Elevation of cytokine levels in cachectic patients with prostate carcinoma
Abstract
BACKGROUND
Approximately 60–70% of patients with advanced prostate carcinoma (CaP) suffer from cachexia, one of the most devastating conditions associated with advanced malignant disease. The pathophysiology of cachexia is multifactorial, and several cytokines, such as tumor necrosis factor α (TNFα) and interleukin 1 (IL-1), IL-6, and IL-8, may be involved. The objective of the current study was to determine whether cachexia associated with advanced CaP is accompanied by increased serum levels of TNFα, IL-1β, IL-6, and IL-8.
METHODS
The levels of TNFα, IL-1β, IL-6, IL-8, and prostate specific antigen (PSA) were examined in serum samples from normal donors (n = 10 donors), from patients with organ-confined CaP (n = 19 patients), from patients with advanced CaP without cachexia (n = 17 patients), and from patients with cachectic CaP (n = 26 patients). DPC Immulite and Abbott IMx Total-PSA assays were used to determine cytokine and PSA levels, respectively.
RESULTS
Levels of TNFα, IL-6, and IL-8 were elevated significantly in the group of patients with advanced, cachectic CaP compared with patients who were without cachexia. In the cachectic patients, levels of TNFα were correlated positively with IL-8, and there was no correlation between PSA levels and any of the cytokine levels. IL-1β levels were below the limit of detection in all samples.
CONCLUSIONS
The current results show that levels of TNFα, IL-6, and IL-8 were increased in CaP patients with cachexia. Increased levels of these cytokines were not correlated with PSA levels, suggesting that they are regulated by a mechanism that is independent of PSA synthesis. Additional fundamental research is needed to determine the mechanisms involved and to identify potential therapeutic targets in patients with cachexia. Cancer 2003;97:1211–6. © 2003 American Cancer Society.
DOI 10.1002/cncr.11178
Cancer cachexia is one of the most serious conditions associated with advanced malignant disease and has a substantial impact on the quality of life and survival of affected patients; it is suspected that ≈ 20% of patients with malignant disease die from the effects of cachexia rather than tumor burden.1 Cachexia is characterized by weight loss and anorexia that involve the depletion of host adipose tissue and skeletal muscle.2 Nutritional supplementation has little or no impact on cachexia. Cachexia is associated with a lower quality of life, a poorer response to chemotherapy, and reduced survival.3, 4 Cachexia is of especially high significance in patients with prostate carcinoma (CaP): 60–70% of patients with advanced CaP suffer and die from cachexia.1 However, the etiology of cachexia in patients with CaP, as in patients with other types of malignant disease, is not well understood, and intervention strategies to alleviate the condition are not yet available.
The pathophysiology of cachexia associated with progression in patients with malignant disease is multifactorial. Several cytokines, including tumor necrosis factor α (TNFα) and interleukin 1 (IL-1), IL-6, and IL-8, have been proposed as agents that cause or contribute to the condition.3, 5-8 Oliff et al.9 demonstrated that TNFα induced a cachexia-like syndrome in mice, and other studies have supported these observations.10, 11 Increased TNFα levels were also associated with weight loss in patients with pancreatic carcinoma12 and gastrointestinal carcinoma.13 Nakashima et al.14 reported that TNFα contributed to the complex syndrome of cachexia in patients with CaP: TNFα was detected in 75% of patients with recurrent disease, and the body mass indices of patients with elevated serum TNFα levels were significantly lower compared with the corresponding values in patients with undetectable serum TNFα levels. However, some clinical studies failed to correlate TNFα levels with the cachectic state.15-17
IL-6 is a powerful cytokine that mediates several cellular events. Strassmann et al.18 reported that IL-6 caused cachexia in an animal model, and recent studies in transgenic mice have demonstrated that IL-6 participates in muscular atrophy.19 Again, translation from the murine model to humans has not been straightforward, and the evidence of a role for IL-6 in clinical cachexia is mixed.20-22 In patients with CaP, Adler et al.23 detected significantly higher IL-6 levels in patients with clinically evident metastases. Shariat et al.24 reported increased levels of IL-6 in patients with CaP who had bone metastases, and Wise et al.25 reported that patients who had hormone-refractory CaP had significantly higher IL-6 levels compared with patients who had hormone-sensitive tumors. Similar observations were reported by De Vita et al.,26 who detected elevated levels of IL-6 in patients with gastrointestinal malignancies that were correlated negatively with overall survival and time to disease progression. However, none of those studies divided patients with advanced disease into cachectic and noncachectic groups.
It has also been shown that IL-8 is associated with the depletion of adipose tissue in vitro. Gerhardt et al.27 showed that prolonged stimulation of cultured human adipocytes with exogenous cytokines, including IL-8, leads to a decrease in lipid content. Ferrer et al.28 reported the presence of IL-8 in tissue sections from patients with CaP, but not in patients with benign prostatic hyperplasia. Kim et al.29 and Inoue et al.30 showed that expression of IL-8 in human CaP cells is associated with angiogenesis, tumorigenicity, and metastasis in vitro and in vivo. In our collaborative study with Veltri et al.,31 levels of IL-8 were 11-fold higher in patients with metastatic CaP compared with control patients who had benign prostatic hyperplasia. In patients with melanoma, Scheibenbogen et al.32 detected a significant correlation between IL-8 levels and tumor burden. However, there have been no clinical studies of IL-8 levels in cachectic patients.
It has been suggested that IL-1 may play a role in cachexia based on its similarity to TNFα.33, 34 However, neutralization of IL-1 by antibodies,35 administration of an IL-1 receptor antagonist,36 and transfection of an IL-1 receptor antagonist37 failed to prevent cachexia in mice. The objectives of the current study were to determine the levels of TNFα, IL-1β, IL-6, and IL-8 in patients with CaP and to determine whether higher levels of these cytokines are associated with cachexia.
MATERIALS AND METHODS
Serum Samples and Assays
Serum specimens were obtained from consented patients at the University of Washington Hospital and the Seattle Veterans Administration Hospital between May 1997 and October 2000. Samples were processed within 6 hours of blood draw. The blood was allowed to clot at room temperature for 15 minutes and was stored at 4 °C until centrifugation at 3500 rpm for 15 minutes. After processing, all samples were aliquoted and stored immediately at −80 °C. We and others have observed38, 39 that sera stored at −70 °C and lower for 9 months or more showed no significant differences in prostate specific antigen (PSA) levels compared with fresh samples.40 Similar observations have been made regarding the stability of cytokine serum levels.41 The specimen aliquots had not been thawed prior to this study. TNFα, IL-1β, IL-6, and IL-8 DPC Immulite assays (gifts of Diagnostic Products Corporation, Los Angeles, CA) and the IMx Total-PSA assay (gift of Abbott Laboratories, Abbott Park, IL) were used to determine serum cytokine levels and PSA levels, respectively. All assays were performed according to the manufacturer's instructions.
Patient Groups
Serum levels of TNFα, IL-1β, IL-6, IL-8, and PSA were determined in 72 serum samples from patients with CaP in various stages of the disease and from a control group of volunteers.
Group A: control group
Group A included 10 volunteers with no history of benign or malignant prostatic disease. The median age of this group was 55.0 years.
Group B: organ-confined CaP
Group B included 19 patients with newly diagnosed, organ-confined CaP. The samples were obtained prior to primary therapy. None of the patients in this group suffered from peripheral edema, weight loss, appetite loss, or anemia. The median age was 65.0 years.
Group C: advanced CaP
Group C included 43 patients with advanced CaP. Their median age was 70.0 years. For further analysis, this group was divided into two subgroups: Group Cα and Group Cβ.
Group Cα: advanced CaP without cachexia.
Clinical information on the 17 patients in Group Cα is shown in Table 1. Weight loss was seen in 4 of 17 patients (24%), with a mean ± standard error weight loss of 0.94 ± 0.45 kg (median, 0 kg; range, 0–6.2 kg). Despite the weight loss in some patients, they were not considered cachectic according to their clinical status, as determined by the attending oncologist (see the criteria outlined below for Group Cβ).
| Characteristic/treatment | Group Cα | Group Cβ |
|---|---|---|
| No. of patients | 17 | 26 |
| Median age (yrs) | 72 | 68 |
| Radical prostatectomy (%) | 29 | 27 |
| Prostatic radiation therapy (%) | 24 | 31 |
| Stage D2 (%) | 100 | 100 |
| Androgen suppression (%) | 88 | 96 |
| Chemotherapy (%) | 24 | 58 |
| Palliative radiation therapy (%) | 0 | 46 |
| Bisphosphonate treatment (%) | 35 | 12 |
| Peripheral edema (%) | 24 | 62 |
| Anemia treatment (%) | 24 | 42 |
| Corticosteroid treatment (%) | 41 | 46 |
| Median albumin level (g/dL) | 3.45 | 3.30 |
- Group Cα: noncachectic group; Group Cβ: patients with cachectic prostate carcinoma.
Group Cβ: advanced CaP with cachexia.
Clinical information on the 26 patients in Group Cβ is shown in Table 1. The decision whether a patient with advanced CaP was cachectic or noncachectic was based on his clinical status and was determined by the oncologist prior to the start of this study: A patient was determined to be cachectic when he presented with loss of weight over a 3–6 month period and loss of appetite (anorexia). If both symptoms were absent, then the patient was considered noncachectic. If only one of these symptoms was detected, then symptoms such as low energy level, fatigue, and low exercise tolerance were evaluated. If one of those symptoms was positive, then the patient was categorized as cachectic. Only one patient was placed in the cachectic group on the basis of the secondary criteria. That patient died 2 months after the blood sample was taken with further disease progression. All patients in this group had lost weight. The mean weight loss ± standard error was 7.1 ± 0.95 kg (median, 6.0 kg; range, 1.5–21.9 kg). Appetite loss was observed in 25 patients.
Statistical Methods
All statistical analyses were performed using GraphPad Prism software (version 3.0). Comparisons between the serum cytokine levels of the different groups were performed by the Mann–Whitney test. Correlations among the different markers were calculated using the Spearman rank correlation. For statistical evaluation, we set the cytokine levels in the nondetectable samples to half the value of the limit of detection of each assay.
RESULTS
The current results are presented in Figure 1 and are summarized in Table 2. IL-8 levels were elevated significantly in patients with advanced CaP compared with healthy volunteers and patients with organ-confined CaP, whereas levels of IL-6 and TNFα were higher in patients with advanced CaP but did not reach significance. Importantly, levels of TNFα, IL-6, and IL-8 were increased in cachectic patients (Group Cβ) compared with noncachectic patients with advanced CaP (Group Cα) (TNFα, P = 0.031; IL-6, P = 0.002; IL-8, P = 0.018). The IL-1β levels were not detectable in any of the groups with the assay used (the limit of detection was 5 pg/mL). PSA levels exhibited significant pairwise differences among all groups (P < 0.0001) except Cα to Cβ (P = 0.019). The albumin levels in patients with cachectic and noncachectic advanced CaP were not significantly different (P = 0.106).
Levels of (A) prostate specific antigen (PSA), (B) tumor necrosis factor α (TNFα), (C) interleukin 6 (IL-6), and (D) IL-8 during disease progression in patients with prostate carcinoma. Group A, control samples (n = 10 patients); Group B, patients with organ-confined prostate carcinoma (n = 19 patients); Group C, patients with advanced prostate carcinoma (n = 43 patients); Group Cα, patients with noncachectic advanced prostate carcinoma (n = 17 patients); and Group Cβ, patients with advanced prostate carcinoma with cachexia (n = 26 patients). Results are plotted as the mean ± standard error of the mean. Levels of IL-1β were below the limit of detection in all samples.
| Level | Control (A) | Localized CaP (B) | Advanced CaP (C) (Cα + Cβ) | Noncachectic CaP (Cα) | Cachectic CaP (Cβ) | P value (Mann–Whitney test; Cα vs. Cβ)b |
|---|---|---|---|---|---|---|
| PSA (ng/mL) | 0.89 ± 0.16 | 9.67 ± 2.09 | 1406 ± 483.1 | 336.8 ± 92.95 | 2104 ± 771.7 | 0.019 |
| TNFα (pg/mL) | 3.16 ± 0.49 | 5.70 ± 0.84 | 5.12 ± 1.04 | 2.95 ± 0.45 | 6.54 ± 1.66 | 0.032 |
| IL-1β (pg/mL) | < 5.0 | < 5.0 | < 5.0 | < 5.0 | < 5.0 | n.p. |
| IL-6 (pg/mL) | 7.24 ± 1.00 | 8.58 ± 0.60 | 22.81 ± 7.52 | 7.32 ± 1.02 | 32.94 ± 12.10 | 0.002 |
| IL-8 (pg/mL) | 4.0 ± 0.63 | 5.58 ± 0.80 | 43.41 ± 14.24 | 10.84 ± 1.46 | 64.70 ± 22.74 | 0.019 |
- CaP: prostate carcinoma; Cα: patients with noncachectic CaP; Cβ: patients with cachectic CaP; PSA: prostate specific antigen; TNFα: tumor necrosis factor α; group; IL-1β; interleukin 1β; n.p.: not performed.
- a TNFα, IL-1β, IL-6, and IL-8 DPC Immulite assays and the Abbott IMx-Total PSA assay were used for analyses. Results are presented as the mean ± standard error of the mean.
- b The Mann–Whitney test was used to calculate the significance of the differences between Group Cα and Group Cβ.
Because weight loss is one of the hallmarks of cachexia, we compared the weight loss in Group Cα with the weight loss in Group Cβ and observed, as expected, that the patients with cachexia lost significantly more weight compared with the patients without cachexia (P < 0.0001). Our analysis also showed that the patients in Group C with weight loss had significantly higher levels of IL-6 compared with the patients who did not loose weight (P = 0.034). Also in Group C, levels of TNFα were correlated with weight loss (correlation coefficient [r] = 0.353; P = 0.020).
Because more of the patients with cachexia had received chemotherapy or palliative radiation (see Table 1), we examined the possibility that the increased levels of TNFα, IL-6, and IL-8 were caused by the treatments administered to these patients. We compared levels of TNFα, IL-6, IL-8, and PSA in cachectic patients and noncachectic patients who had not received chemotherapy or radiation. The results are summarized in Table 3. In addition, we compared levels of TNFα, IL-6, and IL-8 in cachectic patients who had received chemotherapy or radiation with levels in patients who did not receive either chemotherapy or radiation and detected no significant differences.
| Group Cα vs. Group Cβ | P value | |||
|---|---|---|---|---|
| PSA | TNFα | IL-6 | IL-8 | |
| No radiation | 0.025 | 0.155 | 0.020 | 0.191 |
| No chemotherapy | 0.4173 | 0.0289 | 0.008 | 0.040 |
- Group Cα; patients with noncachectic prostate carcinoma; Group Cβ: patients with cachectic prostate carcinoma; PSA; prostate specific antigen; TNFα: tumor necrosis factor α; IL-6: interleukin 6.
- a To evaluate whether changes in levels of cytokines were observed in the absence of variations in treatment status, levels of TNFα, IL-6, and IL-8 were compared in patients with advanced prostate carcinoma who had cachectic disease (Group Cβ) and noncachectic disease (Group Cα) and who did not receive palliative radiation therapy or chemotherapy using the Mann–Whitney test.
Our data showed that, in Group Cβ, high levels of TNFα were correlated significantly with IL-8 (r = 0.411; P = 0.037). In Group C, higher levels of TNFα also were correlated with higher levels of IL-6 (r = 0.324; P = 0.034) and IL-8 (r = 0.483; P = 0.001). There were no significant correlations between TNFα and the other cytokines (IL-1β, IL-6, and IL-8) in Groups A, B, and Cα. PSA levels were not correlated significantly with levels of TNFα, IL-1β, IL-6, and IL-8 in any of the patient groups.
DISCUSSION
Effective treatment of cachexia is likely to come about only when the causative factors have been identified, although this effort is hampered by a lack of animal models that are truly analogous to human cachexia. However, data from various sources, including some animal studies, have suggested the involvement of certain cytokines in cachexia. Thus, on the basis of earlier studies, we examined patients with CaP who had cachexia for elevated levels of cytokines that have high a priori probabilities of involvement in cachexia.
In our study, serum levels of TNFα, IL-6, and IL-8 were elevated significantly in patients with advanced CaP who had cachexia compared with patients who were without cachexia. We also found pairwise correlations between levels of TNFα and IL-8, suggesting a coordinated regulation of the expression of these cytokines.
The cachectic patients in Group Cβ had higher serum PSA levels compared with patients in Group Cα. Thus, we wanted to test the hypothesis that the higher levels of cytokines observed were associated directly with the extent of disease, as reflected in the PSA measurement, rather than the cachectic status. We found that PSA levels were not correlated with levels of any of the cytokines measured in Group Cβ, suggesting that disease extent is not the major factor affecting levels of these cytokines. Examination of the subsets of cachectic and noncachectic patients who were not treated with chemotherapy indicated that levels of PSA were not significantly different between these two groups, although the levels of TNFα, IL-6, and IL-8 remained significantly higher. Conversely, in the subset of patients who did not receive palliative radiation therapy, cachexia was associated with higher PSA levels; whereas cytokine levels were not significantly different, except for IL-6 (Table 3), again indicating a lack of association between PSA levels and levels of the cytokines studied. However, it should be noted that PSA does not necessarily track tumor volume, especially in patients with advanced CaP.
We also examined whether the increases in levels of cytokines detected in cachectic patients were related to the treatments administered. Our results showed that neither chemotherapy nor palliative radiation caused increased levels of cytokines in Group Cβ.
According to our data, IL-1β is not involved in cachexia in patients with advanced CaP. These results confirm reports from in vivo studies in which cachexia could not be prevented by neutralization of IL-1 by antibodies,35 administration of an IL-1 receptor antagonist,36 or transfection of an IL-1 receptor antagonist37 in mice.
It should also be noted that there may be significant differences between the development of cachexia in patients with CaP, which metastasizes predominantly to the bone marrow, and cachexia associated with gastrointestinal tumors (e.g., pancreatic or colon carcinoma), which usually metastasize to the liver and, thus, may lead to symptoms associated with impaired hepatic function or direct involvement of the gastrointestinal tract. The impaired liver function generally is followed by a decreasing protein concentration, including albumin, and causes severe problems, such as edema and bleeding, which contribute to the condition of cachexia. Because CaP rarely spreads to the gastrointestinal tract, cachexia in patients with CaP most likely is triggered by a pathway independent of that seen in patients with gastrointestinal tumors. This is supported by our results showing that albumin levels in our patient groups were not significantly different in cachectic and noncachectic patients.
We hypothesize that there are two (or more) biochemical pathways that can lead to cachexia, only one of which involves elevation of the tested cytokines. For example, proteins like proteolysis-inducing factor (PIF42), which has been shown to cause cachexia in animals, and human cachexia associated protein (HCAP), a probable human analog of PIF, may be involved in one or more of these other pathways. Expression of HCAP in CaP and its potential involvement in CaP-associated cachexia are under investigation in our laboratory.
In conclusion, our results show that there are significantly higher levels of TNFα, IL-6, and IL-8 in patients with CaP who have cachexia. We hypothesize that these three elevated cytokines may play roles in the development of cachexia in patients with advanced CaP. However, further investigations are needed to clarify their putative roles.
