levels in obesity, as observed for cathepsin S (2). This study was designed to test this hypothesis.

To this aim, we measured serum cystatin C in a large population. We separated the study participants in two groups of nonobese and obese subjects to investigate specifically the relationship between increased adipose tissue mass and serum cystatin C. To account for the effect of reduced renal filtration, we used the abbreviated equation from the Modification of Diet in Renal Disease study (22) to estimate GFR in these subjects. This allowed measuring the respective effect of obesity and reduced GFR to increase the circulating concentrations of cystatin C. In addition, to evaluate the capacity of adipose tissue to produce cystatin C, we performed a series of in vitro experiments to determine cystatin C gene expression and secretion in human adipose tissue obtained from a distinct group of subjects.

Methods and Procedures

  1. Top of page
  2. Abstract
  3. Introduction
  4. Methods and Procedures
  5. Results
  6. Discussion
  7. Acknowledgments
  8. Disclosure
  9. REFERENCES

BMI was used to segregate a previously described population (23) in two groups of 250 nonobese (BMI ≤25 kg/m2) and 250 obese subjects (BMI ≥30 kg/m2). One obese and 11 nonobese subjects were excluded on the criteria of age <20 or >80 years. GFR was estimated by the abbreviated equation from the Modification of Diet in Renal Disease study: estimated GFR (eGFR) = 175 × serum creatinine (µmol/l)−1.154 × age−0.203 × 0.742 (if female) (22). Three subjects with discordant, very low eGFR (<30 ml/min/1.73 m2) were excluded. The final two groups of participants comprise 237 nonobese and 248 obese subjects, whose characteristics are shown in Table 1. Local ethics committees approved the investigations and all subjects gave informed consent.

Table 1.  Clinical and biological characteristics of the study participants
inline image

Blood samples were collected after an overnight fast. Plasma glucose, triglycerides, total cholesterol, and high-density lipoprotein–cholesterol levels were measured enzymatically. Frozen serum (−20 °C) was used for the following determinations. Insulin was measured with an IRMA kit (Bi-INSULIN IRMA; CisBio International, Gif-sur-Yvette, France). Insulin sensitivity was evaluated by the quantitative insulin sensitivity check index (24): QUICKI = 1/(log fasting insulin (µU/ml) + log fasting glycemia (mg/dl)). Creatinine was determined in ARCHITECT analyzer (Abbot Diagnostics, Rungis, France). Cystatin C was measured by a particle-enhanced immunonephelometric assay (DakoCytomation, Trappes, France). The assay analytical sensitivity was 0.07 mg/l. The intra- and interassay variability was 2.4 and 2.2%, respectively.

We used a distinct group of subjects, including healthy nonobese individuals and obese subjects recruited from a large population involved in a gastric surgery program (Department of Nutrition, Hôtel-Dieu Hospital, Paris, France), as described in refs. 2 and 25. Paired subcutaneous and visceral adipose tissue biopsies were collected in obese subjects (age: 45 ± 2.2 years; BMI: 47.0 ± 1.5 kg/m2, n = 35) at the time of gastric surgery. Surgical subcutaneous adipose tissue biopsies were obtained during elective surgery in nonobese subjects (age: 37 ± 2.4 years; BMI: 22.7 ± 0.7 kg/m2, n = 16). Adipose tissue samples were stored at −80 °C before RNA extraction. Explants were prepared from fresh subcutaneous adipose tissue biopsies. Approximately 300 mg of minced adipose tissue were incubated in 2 ml of endothelial cell basal medium (Promocell, Heidelberg, Germany) containing 3% bovine serum albumin and antibiotics. Each time point was done in triplicate. Conditioned media were stored at −80 °C before measurement of cystatin C and leptin concentrations by enzyme-linked immunosorbent assay (Human cystatin C Quantikine and human leptin Quantikine enzyme-linked immunosorbent assay kits; R&D Systems Europe, Lille, France). Adipocytes and nonadipose cells of the stroma vascular fraction (SVF) were separated by collagenase digestion when at least 2 g of adipose tissue were available (2). Preadipocytes, endothelial cells, and macrophages were further isolated from the SVF after two passages in culture or immuno-isolation as described previously (26,27,28). Cells were stored at −80 °C before RNA extraction. The Ethics Committees of the Hôtel-Dieu Hospital approved these investigations both for obese and nonobese persons. All subjects gave a written informed consent.

RNA extraction was performed on adipose tissue or isolated cells using the RNeasy RNA Mini Kit (Qiagen, Courtaboeuf, France). Total RNA concentrations and quality were confirmed using the Agilent 2100 bioanalyzer (Agilent Technologies, Massy, France). RNA from several human tissues included in a commercial kit (First Choice human total RNA survey Panel; Ambion, Courtaboeuf, France) was also analyzed. Total RNA (1 µg) was reversed transcribed using random hexamers and Supercript II reverse transcriptase. Real-time PCRs were conducted with 25 ng cDNA using the Taqman universal PCR mix and assessed in a detection system instrument (Applied Biosystems, Minneapolis, MN). mRNA values were normalized to 18S rRNA (Eurogentec, Angers, France). Primers for human genes were as follows: cystatin C: forward 5′-cttggacaactgccccttc-3′, reverse 5′-AAGGCACAGCGTAGATCTGG-3′; leptin: forward 5′-TTGTCACC AGGATCAATGACA-3′, reverse 5′-GTCCAAACCGGTGA CTTTCT-3′.

Data are expressed as mean ± s.e.m.. The Shapiro–Wilcoxon test was used to test the Gaussian distribution of clinical and biological measures. Skewed variables were log-transformed before statistical analyses. Student's t-test and χ2-test were used to assess for differences between lean and obese subjects for continuous and categorical variables, respectively. Comparison between means was performed by ANOVA, followed by p for linear trend post-test when appropriate or by Student' t-test. For comparison between nonobese and obese subjects, we adjusted P values for age, gender, and/or cardiovascular risk factors by multiple regression modeling. Statistical analysis was performed with JMP (SAS Institute, Cary, NC). A P value ≤ 0.05 was considered significant.

Results

  1. Top of page
  2. Abstract
  3. Introduction
  4. Methods and Procedures
  5. Results
  6. Discussion
  7. Acknowledgments
  8. Disclosure
  9. REFERENCES

As expected, the obese subjects display a more severe phenotype than nonobese subjects, including hyperinsulinemia, increased serum triglycerides, and reduced high-density lipoprotein–cholesterol (Table 1). Serum cystatin C concentrations were significantly increased in the obese group after adjustment for age and gender. This effect of obesity remained significant, when adjusted for smoking, dyslipidemia, hypertension, and type 2 diabetes in a multivariate model, and after exclusion of participants under drug therapy (P < 0.0001). Different from serum cystatin C, serum creatinine and, in turn, the creatinine-based eGFR were not significantly changed in obesity (Table 1).

To account for renal status, we choose to segregate the study participants in tertiles of eGFR. The subjects assigned to each tertile were then classified by gender and BMI (Figure 1). Whatever the gender, subjects with the lowest eGFR were ∼10 years older than those classified in the highest eGFR group. As expected, serum creatinine concentrations increased regularly with each step of eGFR reduction, reflecting the inverse relationship between the two variables in the Modification of Diet in Renal Disease equation (Figure 1a). Similarly, serum cystatin C increased significantly along with decreased eGFR, with a roughly parallel trend for each gender and BMI category (Figure 1b; P for trend <0.001 for each group). Although serum creatinine was systematically higher in male than in female subjects, a significant male gender effect increasing cystatin C concentrations was observed only in the group with the lowest eGFR. Different from serum creatinine, which was strikingly unaffected by obesity, serum cystatin C was consistently higher in obese than in nonobese subjects. Moreover, this obesity-linked elevation was in the same order of magnitude (∼ +0.070 mg/l) in each tertile of eGFR and whatever the gender.

image

Figure 1. Serum cystatin C and creatinine concentrations in nonobese and obese subjects classified in tertile of eGFR. (a) Serum creatinine; (b) serum Cystatin C. Cystatin C and creatinine concentrations were measured in the serum of nonobese (white and dotted bars) and obese subjects (gray and dark bars). The participants were segregated in tertiles of glomerular filtration rate estimated by the MDRD equation (eGFR). BMI and gender identify each group, including the number of subjects (n) indicated. Results are mean ± s.e.m. *P < 0.001 vs. nonobese subject of same gender; †P < 0.001 vs. female of same adiposity category. eGFR, estimated glomerular filtration rate; MDRD, Modification of Diet in Renal Disease.

Download figure to PowerPoint

Next, we sought to evaluate the capacity of human adipose tissue to produce cystatin C. First, by measuring cystatin C mRNA levels in a commercial set of human tissues, we found that adipose tissue was amongst the tissues expressing the highest level of cystatin C mRNA (Figure 2). Second, in our human adipose tissue samples, cystatin C mRNA increased threefold in obesity, an elevation slightly less than that measured for the obesity-induced leptin gene (Figure 3a). Cystatin C was similarly expressed in omental as in subcutaneous adipose tissue of the same obese subject (omental/subcutaneous: 1.53 ± 0.26, n = 25, P = 0.254), different from leptin, which was predominantly expressed in the subcutaneous depot (omental/subcutaneous: 0.45 ± 0.12, n = 25, P = 0.002). Third, we measured cystatin C and leptin concentrations in the media of human adipose tissue explants, showing that both proteins were recovered in significantly larger amounts in obese than in nonobese adipose tissue conditioned medium (Figure 3b). In time course experiments, the amounts of cystatin C recovered in the medium increased linearly during the first 9 h of culture, both in obese and nonobese adipose tissue explants (Figure 3b, inset). Of note, the amounts of cystatin C released per g of tissue and per 24 h was 10- to 20-fold higher than the amounts of leptin in the same media. Finally, adipose tissue cell fractionation showed that cystatin C is expressed in adipocytes and, at a twice higher level, in SVF cells, while leptin mRNA is exclusively expressed in adipocytes, as expected (Figure 3c). The selection of cell types contained in SVF further revealed that cystatin C is equivalently expressed in preadipocytes, endothelial cells, and macrophages.

image

Figure 2. Differential expression of cystatin C mRNA in various human tissues. Data represent the mean of duplicate measurements of cystatin C mRNA in each tissue included in a total RNA commercial kit (First Choice human total RNA survey Panel; Ambion).

Download figure to PowerPoint

image

Figure 3. Cystatin C and leptin gene expression and release in human adipose tissue. (a) Cystatin C and leptin mRNA levels were determined in surgical subcutaneous adipose tissue biopsies obtained in nonobese (white bars, n = 12) and obese subjects (black bars, n = 25). Data are shown as mean ± s.e.m. of fold change over nonobese. *P < 0.05 vs. nonobese. (b) Cystatin C and leptin concentrations were measured in 24-h culture media of subcutaneous adipose tissue explants from nonobese (white bars, n = 4) or obese (black bars, n = 10) subjects. Inset: Cystatin C concentrations were measured in 6-, 9-, and 24-h culture media of subcutaneous adipose tissue explants from nonobese (open square, n = 2) and obese (black square, n = 3) subjects. Data are mean ± s.e.m. *P < 0.05 vs. nonobese. (c) Cystatin C and leptin mRNA levels were measured in different cell types isolated from subcutaneous adipose tissue. Data show fold change over adipocytes, as mean ± s.e.m. (n = 4–9). *P < 0.05 vs. adipocytes. SVF, stroma vascular fraction.

Download figure to PowerPoint

Discussion

  1. Top of page
  2. Abstract
  3. Introduction
  4. Methods and Procedures
  5. Results
  6. Discussion
  7. Acknowledgments
  8. Disclosure
  9. REFERENCES

This study confirms the association between obesity and elevated serum cystatin C in humans (20). Here, we attempted to discriminate the influence of reduced renal function, evaluated by creatinine-based eGFR, on this elevation. The eGFR values of the study participants were in the range associated with normal to moderately impaired renal function and none of the subjects has stage 3–5 chronic kidney disease (22,29). The main relevant finding is that obesity-induced elevation of circulating cystatin C was similar for each tertile of eGFR. Even in the group with the lowest eGFR, the expected elevation of cystatin C does not conceal this effect of obesity. Besides reduced eGFR, several nonrenal factors are known to influence cystatin C circulating concentrations, including older age (17). A recent report indicates that circulating concentrations increase by 0.047 mg/l for every 10 years of age in Japanese adult population (19). Here, a nearly twice higher increase was associated with obesity in age-matched subjects whatever the gender and eGFR category. Moreover, the difference in serum cystatin C between obese and nonobese subjects remains significant when adjusted for potential sources of variation, including smoking, dyslipidemia, hypertension and type 2 diabetes, as well as drug therapy. These observations suggest that higher BMI is the main determinant of obesity-linked increase in serum cystatin C.

This analysis further shows that serum cystatin C increased regularly with reduction of eGFR, following a parallel trend in the obese and nonobese groups. This suggests that circulating cystatin C concentrations reflect the renal status whatever the degree of obesity, at least in the range of BMI and eGFR values found in the groups of subjects studied. One limitation of our study, however, is that GFR was not directly measured given the cost and time consuming procedures required (29). Alternatively, we used the Modification of Diet in Renal Disease equation, which has been validated to estimate GFR in subjects in the range of BMI of our study participants (30). Clinical studies with direct measurement of GFR are needed to establish whether BMI has to be taken into account to adjust the cystatin C reference intervals, similar to male gender for creatinine. Whether cystatin C could be a suitable marker of eGFR in severely obese patients (BMI >40 kg/m2) in whom the evaluation of renal function is difficult, remains to be determined.

Increased serum cystatin C consistently observed in obese individuals, whatever their renal status, strongly suggest a role for adipose tissue as a contributor to circulating concentrations of this protein in obesity. In support of this hypothesis, we show that cystatin C is highly expressed in human adipose tissue, equivalently in subcutaneous and omental fat depots, and that adipose tissue expression of cystatin C is increased in obesity. Both adipose cells and several SVF cell types express cystatin C and might contribute to adipose tissue production of the protein. High amounts of cystatin C protein are secreted by human adipose tissue explants in vitro, at a much higher rate than leptin, which is virtually exclusively expressed in adipocytes. Because both proteins are in the same range of molecular weight (13 kDa for cystatin C and 16 kDa for leptin), the same difference exists on a molar basis. These greatly distinct rates of adipose tissue production reflect similar differences in circulating concentrations. Indeed, cystatin C circulates at concentrations in the range of 800–1,100 ng/ml, much higher than that of leptin amounting 10–60 ng/ml (ref. 2). The increased release of cystatin C by obese adipose tissue explants argues for a higher production of the protein in vivo, potentially by both subcutaneous and omental adipose tissue. This increase could arise from enlarged adipocytes and/or from SVF cells, including macrophages, which express cystatin C mRNA and infiltrate the adipose tissue in obesity (31). These in vitro data suggest that adipose tissue production of cystatin C contributes to the elevation of circulating concentrations in obesity, although the participation of other cells or tissues where cystatin C is expressed cannot be excluded.

In summary, our data show that serum cystatin C increases with adiposity and strongly argue for the contribution of adipose tissue to this elevation. The consequences of elevated serum cystatin C are still hypothetical. Based on its function as inhibitor of cysteine proteases, cystatin C has the potential to influence pathological processes relying, at least in part, on dysregulation of cathepsins. This includes atherosclerosis and other inflammatory-related diseases (32). Increased serum cystatin C might be part of regulatory mechanisms engaged to control the proatherogenic capacity of specific cathepsins such as cathepsin S (33). In this context, the contributing influence of systemic cystatin C and of its local action needs to be deciphered. In addition, recent observations from our and other groups suggest that several cathepsins promote adipose differentiation in vitro and contribute to adipose tissue development in mice, as indicated by reduced fat mass resulting from gene deletion (21,34,35,36). This suggests that increased adipose tissue production of cystatin C might participate locally into a protective mechanism to control adipose tissue mass through cathepsins inhibition. Our study adds cystatin C to the list of adipose-secreted factors with the potential to affect adipose tissue biology and obesity-linked complications through local and/or systemic actions.

Acknowledgments

  1. Top of page
  2. Abstract
  3. Introduction
  4. Methods and Procedures
  5. Results
  6. Discussion
  7. Acknowledgments
  8. Disclosure
  9. REFERENCES

We thank Danièle Lacasa for immunoselection of cells isolated from adipose tissue biopsies, Florence Marchelli, who contributed to the clinical database constitution and Jean-François Bedel, Guillaume Ladrange, Sandrine Rallier and Marie-Christine Rincon for their excellent technical assistance. This work was supported by the “Programme Hospitalier de Recherche Clinique,” Assistance Publique des Hôpitaux de Paris (AOR 02076), a grant from the French National Agency of Research (OBCAT, program ANR-05-PCOD-026-01), ALFEDIAM and Commission of the European Communities (Collaborative Project ADAPT, contract number HEALTH-F262008-201100).

REFERENCES

  1. Top of page
  2. Abstract
  3. Introduction
  4. Methods and Procedures
  5. Results
  6. Discussion
  7. Acknowledgments
  8. Disclosure
  9. REFERENCES