Fax: (011) 46 8 58582407.
Lipolysis—Not inflammation, cell death, or lipogenesis—Is involved in adipose tissue loss in cancer cachexia†
Article first published online: 14 AUG 2008
Copyright © 2008 American Cancer Society
Volume 113, Issue 7, pages 1695–1704, 1 October 2008
How to Cite
Rydén, M., Agustsson, T., Laurencikiene, J., Britton, T., Sjölin, E., Isaksson, B., Permert, J. and Arner, P. (2008), Lipolysis—Not inflammation, cell death, or lipogenesis—Is involved in adipose tissue loss in cancer cachexia. Cancer, 113: 1695–1704. doi: 10.1002/cncr.23802
We are grateful for the skilful technical assistance of Kerstin Wåhlén, Gaby Åström, Katarina Hertel, and Britt-Marie Leijonhufvud. We also thank Dr. Hans Wahrenberg for help in the statistical power calculations of the study.
- Issue published online: 17 SEP 2008
- Article first published online: 14 AUG 2008
- Manuscript Accepted: 2 JUN 2008
- Manuscript Revised: 30 MAY 2008
- Manuscript Received: 9 JAN 2008
- Swedish Cancer Society and the Swedish Research Council
- interleukin 6;
- tumor necrosis factor α;
Cancer cachexia is an important, negative prognostic marker that has been linked to systemic inflammation and cell death through unclear mechanisms. A key feature of cancer cachexia is loss of white adipose tissue (WAT) because of increased adipocyte lipolysis and possibly reduced lipid synthesis (lipogenesis). In this study, the authors investigated whether alterations in fat cell numbers, lipogenesis, or cytokine and/or leukocyte infiltration could account for some of the functional changes observed in WAT in cancer cachexia.
Blood and subcutaneous WAT samples were obtained from a 10 weight-stable patients, from 13 weight losing (cachexia) patients with cancer, and from 5 patients without cancer (noncancer patients) who initially were classified with cancer.
Systemic inflammation (increased circulating levels of interleukin 6 [IL-6]) and enhanced lipolysis were confirmed in the cachectic patients compared with the other patients. However, the messenger RNA expression of IL-6 and other cytokine or leukocyte markers, as well as WAT secretion of IL-6, were not altered in the patients with cachexia. Thus, the elevated serum levels of IL-6 that were observed in cachexia were not derived from WAT. Insulin-induced lipogenesis in adipocytes from patients with cachexia was the same as that in adipocytes from patients with weight-stable cancer and from noncancer patients (2.5-fold maximal stimulation; half-maximum effective concentration, ∼0.03 nmol/L). Fat cell size was decreased but adipocyte numbers were normal in cancer patients with cachexia, suggesting that there was no major fat cell death.
The current findings indicated that subcutaneous WAT does not contribute to the systemic inflammatory reaction and does not induce adipocyte insulin resistance in cancer cachexia. Moreover, increased fat cell lipolysis, not reduced lipogenesis or adipocyte cell death, appeared to be the primary cause of fat loss in this condition. Cancer 2008. © 2008 American Cancer Society.
Cancer cachexia is a wasting syndrome characterized by loss of skeletal muscle and adipose tissue1 and is a negative prognostic marker associated with poorer survival and considerably increased risk of developing complications to therapeutic interventions.2 It is well established that there is a link between cachexia and systemic inflammation. However, the mechanisms and effects of this inflammatory response are not clear. The factors behind muscle wasting (sarcopenia) have been studied intensely, whereas much less is known regarding the loss of adipose tissue in cancer patients.3, 4 It has been proposed but not demonstrated that reduced de novo synthesis of lipids (lipogenesis) may be an etiologic factor.5 Conversely, it has been demonstrated that increased in vivo lipolysis, resulting in loss of lipids from adipose tissue, is a key factor promoting adipose cachexia in cancer patients3, 4, 6, 7 that is independent of malnutrition.8–11 Several mechanisms promoting adipocyte lipolysis have been proposed. In animal models, tumor-derived lipolytic factors (eg, zinc-alpha 2 glycoprotein) or adipose tissue-derived lipolytic factors, such as interleukin 6 (IL-6) and tumor necrosis factor alpha (TNFα), have been suggested as etiologic factors.11, 12 In humans, the hormonal stimulation of lipolysis in human fat cells is enhanced in cancer cachexia because of an increased expression and action of the rate-limiting lipolytic enzyme hormone-sensitive lipase (HSL).13 This alteration is secondary to the cachectic state and is not a primary etiologic factor in the development of cachexia. The human study13 did not assess lipogenesis in fat cells from cachectic individuals.
White adipose tissue (WAT) is a highly active organ that, apart from its metabolic function, releases an array of secreted products with both systemic and local effects on lipid and glucose metabolism (for a review, see Ronti et al14). Increased adipose tissue inflammation is an important factor in the development of obesity-related complications, such as insulin resistance.15, 16 The infiltration of inflammatory cells, primarily macrophages, into WAT promotes the local production of inflammatory mediators, such as cytokines, from both leukocytes and adipocytes, which, in turn initiates a negative set of effects on adipocyte function and insulin responsiveness and also may induce fat cell death. For instance, it is well established that IL-6 and TNFα promote adipocyte lipolysis and attenuate insulin signaling through several mechanisms.17 Moreover, TNFα can induce apoptosis in human adipocytes.18
Metabolic alterations, including reduced insulin sensitivity, also can be observed when fat mass is reduced considerably, eg in lipodystrophy.19 It is not known whether the remaining adipose tissue in this or other conditions with reduced fat mass, such as cancer cachexia, displays increased inflammation.
We investigated whether subcutaneous WAT in patients with cancer cachexia could display an altered inflammatory reaction that could promote significant changes in fat cell function, including increased lipolysis and possibly reduced lipogenesis. To this end, we studied the expression of inflammatory markers (cytokines and leukocyte markers) at the systemic (ie, circulation) and local levels in WAT from cancer patients with or without cachexia. Because an excess of cytokines in cancer cachexia putatively could induce increased cell death, affecting WAT as well, we also investigated total adipose tissue cellularity.
MATERIALS AND METHODS
We included patients who were scheduled for suspected gastrointestinal cancer operation who were 1) fit despite their cancer disease, 2) had not received prior anticancer treatment, and 3) were willing to participate (n = 28 patients). The study was approved by the Ethics Committee of the Karolinska Institute. After the study was explained in detail to each patient, written consent was obtained. All patients who were included in the study were interviewed regarding body weight changes during the last year. Then, the patients were divided into 3 groups based on their diagnosis after surgery. Cancer cachexia (n = 13 patients) was defined as a confirmed diagnosis of gastrointestinal cancer in combination with unintentional weight loss of >5% of the habitual weight during the last 3 months or >10% weight loss during the last 6 months. The definition of weight loss in cachexia is not exact, but the inclusion limits that we used in this study were chosen based on previous findings discussed in a recent overview.20 Previous21, 22 studies have demonstrated that the 95% confidence interval for body weight change in healthy individuals is ±3.5% in 3 months and 5% during a 6-month follow-up. Therefore, cachexia should be suspected if a patient reports an involuntary weight loss >5% within a 6-month period. To increase the specificity in the definition of cachectic individuals, we chose to use an even stricter definition for this investigation. These patients had adenocarcinoma of the pancreas (n = 6 patients), colon (n = 1 patient), or gastric cardia (n = 1 patient), or they had esophageal cancer (n = 5 patients). A weight-stable group with confirmed cancer consisted of 10 patients who reported no important weight change during the last year. They were diagnosed with pancreatic adenocarcinoma (n = 7 patients), gastric adenocarcinoma (n = 1 patient), colon adenocarcinoma (n = 1 patient), or esophageal dysplasia (n = 1 patient). Postoperatively, 5 patients were reclassified, because histopathology showed a benign diagnosis (chronic pancreatitis or chronic cholecystitis). These patients constituted a third, noncancer group.
The patients came to the laboratory after an overnight fast. Height, weight, waist-to-hip ratio, and body composition with bioimpedance using Bodystat Quad Scand 4000 (Bodystat LTD, Isle of Man, UK) were determined. A venous blood sample was obtained for the determination of hemoglobin, transferrin, glycerol, fatty acid, albumin, glucose, and insulin levels by the hospital's accredited routine chemistry laboratories. TNFα and IL-6 levels were determined as described previously.23 Plasma glycerol and fatty acid levels were related to the total amount of body fat, which served as an in vivo index of lipolysis.13 The nutritional status was assessed by using a standardized questionnaire for oncology called the Subjective Global Assessment (SGA).24 Tumor stage was classified postoperatively according to the TNM Classification of Malignant Tumors from the International Union Against Cancer (6th edition).25
After the clinical examination, an abdominal subcutaneous fat biopsy (1-1.5 g) was obtained by needle biopsy, as described previously.26 The tissue pieces were rinsed rapidly in saline and submitted to lipolysis investigations. One 300-mg portion of the collected adipose tissue was frozen in liquid nitrogen and kept at −70°C for later gene expression studies. Another 300-mg portion of the collected adipose tissue was used for protein secretion. Finally, 1 300-mg portion of the collected adipose tissue was used for isolation of fat cells and lipogenesis studies (see below). We previously demonstrated that the tissue pieces removed and frozen in this way are free from damaged cells and blood.27 It should be pointed out that WAT samples were obtained by needle biopsies under local anesthesia preoperatively, and not during laparotomy, because we previously demonstrated that general anesthesia influences adipose tissue function.28
Isolation of Adipocytes From Adipose Tissue
Fat cells were isolated from the remaining components of the tissue according to the collagenase procedure described by Rodbell.29 In brief, tissue was cut into ∼20-mg pieces and incubated (1 g tissue/mL medium) in Krebs-Ringer phosphate (KRP) buffer (pH 7.4) containing 4% bovine serum albumin (BSA) and 0.5 mg/mL collagenase Type I for 60 minutes at 37°C in a shaking water bath. The isolated fat cells were collected on a nylon mesh filter and were washed 4 or 5 times with 0.1% KRP-BSA buffer. The purity of the isolation procedure was estimated by investigating 200 cells under a light microscope in each sample. The number of isolated cells that did not resemble fat cells or cell material that was stuck to a fat cell was always 0 to 2 per 200 counted cells.
Measurement of Fat Cell Weight, Volume, and Number
Mean fat cell volume and weight were determined on isolated fat cells as follows: The greatest dimension of a fat cell was determined during direct microscopy, and the greatest mean dimension of 100 cells in each individual was determined. The mean fat cell volume and weight were calculated by using the formulas developed by Hirsch and Gallian.30 The coefficient of variation for this method is 2% to 3%. The mean values are essentially the same as those determined from fat cell sizing of intact pieces of human adipose tissue.31 The total number of fat cells in the incubated sample was calculated as the lipid weight of the incubated fat cells divided by the mean fat cell weight. The total number of fat cells in the body was calculated as the amount of body fat divided by the mean fat cell weight. It is well known that mean fat cell volume and weight differ between various adipose regions in humans. However, the differences are small and introduce only a marginal error when a single depot is used for the calculation of total fat cell numbers, as discussed previously.32 In the whole cohort, total fat cell numbers (mean ± standard error) were 45 ± 4 × 109, which is in the same range as several other studies.31, 32 In a separate methodological study, we evaluated the bioimpedance method to determine body fat by comparing it with dual-x-ray absorbmetry (DEXA) in 7 men and 14 women with a body mass index (BMI) from 17 kg/m2 to 46 kg/m2 and ages 22 to 79 years. There was an excellent correlation between the 2 measures (r = .92; linear regression) with slope and intercept near 1.0 and zero, respectively. The coefficient of variation for the 2 measures was 3.6% in nonobese patients and 3.9% in obese patients (the group was divided at a BMI of 30 kg/m2). Taken together, these methodological data strongly suggest that the determination of total number of fat cells in the body is accurate.
Lipogenesis in isolated fat cells was performed as described in detail previously.33 In brief, isolated fat cells were incubated for 2 hours at 37°C in a buffer containing 3H-glucose without (basal) or with different concentrations of insulin. After incubation, lipids were extracted, and radioactive glucose incorporation into total lipid was used as an index of lipogenesis. The ability of insulin to stimulate basal lipogenesis was calculated. In addition, the half-maximum effective concentration (EC50) of insulin for each individual concentration response curve was determined.
Gene expression analyses were performed as described previously.23 Total RNA was extracted from 300 mg of adipose tissue using the RNeasy Mini Kit (Qiagen GmbH, Hilden, Germany). The Agilent 2100 Bioanalyzer (Agilent Technologies, Kista, Sweden) was used to confirm the integrity of the RNA. One microgram of total RNA from each sample was reverse transcribed to combinational DNA (cDNA) using the Omniscript RT Kit (Qiagen) and random hexamer primers (Invitrogen, Tastrup, Denmark). cDNA synthesis was performed simultaneously for all patients using the same mix of primers and the RT Kit. In a final volume of 25 μL, 5 ng of cDNA were mixed with 2 times SYBR Green polymerase chain reaction (PCR) Master Mix (Bio-Rad Laboratories Inc., Hercules, Calif) and primers (Invitrogen). The primer pairs were selected to yield a single amplicon based on dissociation curves and analysis by agarose gel electrophoresis. Messenger RNA (mRNA) was analyzed for IL-6 (NM_000600.2), TNFα (NM_000594.2), CD68 (NM_001251.2), and CD3D (NM_000732.4). Quantitative real-time PCR was performed in an iCycler IQ (Bio-Rad Laboratories Inc.) and was quantified using TaqMan kits (Applied Biosystem, Foster City, Calif) for Hs00174131_m1 (IL-6), Hs00174128_m1 (TNFα), Hs00154355_m1 (CD68), and Hs00174158_m1 (CD3D). Expression of mRNA was normalized to the 18S internal control by using a comparative Ct method. The patient with the highest Ct value was used as a reference from which all other Ct values for the target gene and reference gene, respectively, were subtracted from this Ct value. The Ct values then were normalized to ribosomal RNA for 18S using the formula 2ΔCt-target gene/2ΔCt-reference gene. Previous studies have demonstrated that these markers correlate well with infiltration of macrophages (CD68) and T-lymphocytes (CD3).34, 35
Protein Secretion From Adipose Tissue
TNFα and IL-6 secretion in the medium was measured exactly as described previously.23 In brief, a 300-mg portion of adipose tissue was incubated in an albumin-containing buffer for 2 hours at 37°C. After incubation, an aliquot of the medium was removed and stored at −70°C for subsequent analysis of TNFα and IL-6. The total amount of protein secreted was related to wet weight, lipid weight, and the number of fat cells in the incubated samples.
The reported values are the mean ± standard error or the median and range for SGA and tumor scores, which were not continuous variables. Results were compared using analysis of variance (ANOVA) and appropriate post hoc tests: Student t tests for paired or unpaired data or linear regression by the method of least squares. Kruskal-Wallis and Mann-Whitney tests (which are nonparametric) were used to compare SGA and tumor score values. A P value <.05 was considered statistically significant. The patients were recruited in a consecutive manner as 2 groups; weight-stable cancer and cancer cachexia. A power calculation was made before patient collection using Sample Power (SPSS Inc., Chicago, Ill). It was based on the 3 major measures in adipose tissue and on circulation in the current study (ie, gene expression, protein secretion, and circulating levels of TNFα and IL-6). Except for circulating TNFα and IL-6 levels, which were determined with methods other than those used in our study, to our knowledge, there have been no reports on the adipose tissue values in cancer patients. Therefore, we used previously published data obtained in individuals without cancer23 for the parameters detailed above, which were generated in our own laboratory using the same methods that we used in the current study. Lg-values for means and standard deviations were used to obtain a normalized distribution. We conducted a 2-group analysis, setting the weight-stable group values as the historic mean and standard deviation values and the cancer cachexia values as 1.3 times the historic values, thus setting the effect size at 30%.This effect size was chosen for 2 reasons. First, a smaller effect size would necessitate recruitment of a very large number of patients. Second, a dietary challenge caused a 25% to 30% significant change in the measured TNFα and IL-6 parameters in our previous study in individuals without cancer.23 According to our calculations, we could detect a 30% increase (or decrease) in the cachexia group compared with the other group at a P = .05 or better (2-tailed Student t test) for mRNA expression, protein secretion, and serum levels of TNFα and IL-6 if we recruited 14 patients in each group with a power ranging from 94% to 99% (mean power, 97%). We initially assumed that the distribution of patients should be equal so that we would obtain 14 patients in each group by recruiting 28 patients in a consecutive fashion. However, after recruitment, we ended up with 13 cachectic patients and 15 noncachectic patients; however, but this had no important effect on the power calculation when the tests were repeated on the actual number of patients. A postoperative reclassification revealed that 5 of the recruited patients did not have malignant disease, thus constituting a third group. Therefore, we performed a new power calculation for a 1-way fixed-effect ANOVA with 3 groups; cancer cachexia (n = 13 patients), weight-stable cancer (n = 10 patients), and no cancer (n = 5 patients). Again, we used the historic values described above for the weight-stable cancer group and 1.3 times those values for the cachectic group. For the noncancer group, we postulated a 30% lower value than the historic means. Thus, the relative values for the 3 groups were set at 1, 1.3, and 0.7 for the weight-stable cancer, cancer cachexia, and noncancer groups, respectively. The new calculation revealed that, for the different TNFα and IL-6 parameters, we would have an average power of 87% (range, 85%-93%) to detect a 30% effect size at P = .05 or better.
The initial determination of how many patients we had to recruit for this study was based on a set of assumptions that we calculated for each TNFα and IL-6 parameter and the correlation between effect size, statistical power, and the number of patients we needed to recruit. We used our previously published data mentioned earlier, assuming 1 control group and 1 case group of the same size (Fig. 1). The results indicated that, if we had chosen a smaller effect size (eg 20%), we would have had to recruit nearly 30 patients in each group to obtain a power of 97%. With 14 patients in each group (or 13 patients in 1 group and 15 patients in the other), the power would only be 80% to detect a 20% effect, and we would have obtained close to full power (100%) to detect an effect size of 40%. Figure 1 depicts data for historic circulating IL-6 values, but the curves were essentially the same for the other measured variables of TNFα and IL-6 (graphs not shown).
Patients enrolled in the study were divided into 3 groups after final diagnosis: a weight-losing cancer group (cachexia), a weight-stable cancer group, and a third noncancer group. Clinical characteristics are listed in Table 1. The groups were matched for age and sex distribution. Tumor severity scores were similar in the 2 cancer groups. The group with cachexia displayed a significantly lower BMI, adipose mass, fat cell volume, and plasma levels of transferrin and albumin than the weight-stable cancer group and the noncancer group. Insulin levels tended to be lower in the cancer cachexia group than in the other 2 groups (P = .056), but there was no in-between group difference in plasma glucose or hemoglobin. SGA scores were significantly higher in the weight-losing group compared with the other 2 groups. Furthermore, plasma glycerol and fatty acids per kg of body fat (an index of lipolysis) were increased significantly in the cachectic group compared with the other 2 groups. However, lean body mass and total fat cell numbers were similar all 3 groups. To determine whether the weight-losing group displayed any signs of increased systemic inflammation, serum levels of IL-6 (Fig. 2A) and TNFα (Fig. 2B) were measured. Whereas TNFα levels were similar in the 3 groups, there was a significant increase (P = .018) in IL-6 levels in the cachectic group.
|CC||WS||NC||Overall||CC vs WS†||CC vs NC†||WS vs NC†|
|Body mass index, kg/m2||22.2±0.6||26.5±1||28.0±6.2||.005||.0074||.0045||.45|
|Body fat, kg||16.8±1.0||24±1||31±5.4||.0002||.0056||<.0001||.034|
|Lean body mass, kg||54±3.9||58.4±5.3||49.4±4.8||.48||.45||.52||.24|
|Weight loss, % of habitual weight||11.7±1.8||1±1.2||3±1.8||.003||.001||.043||.42|
|Plasma glucose, mmol/L||6.1±0.4||6.8±0.6||7.3±1.1||.37||.36||.18||.55|
|Plasma insulin, mU/L||6.5±0.9||12.8±3.2||16.2±5.2||.056||.074||.031||.45|
|Plasma glycerol, μmol/L/kg body fat||9.8±2||3.3±0.3||3.2±0.4||.007||.0046||.017||.95|
|Plasma fatty acids, μmol/L/kg body fat||87±2||30±0.4||34±1||.014||.007||.038||.87|
|SGA score, points||11 (5–18)||2 (1–3)||4 (1–10)||.0004||NA||NA||NA|
|Tumor score||3 (1–4)||3 (1–4)||0||.57||NA||NA||NA|
|Fat cell volume, picoliters||339±30||574±40||678±113||.0003||.0013||.0003||.21|
|Fat cell number, ×109||42±6||45±2||49±10||.73||.73||.43||.61|
To determine whether there were local signs of increased inflammation in adipose tissue from the 3 groups, abdominal subcutaneous WAT was assessed for mRNA expression. In contrast to the elevated serum IL-6 levels, IL-6 mRNA levels did not differ between the groups (Fig. 3A) (P = .12; repeated-measures ANOVA). If anything, IL-6 expression tended to be lower in the cancer cachexia group. Similarly, no differences in TNFα expression could be observed (Fig. 3B) (P = .87). To assess whether there were signs of increased leukocyte infiltration, mRNA expression for CD68 (a macrophage marker) and CD3 (a T-lymphocyte marker) were measured. Similar to the findings with cytokine mRNA, the levels of CD68 (Fig. 3C) (P = .78) and CD3 (Fig. 3D) (P = .88) did not differ between groups.
To determine whether the secretion of IL-6 and TNFα from fat tissue corresponded to the findings at the mRNA level, in vitro incubations of WAT specimens were analyzed. The secretion rate of IL-6 did not differ significantly between specimens from the 3 groups irrespective of whether the rate was expressed per gram of wet weight fat tissue (Fig. 4A) (P = .65) or per 107 fat cells (Fig. 4B) (P = .36). Similarly, there was no significant difference in TNFα secretion expressed in either of the 2 ways (Fig. 4C,D) (P = .81 and P = .96, respectively). There was also no in-between group difference when protein secretion was expressed per lipid weight of adipose tissue (values not shown).
To assess possible alterations in lipogenesis, fat cells from WAT were isolated from the 3 groups and were assessed for the ability of insulin to stimulate the basal (no insulin present) incorporation of radioactive glucose into lipids. In these experiments, insulin stimulated lipogenesis in all 3 groups in a concentration-dependent manner (Fig. 5). The concentration-response curves in the 3 groups were superimposed almost completely. We also calculated the maximal insulin effect and EC50 in each patient. The maximal insulin effect in the 3 groups varied between 2.3- and 2.6-fold stimulation (P = .51). EC50 for insulin varied between 0.02 and 0.04 nmol/L in the 3 groups (P = .94). Thus, insulin-stimulated lipogenesis was almost identical in the 3 studied groups. Basal lipogenesis (expressed per number of isolated fat cells) did not differ between groups (P = .35).
In this study, we assessed the influence of cancer cachexia on the inflammatory status and adipocyte lipogenesis in subcutaneous adipose tissue. To include an additional feature of inflammation, we assessed adipocyte loss.
Cancer cachexia is regarded as an inflammatory condition.2 We were able to confirm that circulating IL-6 levels were increased significantly in cachexia. This was accompanied by clear signs of catabolism, as evidenced from several independent measures, including SGA scores and reduced albumin and transferrin levels in the cachexia group. However, we observed no changes in the mRNA expression or protein secretion (IL-6 and TNFα) of inflammatory markers in adipose tissue from cachectic patients. In addition, there were no signs of inflammatory macrophages or lymphocytes, as judged by CD68 and CD3 mRNA levels. Ideally, CD68- and CD3-positive cells also should have been determined (eg, by immunohistochemistry). However, the extremely small amounts of adipose tissue obtained by needle biopsy in these patients precluded such studies. Nevertheless, our data clearly suggest that, at least in patients with gastrointestinal cancer, cancer cachexia is not associated with an increased inflammatory reaction in adipose tissue. This implies that the increase in adipocyte lipolysis previously observed in cachexia13 is not dependent on an increased effect of locally produced cytokines. Furthermore, the finding that fat cell numbers were similar in all 3 groups rules out any significant increase in adipocyte cell death because of apoptosis and/or necrosis secondary to an inflammatory response in cancer cachexia. This finding, together with the demonstration that fat cells in cachectic patients display a normal insulin-stimulated lipogenesis, firmly establishes that increased lipolysis is the primary factor for fat mass loss in cachexia.
We recently elucidated the putative mechanisms underlying increased adipocyte lipolysis in cancer patients with cachexia.13 We were able to demonstrate that, as described above, the mRNA and protein expression of HSL were increased significantly in this condition. However, our data could not decipher the primary cause for these increased levels. Because our current results do not point to a change in adipokine secretion, local factors in adipose tissue most likely are not responsible for regulating HSL expression. Instead, it appears reasonable to speculate that tumor-secreted factors may stimulate HSL expression, although this remains to be established. In vitro coculture systems using human adenocarcinoma cells and fat cells may be an interesting model in which to study this possible effect.
In healthy individuals, subcutaneous WAT contributes significantly to in vivo circulating IL-6 levels in humans.36 Furthermore, diet-induced reductions in circulating IL-6 occur in parallel with reduced WAT secretion of the cytokine.23 Because IL-6 secretion was not increased in WAT from cachectic patients despite elevated circulating levels of the cytokine, it is apparent that fat tissue contributes to the circulating IL-6 levels only under certain conditions. In cancer cachexia, other sources that were not examined in this study (eg, the liver and pancreas) may play dominant roles in plasma IL-6, a hypothesis that is supported by previously reported findnings.37, 38 In addition, the tumor, or the host cells, or a combination of the 2 may be responsible for the production of proinflammatory cytokines. In a study of gastroesophageal cancer, both mRNA and protein expression of IL-1, IL-6, IL-8, and TNFα were significantly higher in the tumor than in the adjacent benign tissue.39
A prominent finding was the markedly smaller fat cells in cancer cachexia. Lipids constitute 95% of the fat cell volume. A decrease in cell volume can be caused by lipid synthesis (lipogenesis) or increased lipid mobilization (lipolysis). Our cachectic patients displayed a marked increase in in vivo lipolysis, in accordance with previous observations.3–7, 13 In contrast, adipocyte lipogenesis was unaltered. The latter finding rules out lipogenesis as a factor responsible for the reduced fat cell volume in cancer cachexia. The results on lipogenesis also contradict the notion that inflammation in cancer is associated with insulin resistance in adipose tissue. Both insulin sensitivity (EC50) and maximum insulin-induced lipogenesis were normal despite significantly higher levels of circulating IL-6, a cytokine that induces insulin resistance in human fat cells in vitro.40
We observed no difference in lean body mass between the groups, indicating that we studied patients in an early phase of cachexia, when there is loss predominantly of adipose tissue. A recent study demonstrated that body fat was lost more rapidly than lean tissue in progressive cancer cachexia.41
Cancer cachexia may have several etiologies, including malnutrition, anorexia and tumor-specific effects. Our studies were conducted on patients with gastrointestinal adenocarcinoma, predominantly pancreatic cancer, which is well known to associate with significant weight loss. Nevertheless, it is possible that the mechanisms underlying cancer cachexia may be different in patients with different types of gastrointestinal cancer. For example, some patients with esophageal cancer may have obstruction and dysfagia as a cause of weight loss. The current study was designed to assess differences between cancer patients with or without cachexia but was not powered to detect differences between selective cancer forms. Therefore, if we had had a significantly larger cohort, then we cannot exclude the possibility that a subdivision of patients according to tumor phenotype and possibly a stricter definition of cachexia than just unintentional weight loss could have resulted in other differences. However, in our experience13 it is very difficult to recruit patients with cancer cachexia to this type of study. It took us almost 3 years and 4 years, respectively, to recruit patients with cancer cachexia to the previous study13 and to the current study. Therefore, a multicenter study of these types of patients would be of great value. However, to our knowledge, other than our group, there are no investigators working with both cancer cachexia and detailed functional assessments of adipose tissue function.
In the initial power calculation, we chose an effect size of 30% for the 3 parameters measured for TNFα and IL-6, and we had very strong power (94%-99%) to detect such changes in each cytokine parameter investigated in cachectic patients versus weight-stable patients. This effect size was relevant biologically, because a dietary intervention caused a 25% to 30% change in several of the investigated TNFα/IL-6 parameters in WAT.23
In conclusion, the findings of the current study indicated that adipose tissue does not contribute to any significant degree to an inflammatory reaction in cancer cachexia. Increased lipolysis, and not reduced lipogenesis or increased fat cell death, appears to be the primary causative factor for fat mass loss in cancer cachexia.
- 25Sobin LH,Wittekind C, eds. TNM Classification of Malignant Tumors.6th ed. New York, NY: John Wiley & Sons, Inc.; 2002.
- 41Body composition and time course changes in regional distribution of fat and lean tissue in unselected cancer patients on palliative care—correlations with food intake, metabolism, exercise capacity, and hormones. Cancer. 2005; 103: 2189–2198., , , , , .