Fax: (011) 31-20-6968833
Plasma cytokine and P-selectin levels in advanced malignancy
Prognostic value and impact of low–molecular weight heparin administration
Version of Record online: 7 OCT 2005
Copyright © 2005 American Cancer Society
Volume 104, Issue 10, pages 2275–2281, 15 November 2005
How to Cite
Di Nisio, M., Niers, T. M., Reitsma, P. H. and Buller, H. R. (2005), Plasma cytokine and P-selectin levels in advanced malignancy. Cancer, 104: 2275–2281. doi: 10.1002/cncr.21485
- Issue online: 31 OCT 2005
- Version of Record online: 7 OCT 2005
- Manuscript Accepted: 8 JUL 2005
- Manuscript Revised: 1 JUN 2005
- Manuscript Received: 29 NOV 2004
- innate immunity;
- low molecular weight heparin
The survival benefit described in patients with cancer treated with low molecular weight heparin (LMWH) may result from a LMWH-mediated effect on the immune system or on the cross-talk between platelets and tumor cells.
Plasma levels of interleukin (IL)-10, IL-6, interferon (IFN)-γ, and P-selectin were measured in patients with advanced stage malignancy who were randomized to receive standard cancer care with or without the addition of LMWH.
Patients with IL-6 levels above the median had a median survival of 6.5 months versus 8.8 months for those with values below this cutoff (P = 0.02). IL-10 levels were found to be similarly correlated with survival such that IL-10 concentrations above the detection limit of the assay were associated with a doubled risk of dying in comparison to undetectable IL-10 (P = 0.02). No significant association was found between survival and circulating levels of IFN-γ. For P-selectin, patients with values below the fourth quartile had a median survival of 8.8 months versus 6.5 months for patients with levels above the fourth quartile (P = 0.02). In multivariate analysis, IL-10 remained an independent unfavorable prognostic factor (hazard ratio, 2.13; 95% confidence interval, 1.08–4.20). In patients treated with LMWH, the plasma levels of IL-6, IL-10, IFN-γ, and P-selectin demonstrated similar correlations with survival. However, none of the markers was associated with the beneficial survival effects observed with the administration of LMWH.
IL-10, IL-6, and P-selectin levels predicted a poor outcome in patients with advanced stage malignancy. The prolongation in survival observed with LMWH in patients with cancer apparently cannot be explained by a LMWH effect on these circulating markers. Cancer 2005. © 2005 American Cancer Society.
Inflammation and innate immunity are considered essential in the defense against cancer.1–4 Several studies have suggested that host responses are often defective in patients with cancer favoring, rather than opposing, the progression of the tumor.1, 2 Both tumors and innate immunity cells can produce immunomodulating agents that divert the host-protective mechanisms and suppress tumoricidal activity leading to a predominant humoral immunity, which is ineffective against the tumor.1–4
Accumulating data suggest that the plasma levels of some cytokines might reflect the activity of the immune system against the tumor and correlate to the extent of disease and prognosis.5–19 Interleukin (IL)-6, for example, can stimulate cell growth and angiogenesis, and induce resistance to therapy in cancer cells1, 3, 20 and high serum IL-6 levels have been found to predict a poor clinical outcome.5–8, 10–16 In addition, animal studies have shown an impairment in the host response against the tumor due to an abnormal production of IL-10 by the malignant and host immune cells.21, 22 To our knowledge, he prognostic value of IL-10 levels in humans remains uncertain with conflicting data reported in the literature.9, 17–19, 23 Similarly, the possible prognostic value of other cytokines, such as interferon (IFN)-γ, that promote a cell-mediated immune response, has not been clearly established to the best of our knowledge.24–26
Recently, a role for low molecular weight heparin (LMWH) in the management of patients with cancer has been claimed after several clinical studies demonstrated a prolongation of survival in patients with cancer who were treated with LMWH in addition to standard cancer care.27–32 The beneficial effects of LMWH on survival could be related to an effect on the host immune response, although data are limited, with discordant results published across the studies, possibly due to differences in experimental conditions.33–38
Another possible mechanism with which to explain the anticancer activity of LMWH could be that LMWH interferes with the cross-talk between platelets and tumor cells. Platelets have the potential to promote several steps of the tumor progression and markers of platelet activation have been correlated with a worse prognosis in patients with cancer.39–41 It has been suggested that heparin can inhibit tumor metastasis by blocking P-selectin,42 a marker of platelet activation that has been associated with mortality and recurrent disease risk in patients with malignancy.43
The aim of the current study was, first, to evaluate whether the plasma levels of P-selectin, IL-6, IL-10, and IFN- γ predict survival in patients with advanced stage cancer. Second, we assessed whether the levels of these markers respond to LMWH treatment.
MATERIALS AND METHODS
Plasma samples were obtained from patients participating in the malignancy and low molecular weight therapy (MALT) trial.28 In that study, 302 patients without signs or symptoms of venous thromboembolism and with a diagnosis of advanced stage solid malignant tumor that was not curable with the standard available treatment were randomized to receive a 6-week cycle of subcutaneous nadroparin (Sanofi-Synthelabo, Paris, France) or placebo. The characteristics of the MALT patients have been described in more detail elsewhere.28 Briefly, patients with a life expectancy < 1 month, an indication for anticoagulant treatment, a contraindication for LMWH, thrombocytopenia (defined by a platelet count of < 50,000 platelets/mm3), or who were pregnant were excluded from the study. At baseline, data were collected concerning the demographic characteristics, as well as information regarding the type, histology, stage, and duration of cancer. Moreover, the World Health Organization (WHO) performance status and the physician's assessment of life expectancy (< 6 mos vs. ≥ 6 mos) were determined. Patients were followed until death or until the end of the study, with a median follow-up of 12 months. In the intention-to-treat analysis, treatment with nadroparin was associated with a significantly prolonged survival compared with placebo with the greatest effects noted in the subgroup of patients with a better prognosis at baseline. The overall hazards ratio (HR) for mortality was 0.75 (95% confidence interval [95% CI], 0.59–0.96) in favor of LMWH.
In the current analysis, IL-6, IL-10, IFN-γ, and P-selectin levels were determined in an unselected group of 141 patients from the MALT study for whom plasma samples were available. Of these patients, 75 were randomized to receive nadroparin.
We sought to evaluate 1) the prognostic value for survival of circulating levels of IL-6, IL-10, IFN-γ, and P-selectin in all the 141 patients at the time of entry into the study; 2) the association between these circulating markers and prognosis in the group of patients treated with LMWH; and 3) whether the beneficial survival effects observed in the MALT study were related to the influence of LMWH on plasma levels of soluble P-selectin or cytokines. This latter effect could be reflected by changes in immune mediators and plasma concentrations. Given the role of IL-6, IL-10, and IFN-γ in the host response and in the promotion of tumor progression, an increase in IFN-γ and a decrease in IL-6 and IL-10 was hypothesized to result from the administration of LMWH.
Blood Sampling and Sample Analysis
At the start of study treatment (Time 0) and at 6 weeks (the end of LMWH treatment phase), a blood sample was obtained and anticoagulated with sodium citrate (0.109 M, 1/10 volume/volume). Platelet-poor plasma samples were frozen in small aliquots and stored at −70 °C until analysis. IL-6, IL-10, and IFN-γ levels were measured with the Bio-Plex cytokine assay (Bio-Rad, Veenendaal, The Netherlands). The detection range was 0.49––32,000 pg/mL. P-selectin plasma levels were measured by DuoSet enzyme-linked immunoadsorbent assay (R&D Systems, Abingdon, U.K.) with a detection limit of P-selectin concentrations of 0.21 ng/mL.
The chi-square and the Mann–Whitney U tests were used for descriptive purposes. Survival estimates were calculated according to the Kaplan–Meier method with the analysis based on the time from randomization to death. Patients alive at the end of follow-up were censored. The Cox regression model was used to adjust for potential confounding variables: life expectancy (< 6 mos vs. ≥ 6 mos), WHO performance status (≤ 1, 2, ≥ 3), concomitant treatment (chemotherapy, radiotherapy [RT], hormonal therapy, or other antineoplastic treatment), type of carcinoma (breast, colorectal, cervical, or other), and histology (adenosquamous carcinoma, squamous carcinoma, or other). Ninety-five percent CIs were calculated when appropriate. IL-6, IL-10, IFN-γ, and P-selectin demonstrated a nonnormal distribution and therefore median values were calculated. Because plasma IL-10 was detectable in only a fraction of patients, the predictive value of IL-10 was assessed with IL-10 as a dichotomous variable taking the value 0 when below the detection limit or 1 otherwise. Finally, the association between IL-6, IL-10, IFN-γ, and P-selectin levels and prognosis as well as the effects of LMWH on these circulating markers (Mann–Whitney–Wilcoxon test) were evaluated in the group of patients treated with nadroparin.
The statistical analysis was performed using the SPSS package for Windows, version 11.0 (SPSS Inc., Chicago, IL).
The Entire Study Group
The characteristics of the study population are detailed in Table 1. The median levels at the time of entry into the study for IL-6, IL-10, IFN-γ, and P-selectin were, respectively, 9.4 pg/mL (range, 0.6–438.8 pg/mL), 1.2 pg/mL (range, 0.6–24.1 pg/mL), 3.6 pg/mL (range, 2.0–322.2 pg/mL), and 4.3 ng/mL (range, 0.7–11.9 pg/mL) (Table 1).
|Characteristics||No. of patients|
|Age at inclusion in yrs (range)||62.3 (38.4–85.7)|
|Weight in kg||73.2 ± 11.5|
|Mos of follow-up (range)||7.2 (0.6–51.1)|
|Mos of cancer duration at baseline (range)||16 (0–217)|
|Mos of metastasis duration at baseline (range)||5 (0–84)|
|0 or 1||90.8%|
|3 or 4||2.1%|
|< 6 mos||48|
|≥ 6 mos||52|
|IL-6 in pg/mL (range)||9.4 (0.6–438.8)|
|IL-10 in pg/mL (range)||1.2 (0.6–24.1)|
|IFN-γ in pg/mL (range)||3.6 (2.0–322.1)|
|P-selectin in ng/mL (range)||4.3 (0.7–11.9)|
The plasma levels of IL-6 predicted a shorter survival, with both the median and the quartiles of the IL-6 distribution dividing patients into groups with a significant difference in prognosis. The median survival for patients with IL-6 concentrations above the median was 6.5 months compared with 8.8 months for patients with IL-6 values below this cutoff (P = 0.02). In other terms, the risk of dying was 56% higher in patients with IL-6 values above the median (HR = 1.56; 95% CI, 1.05–to 2.30) (Fig. 1). It is noteworthy that in the group of patients who were still alive at the end of the 6-week study treatment period (n = 73), IL-6 maintained this prognostic power (8.0 mos vs. 11.0 months, respectively; P = 0.007). The association of IL-6 with an adverse outcome was even more remarkable when the analysis included the extreme values of the IL-6 distribution. The median survivals in patients with IL-6 levels above and below the fourth quartile (24.3 pg/mL) were remarkably different (7.3 mos and 13.4 mos, respectively; P = 0.018).
Circulating IL-10 was similarly correlated with a poor prognosis. In patients in whom IL-10 was detectable, the median survival was lower than in patients with IL-10 values below the detection limit of the assay (3.3 mos vs. 8.2 months; P = 0.02). This difference corresponded to a risk of death that was 2 times higher in patients with measurable IL-10 (HR = 2.14; 95% CI, 1.10–3.89) (Fig. 2A). At the end of the study treatment period, the number of patients with IL-10 within the range of the assay was too small to assess the IL-10 predictive value (n = 5).
No association with survival was evident using the median IFN-γ levels as a cutoff point (P = 0.37). It is noteworthy that prognosis tended to improve with increasing IFN-γ quartiles, and IFN-γ concentrations above the fourth quartile (5.1 pg/mL) were related to a longer median survival (13.4 mos) compared with lower levels (8.4 mos) (P = 0.78).
The median survival was comparable between patients with P-selectin levels above or below the median (8.0 mos vs. 8.2 mos; P = 0.61). Significant differences in prognosis were evident at the extremes of the P-selectin distribution. In particular, patients with P-selectin concentrations higher than the fourth quartile (5.4 ng/mL) had a shorter survival (6.5 mos) than those with lower values (8.8 mos) (HR = 1.72; 95% CI, 1.1–2.7). Such association with poor prognosis was, however, not found to be statistically significant at the end of the study treatment (8.0 mos vs. 10.1 mos, respectively; P = 0.46).
In a univariate analysis, the study treatment (nadroparin or placebo); patient weight; a life expectancy ≥ 6 months; the treatment received during the study period (chemotherapy, surgery, RT, or hormone treatment); and IL-10, IL-6, and P-selectin levels were all found to be significantly associated with survival. When adjusting in a Cox multivariate model for all possible confounders, IL-10 remained an independent prognostic marker (HR = 2.13; 95% CI, 1.08–4.20). After regression analysis, the predictive value of both IL-6 and P-selectin levels was maintained, but was no longer statistically significant (HR =1.44 [95% CI, 0.96–1.44] and HR = 1.65 [95% CI, 0.99–2.73]).
Patients Treated with LMWH
In the group of patients randomized to receive nadroparin, a similar inverse relation with prognosis was evident for IL-6, IL-10, and P-selectin at baseline, as in the whole study group. A shorter median survival was observed for IL-6 above the median and above the fourth quartile in comparison to lower levels (6.4 mos vs. 10.1 mos [P = 0.10] and 3.3 mos vs. 8.8 mos [P = 0.04], respectively). In the same way, IL-6 concentrations at the end of the study treatment period represented an unfavorable prognostic marker (P = 0.001).
Patients with detectable IL-10 were found to have a poorer outcome than those in whom this cytokine was not detectable (median survival of 3.0 mos vs. 8.8 mos; P = 0.0008) (Fig. 2B). For both IFN- γ and P-selectin, there was no difference in prognosis noted with any of the cutoffs. On multivariate analysis, IL-10 remained a predictor of poor prognosis (HR = 10.8; 95% CI, 2.99–39.4). In contrast to what was expected, none of the circulating markers evaluated was affected by nadroparin and none was found to be correlated with the survival benefit that was evident in the patients randomized to receive LMWH. Surprisingly, an increase in IL-6 from 8.1 pg/mL at baseline to 10.4 pg/mL at the end of the study treatment was observed in patients who received nadroparin (P = 0.03). Although IL-10 and P-selectin concentrations were basically unchanged after the administration of nadroparin, plasma IFN-γ levels increased when compared with baseline (1.7 pg/mL vs. 3.6 pg/mL; P = 0.48).
Detectable IL-10 predicted shorter survival in patients with advanced stage malignancy, and also after correction for other potentially confounding variables. These results are in keeping with previous data that suggested an involvement of IL-10 in the immune escape mechanisms of the tumor,21, 22 and a prognostic value of plasma IL-10 levels in patients with cancer.9, 17–19 The correlation between IL-10 and a worse outcome, however, has not been always consistent.5 Differences in the spectrum of included cancers and/or in the disease severity likely explain these discrepant results. The size of the current study sample did not allow us to perform a subgroup analysis for cancer type. However, when correcting for tumor type and tumor histology in the multivariate analysis, the relation between IL-10 and survival remained unchanged, suggesting a similar prognostic role for IL-10 across different types of cancer.
Circulating IL-6 has been found to be associated with an adverse outcome in a variety of tumors.5, 6, 10, 12–16 In agreement with the available literature, the current analysis found shorter survivals in those patients with high IL-6 levels (Fig. 1). The progression of the malignancy could be promoted by IL-6 in several ways, such as an induction of vascular endothelial growth factor release, the activation of the coagulation system, or a modulating effect on the immune system.1–4, 38, 39
The current study results demonstrate an association between poor prognosis and high levels of IL-10, an antiinflammatory cytokine, as well as high values of IL-6, a proinflammatory marker. Therefore, these data appear to support the notion that despite the general activation of the immune system, host responses remain ineffective against the tumor and would indeed favor the progression of the disease.1, 2
IFN-γ represents an important marker of the “cell-oriented” immune response.1, 2 Experimental and preliminary clinical studies suggest that IFN-γ could reverse the defective immune response induced by other cytokines such as IL-10, and favor the development of an effective response against the tumor.24–26 To our knowledge, the current study is the first to evaluate the prognostic value for survival of circulating IFN-γ in patients with cancer. A nonstatistically significant prolongation in survival was found for patients with IFN-γ above the highest quartile. However, the relevance of this association needs to be evaluated further.
Several studies have suggested that platelets can promote tumor progression by regulating angiogenesis40 or favoring tumor metastasis.41 Platelet activation could lead to the release of P-selectin, which, in turn, may facilitate the attachment of tumor cells to the vascular wall.41 Indeed, plasma levels of P-selectin have been associated with survival and disease recurrences in patients with cancer.43 The predictive role of P-selectin is supported by our findings (Fig. 3).
In the current analysis, the influence of LMWH on circulating levels of IL-6, IL-10, IFN-γ, and P-selectin also was assessed. Data from a growing number of clinical trials suggest that LMWH does improve the prognosis and prolongs the survival of patients with cancer.27–32 Although the mechanisms behind the anticancer activity of LMWH remain poorly understood, these could involve an effect on the host immune response.33–38 However, this hypothesis was not confirmed by the results of the current study, in which none of the measured cytokines was found to correlate with the beneficial survival effects of nadroparin.28 Although a change in these cytokine concentrations would have given an indication of the general immune system activity against the tumor, the lack of such an effect cannot exclude a possible influence of LMWH on other immune markers or on immune pathways not reflected in circulating markers. In contrast to what was hypothesized initially, nadroparin treatment was associated with a modest increase in IL-6 levels, the clinical relevance of which to our knowledge remains unknown.
The survival prolongation noted with LMWH also could be explained by an effect on pathways such as angiogenesis or on P-selectin–mediated interactions between platelets and tumor cells,42 although the current data do not support this theory.
Given the relatively small sample size, the results of the current study have to be interpreted with caution and mainly considered as hypothesis generating. Because the investigated group included a broad range of malignancies, it was not possible to determine a hypothetical cancer type-specific effect of LMWH. However, it is reasonable that a potential LMWH anticancer activity directed against general mechanisms of cancer progression, such as the immune system or on platelets, would be less dependent on the type of cancer.
Circulating levels of IL-6, IL-10, and P-selectin appear to be predictive of an adverse prognosis in patients with advanced stage malignancy. Whether IL-10 circulating levels might help in guiding therapeutic decisions in patients with cancer remains to be evaluated. These markers were not sensitive to the LMWH administration, leaving the question of how LMWH positively impacts cancer progression unanswered.
The authors greatly appreciate the scientific input of Dr. C.T. Esmon from the Department of Pathology at the University of Oklahoma Health Sciences Center, Oklahoma City. They thank A. Groot from the Laboratory for Experimental Internal Medicine, Academic Medical Center, Amsterdam, for expert technical assistance.
- 25Clinical and biological effects of intraperitoneal injections of recombinant interferon-gamma and recombinant interleukin 2 with or without tumour-infiltrating lymphocytes in patients with ovarian or peritoneal carcinoma. Clin Cancer Res. 2000; 6: 2268–2278., , , et al.
- 31Low molecular weight heparin (LMWH) increases the efficacy of cisplatinum plus gemcitabine in advanced pancreatic cancer [abstract]. Proc Am Soc Clin Oncol. 2003; 22: 286., , , et al.
- 39Abnormalities in hemostasis in malignancy. In: ColmanRW, HirshJ, MarderVJ, ClowesAW, GeorgeJN, editors. Hemostasis and thrombosis. Philadelphia: Lippincott Williams & Wilkins, 2001: 1131–1152., , .