SEARCH

SEARCH BY CITATION

Keywords:

  • adipocyte;
  • cachexia;
  • cancer;
  • nutrition;
  • obesity;
  • secretion

Abstract

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

Abstract.  Rydén M, Agustsson T, Andersson J, Bolinder J, Toft E, Arner P (Intervention and Technology (CLINTEC), Karolinska University Hospital, Huddinge, Stockholm, Sweden; Norrlands University Hospital, Umeå, Sweden; and Ersta Hospital, Stockholm, Sweden). Adipose zinc-α2-glycoprotein is a catabolic marker in cancer and noncancerous states. J Intern Med 2012; 271: 414–420

Objective.  Zinc-α2-glycoprotein (ZAG) has been proposed as a tumour-derived cancer cachexia factor. However, ZAG is produced by some normal tissues, including white adipose tissue (WAT), and high serum ZAG levels are present in nonmalignant conditions. We determined whether human WAT contributes to serum ZAG levels and how serum and WAT-secreted ZAG levels correlate with catabolism in patients with cancer and in obese subjects undergoing a very low-calorie diet (VLCD) for 11 days.

Design/subjects.  ZAG levels in serum and in conditioned medium from WAT/adipocytes were determined by enzyme-linked immunosorbent assay. ZAG release from WAT in vivo was determined in 10 healthy subjects. The correlation between ZAG and cachexia was studied in 34 patients with newly diagnosed gastrointestinal cancer. The impact of a VLCD on ZAG release and serum levels was assessed in 10 obese women.

Results.  ZAG was released from abdominal WAT and adipocytes in vitro. However, the arteriovenous differences in vivo showed that there was no significant contribution of WAT to the circulating levels. WAT-secreted but not serum ZAG correlated positively with poor nutritional status but not with fat mass (or body mass index) in patients with gastrointestinal cancer. In obese subjects on a VLCD, ZAG secretion from WAT increased significantly whereas serum levels remained unaltered.

Conclusions.  ZAG is released from human WAT, but this tissue does not contribute significantly to the circulating levels. WAT-secreted ZAG correlates with nutritional status but not with fat mass in both cancer and nonmalignant conditions. Adipose ZAG is therefore a local factor activated primarily by the catabolic state per se.


Introduction

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

Fifty per cent of all patients with cancer experience an unintentional weight loss termed cancer cachexia, which is strongly associated with reduced survival and poor response to antitumour treatment. Although both fat mass and lean body mass are depleted in advanced cancer cachexia [1], longitudinal studies have demonstrated that loss of fat mass precedes that of lean body mass [2]. The mechanisms promoting weight loss in cancer are not clear. Because there is no effective treatment for cancer cachexia [3], a better understanding is vital to develop directed therapies. Results in recent years have demonstrated that loss of fat mass in human cancer cachexia is associated with increased breakdown (lipolysis) of the lipids within the fat droplet of adipocytes [4]. In vivo and in vitro studies have confirmed the existence of enhanced hormone-induced lipolysis in patients with cancer cachexia [5].

What activates lipolysis in cancer cachexia? Although this remains unclear, it has been proposed that factors derived from host and/or tumour tissues could be involved [1]. One of the most interesting candidates is the protein zinc-α2-glycoprotein (ZAG) (reviewed in [6]). This protein is identical to the previously described lipid mobilization factor, which was originally described in the urine of patients with cancer. ZAG is expressed and secreted from many normal tissues but is also over-expressed in several types of tumours [6]. It is well established that ZAG, at least in vitro, stimulates lipolysis in both rodent and human fat cells (reviewed in [7]). We recently demonstrated increased gene expression and secretion of ZAG from adipose tissue of patients with cancer cachexia, and the secretion was proportional to the degree of weight loss [8]. It is therefore conceivable that increased levels of ZAG stimulate adipocyte lipolysis in cancer cachexia. However, a number of questions regarding ZAG as a weight loss factor remain unanswered. Is ZAG just a local factor in adipose tissue in cancer cachexia or is circulating ZAG, possibly derived from the tumour and/or other tissues, also involved? Is increased ZAG production related to the nutritional status in cancer or is it a reflection of reduced fat mass? Is increased ZAG expression/activity specific for cancer cachexia or do the levels increase in other catabolic conditions? Does adipose ZAG contribute to the circulating level of ZAG? This study was undertaken to answer these clinically important issues. Levels of circulating and adipose-secreted ZAG were determined in patients with abdominal cancer and correlated with the Patient-generated Subjective Global Assessment (PG-SGA) score, which is a well-established method to assess nutritional status in patients with cancer [9]. As a noncancer model of catabolism, we investigated obese subjects before and after a very low-calorie diet (VLCD). The contribution of adipose-derived ZAG to the circulation was determined by assessing the arteriovenous difference in ZAG concentrations between arterialized blood and the vein draining the abdominal subcutaneous adipose tissue.

Materials and methods

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

Study cohorts and clinical measurements

Cohort 1 included four subjects undergoing cosmetic liposuction who anonymously donated white adipose tissue samples and freshly isolated fat cells. These subjects were not selected on the basis of sex, body mass index (BMI) or age. Cohort 2 included 10 (seven men and three women) otherwise healthy subjects with a large range of age (27–70 years) and BMI values (23–45 kg m−2) for whom the arteriovenous difference across the abdominal subcutaneous adipose tissue was assessed in vivo as described later. Cohort 3 included 34 subjects with different forms of newly diagnosed gastrointestinal cancer. The clinical characteristics of this cohort are described in Table 1. Only patients who were judged to be physically fit and who had not received any previous anti-cancer treatment were included. None of the patients had signs of diabetes or any other concomitant cancer diagnosis. Cohort 4 included 10 obese but otherwise healthy women subjected to caloric restriction with a VLCD as described [10]. Investigations were conducted in these subjects in the morning after an overnight fast, on the day before and during day 11 of the VLCD. On day 11, all subjects displayed markedly elevated levels of ketone bodies.

Table 1. Clinical characteristics of patients with gastrointestinal cancer (cohort 3)
MeasuresOutcome
Sex (male/female), n26/8
Age (years, mean ± SD)65 ± 9
Body mass index (kg m−2, mean ± SD)25 ± 4
Body fat (%, mean ± SD)28 ± 5
SGA-score (mean, range) 7 (1–18)
Colorectal cancer (n) 5
Pancreatic cancer (n)12
Stomach cancer (n) 4
Gall bladder cancer (n) 2
Oesophagus cancer (n) 9
Liver cancer (n) 2

All study subjects gave their informed written consent to participate. Methods for obtaining blood samples and fat biopsies were approved by the regional ethics committees, and all studies were performed in accordance with the Declaration of Helsinki.

Clinical examination

All subjects visited the laboratory after an overnight fast. Height, weight, body composition by bioimpedance using QuadScan 4000 (Bodystat Ltd, Isle of Man, British Isles) and indirect calorimetry using Deltatrac™ (Datex-Engstroms, Helsinki, Finland) were determined. A venous blood sample was obtained for the analysis of standard clinical laboratory parameters as well as ZAG protein levels. Arteriovenous differences were determined as described previously [11]. In brief, the cephalic vein was cannulated in a retrograde direction with the tip of the cannula (Optiva 2 18-20G; Johnson & Johnson, New Brunswick, NJ, USA) at the wrist joint, and the hand was placed in a heated chamber (Biomedical Engineering Department, Huddinge University Hospital, Sweden) with the air warmed to 60 °C to provide arterialized samples. The catheter was infused with saline at 50–100 mL h−1 to maintain flow. The degree of arterialization was confirmed by blood gases (i-STAT 1; Abbott, Abbott Park, IL, USA) and was always greater than 94%. A 20 G 15-cm catheter was sited in a superficial epigastric vein, as previously described [12]. To ensure that blood was collected from subcutaneous adipose tissue and not from deeper structures, O2 saturation was confirmed to be >85%. Blood samples were taken simultaneously from the arterialized hand vein and superficial epigastric vein. The arteriovenous technique used herein has been thoroughly validated (reviewed in [13]). In the patients with cancer (cohort 3), the nutritional status was assessed using the PG-SGA, a standardized oncology questionnaire [9]. Tumour stage was assessed postoperatively as described [14].

Fat biopsies

After the clinical examination, an abdominal subcutaneous fat sample (about 0.5 g) was obtained by needle biopsy as described [15]. The tissue sample was rapidly rinsed in saline and cut into small pieces. A portion of the sample was used to isolate fat cells as described later.

Isolation of adipocytes from adipose tissue

Fat cells were isolated from adipose tissue pieces using the collagenase procedure of Rodbell [16]. In brief, tissue samples were cut into approximately 20 mg pieces and incubated (1 g tissue mL−1 medium) in Krebs-Ringer phosphate (KRP) buffer (pH 7.4) supplemented with 4% bovine serum albumin (BSA) and 0.5 mg mL−1 collagenase type 1 for 60 min at 37 °C in a shaking water bath. The isolated fat cells were collected on a nylon mesh filter and were washed 4–5 times with KRP buffer containing 0.1% BSA. The purity of the isolation procedure was estimated by investigating 200 cells for each subject under light microscopy.

Measurements of glycerol, nonesterified fatty acids and ZAG

Plasma glycerol and nonesterified fatty acids (NEFAs) were determined as described previously [5]. ZAG protein levels in serum and in conditioned media were determined using an enzyme-linked immunosorbent assay kit (BioVendor, Palackeho, Czech Republic). Protein secretion from adipose tissue explants and isolated fat cells was determined as described [17]. In brief, adipose tissue (about 300 mg) or 300-μL packed freshly isolated adipocytes were incubated in 3 mL medium for different time periods at 37 °C. The incubation medium was then collected from each sample and centrifuged at 200 g for 10 min to remove cell debris, and the supernatant was stored at −80 °C until analysis. For the determination of tissue secretion, tissue samples were collected after the incubation periods and subjected to lipid extraction. ZAG secretion in these samples was expressed as the protein concentration in the medium corrected for the lipid weight of the incubated tissue and the incubation time.

Statistical analysis

The reported values are mean ± SD or range. Results were compared using regression analysis or paired t-test. Adipose ZAG values in the VLCD study were log10 transformed before comparison. Based on previously published data on ZAG secretion from adipose tissue [8] and serum ZAG (referenced in [7]), we calculated that we could measure a 20–25% difference in mean values between different groups and conditions. To our knowledge, there has been no previously reported information on arteriovenous sampling of ZAG. For this study, we used our previously published data on the chemokine monocyte chemoattractant protein-1 [11] to show that we could detect a 20% difference with 80% power at < 0.05.

Results

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

ZAG is an adipokine but is not released into the circulation

To determine whether ZAG is released from adipocytes in a time-dependent manner, adipose tissue samples or fat cells were incubated in suspension for 0.5, 1, 1.5, 2 and 3 h (cohort 1, n = 4, data not shown). Under these conditions, there was an almost linear increase in ZAG secretion, with levels 2.5 times higher at 3 h compared with 1 h (< 0.05). To assess whether adipose tissue releases significant levels of ZAG into the circulation, we compared the levels in abdominal subcutaneous venous blood and arterialized blood in 10 healthy subjects (cohort 2). We found that the concentrations were almost identical in venous and arterialized blood (57.5 ± 11.5 and 64.8 ± 7.9 μg mL−1, respectively, = 0.07), and there was no significant fractional extraction of ZAG from any of the blood compartments implying that adipose ZAG is not released to a significant degree into the circulation (Fig. 1). By contrast, the concentrations of NEFAs and glycerol were significantly higher in venous compared with arterialized blood (NEFAs: 1.2 ± 0.5 vs. 0.7 ± 0.2 mmol L−1, = 0.001; glycerol: 224 ± 139 vs. 59 ± 19 μmol L−1, = 0.002; data not shown).

image

Figure 1. Zinc-α2-glycoprotein (ZAG) is not released into the circulation from abdominal white adipose tissue. ZAG concentrations were determined in 10 healthy subjects in the vein draining subcutaneous abdominal adipose tissue as well as in arterialized venous blood from the dorsal part of the hand. (a) Comparison of ZAG levels in venous and arterialized blood presented as a box plot and compared by Student’s pairedt-test. (b) To estimate whether there was any release (above the dotted line) or uptake (below the dotted line) of ZAG in individual subjects, the concentrations in arterialized blood were subtracted from those in venous blood. Values were correlated with body mass index using linear regression analysis.

Download figure to PowerPoint

ZAG secretion from adipose tissue but not circulating levels is associated with nutritional status

Figure 2 shows the relationship between nutritional status and circulating or adipose-secreted ZAG in patients with gastrointestinal cancer (cohort 3) as examined by linear regression analysis. A strong positive correlation between adipose secretion of ZAG and PG-SGA scores was observed by regression analysis (Fig. 2a) or by Spearman correlation analysis (< 0.001). By contrast, there was no relationship between the circulating ZAG levels and PG-SGA scores (Fig. 2b). Furthermore, there was no correlation between adipose-secreted and circulating ZAG levels (r = 0.18, = 0.33; data not shown). Neither adipose nor serum ZAG levels correlated with tumour stage (data not shown).

image

Figure 2. Secretion of zinc-α2-glycoprotein (ZAG) from adipose tissue, but not serum levels, correlates with nutritional status in patients with cancer. Abdominal adipose tissue secretion (a) and serum levels (b) of ZAG were correlated with PG-SGA scores in 34 patients with gastrointestinal cancer using linear regression analysis.

Download figure to PowerPoint

ZAG is not correlated with BMI or fat mass in patients with cancer

There was no relationship between circulating or adipose-secreted ZAG and BMI in the same group of patients with cancer (cohort 3, Fig. 3). Similarly, there was no relationship when body fat mass values determined by bioimpedance were used instead of BMI (r = 0.2, = 0.15). Furthermore, ZAG levels were not related to BMI or body fat when only nutritionally fit patients were included in the analysis (PG-SCA score < 3, n = 16).

image

Figure 3. Zinc-α2-glycoprotein (ZAG) levels do not correlate with body mass index (BMI). Levels of ZAG secreted from adipose tissue samplesin vitro(a) or serum levelsin vivo(b) were compared with BMI in the same subjects as in Fig. 2.

Download figure to PowerPoint

Adipose ZAG is a marker of catabolism

To assess whether ZAG levels in the circulation and in adipose tissue are influenced by weight loss in subjects without cancer, we used samples from a cohort of 10 obese subjects subjected to a VLCD (cohort 4). The clinical findings of this group are shown in Table 2. Following dietary intervention for 11 days, pronounced caloric restriction induced a significant weight loss involving loss of fat but not of lean body mass. As expected, this was accompanied by enhanced plasma levels of glycerol and fatty acids as a result of increased lipolysis. Moreover, the respiratory quotient was decreased because of increased fat oxidation. ZAG levels before and after a VLCD are shown in Fig. 4. There was a strong (i.e. 2-fold) increase in the rate of adipose ZAG release during the period of intervention with the VLCD, but there was no change in the circulating concentration. Furthermore, there was a significant and strong positive correlation between percentage weight loss and the increase in adipose ZAG secretion (r = 0.72, = 0.019; data not shown).

Table 2. Clinical characteristics of the subjects given a very low-calorie diet (VLCD) (cohort 4)
MeasuresBefore VLCDAfter VLCDP-value
  1. Values are mean ± SD, compared by Student’s paired t-test.

Age (years)42 ± 12
Total body weight (kg)115 ± 12110 ± 12<0.001
Body fat (kg)72 ± 1868 ± 170.027
Lean body mass (kg)43 ± 1042 ± 110.47
Body mass index (kg m−2)41 ± 539 ± 5<0.001
Plasma glycerol (μmol L−1)80 ± 2292 ± 200.04
Plasma nonesterified fatty acid (mmol L−1)0.77 ± 0.231.04 ± 0.120.01
Respiratory quotient0.77 ± 0.030.72 ± 0.010.002
image

Figure 4. Adipose tissue–derived zinc-α2-glycoprotein (ZAG), but not the serum level, increases in obese subjects under pronounced caloric restriction. Ten obese but otherwise healthy women were subjected to caloric restriction with a very low-calorie diet for 11 days. Values before and after treatment are presented as box plots and compared by Student’s pairedt-test (a) Adipose tissue secretion of ZAG, Log10 transformed values were compared. (b) Serum ZAG levels were normally distributed.

Download figure to PowerPoint

Discussion

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

We recently found that adipose ZAG release was positively related to reported weight loss in patients with cancer cachexia [8]. In this study, we have shown that there is a strong relationship between adipose ZAG release and nutritional status (i.e. PG-SGA score) in patients with abdominal cancer. However, we have extended our previous findings further and also demonstrate that there are marked similarities in the regulation of ZAG in cancer cachexia and catabolism induced by a VLCD in healthy obese subjects without cancer. Thus, the secretion of ZAG from adipose tissue is markedly enhanced by a relatively short (i.e. 11 days) period of caloric restriction, and there is a strong relationship between weight loss after energy deprivation and increased adipose ZAG release. Our results in the cohort subjected to the VLCD are somewhat different from those of a previous study by Capel et al. in which obese subjects without cancer underwent caloric restriction by a VLCD for 1 month followed by a 5- to 6-month phase of weight stabilization (WS) [18]. Fat biopsies were obtained before and after the VLCD and WS phases, respectively. ZAG mRNA levels determined by microarray analysis and quantitative PCR were found to be significantly increased (by approximately 30%) only after the WS phase [18]. By contrast, the obese subjects in this study were subjected to a shorter VLCD intervention and we evaluated the secretion of ZAG protein.

As ZAG is released from adipose tissue and fat cells in a time-dependent manner, our data suggest that the active secretion of the protein within adipose tissue is stimulated by a catabolic state. This is probably due to gene activation as we previously observed a positive correlation between protein and mRNA expression in adipose tissue in a different cohort of patients with cancer [8].

Several studies have shown that ZAG mRNA and protein levels are decreased in adipose tissue of obese subjects [19–22]. It has therefore been suggested that ZAG inversely reflects body fat mass [23]. However, it is unlikely that increased ZAG secretion from adipose tissue during catabolism is merely a reflection of reduced fat mass. This notion is supported by the fact that there was no relationship between BMI or relative body fat content and release of ZAG from adipose tissue in patients with cancer. This was also the case when analysing the relationship only in cancer patients with a normal nutritional status.

As mentioned earlier, there are many endogenous sources of ZAG besides adipose tissue, including the liver and the tumour itself. Our present findings suggest a specific role of adipose tissue in relation to altered ZAG expression in catabolic conditions. There was no relationship between nutritional status and the circulating ZAG level amongst patients with cancer; in addition, there was no relationship between secreted and circulating ZAG in the same cohort. Furthermore, the serum levels of ZAG did not change during a VLCD. It is therefore likely that adipose tissue has a unique regulation with regard to ZAG production/release. Only this tissue responds to a nutritional deficit, at least under the conditions examined herein. The poor relationship between circulating levels of ZAG and body fat mass is also illustrated by the fact that circulating levels of ZAG in obesity have been reported to be increased [24], unchanged [25] or reduced [22, 26]. The assumption that ZAG in adipose tissue is a local factor and unrelated to the levels in blood is strongly supported by the finding Department of Infectious Immunology and Pediatrics, Shinshu University, Matsumoto, Japan.

The assumption that ZAG concentrations in plasma were almost identical in the abdominal subcutaneous vein (draining adipose tissue) and arterialized blood. Hence, in subjects without cancer displaying a large BMI range, there was no evidence in vivo of a net uptake or release of ZAG from adipose tissue. By contrast, classical adipose-derived circulating metabolites (glycerol and NEFAs) were markedly higher in samples from the vein draining adipose tissue compared with arterialized blood. These assessments were performed in abdominal subcutaneous adipose tissue, and therefore we cannot exclude the possibility that there might be differences in other subcutaneous depots and/or visceral adipose tissue. However, this issue could not be investigated for technical and ethical reasons.

Based on the present and previous results, we propose the following model of regulation of ZAG and fat mass in humans. In catabolic states, such as cancer cachexia, or during insufficient food intake, there is a need for energy-rich lipids to be mobilized from the adipose tissue by increased fat cell lipolysis. This process is initiated/facilitated by an increased local production of ZAG within the adipose tissue which is not released into the circulation. The local increase in ZAG may activate lipolysis so that more fatty acids are available as an energy source.

In summary, the secretion of ZAG from subcutaneous adipose tissue, but not its circulating levels, is strongly and inversely related to the nutritional status of patients with abdominal cancer and is markedly increased during caloric restriction in obese patients without cancer in proportion to the observed weight loss. This suggests that adipose ZAG is a local factor that is activated by a catabolic state per se rather than specifically by cancer cachexia.

Acknowledgements

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

We are grateful to Kerstin Wåhlén, Gaby Åström, Eva Sjölin, Katarina Hertel and Britt-Marie Leijonhufvud for their skilled technical assistance. This study was supported by grants from the Swedish Cancer Society, the Swedish Heart and Lung Foundation, the Swedish Diabetes Association, the Swedish Research Council, the NovoNordisk Foundation and the European Union ADAPT (HEALTH-F2-2008-201100) and COST action (BM0602).

References

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