Abstract
Aim: To determine any correlation between inflammation parameters in blood glioma patients, with some of the established glioma biomarkers and to evaluate the possible prognostic impact of erythrocyte sedimentation rate (ESR) and C-reactive protein (CRP) for patient survival. Patients and Methods: This retrospective study evaluated ESR values in 94 patients and measured CRP values prior to the excision of primary glioma in 165 patients. Overall survival probabilities were determined separately for all patients with glioma in low-grade glioma (LGG), high-grade (HGG) and in glioblastoma multiforme (GBM) using the Kaplan-Meier log-rank test. The correlation between blood ESR and CRP values and between immunohistochemical (IHC) assessment of cluster of differentiation-68 (CD68), cathepsin B and nestin were evaluated. Results: An ESR above 15 mm/h was significant for poor survival prognosis for patients overall (p<0.001) and in the HGG (p<0.01) and GBM (p<0.04) subgroups. A serum CRP level above 5 mg/l was also identified as prognostic in patients overall (p<0.01), and in the HGG (p=0.02) and GBM (p=0.04) subgroups. Conclusion: Correlations of ESR and serum levels of CRP have been revealed with prognostic tissue biomarkers i.e. cathepsin B, nestin, CD68. Moreover, preoperative measurement of both parameters could be used for survival prognosis in patients with glioma.
- Cathepsin B
- C-reactive protein
- erythrocyte sedimentation rate
- glioblastoma
- nestin
- prognosis
- serum biomarkers
Although primary brain tumours are rare relative to other tumour types, in particular carcinomas, they are characterized by much higher mortality rates and increased patient disability, and by great resistance to conventionally-applied therapies (1). Malignant gliomas are among the most challenging of all types of cancers to treat successfully. Despite recent advances in neuroimaging, neurosurgical resection techniques and development of novel adjuvant therapies, glioblastoma multiforme (GBM), the most malignant of gliomas, still has a very poor prognosis, with a median survival of about 14 months after diagnosis, ranging from a few months to several years (2). There is thus a need for new, independent prognostic and predictive factors, determined by non-invasive methods in patient's serum that can enable individualized treatment modalities for patients with unfavourable prognoses (3).
Cancer has been regarded as a wound that does not heal, and inflammation has long been recognized as being important, playing a critical role in various processes related to cancer progression (4, 5). Most tumours are highly infiltrated by immune cells, including macrophages, neutrophils and lymphocytes (6). Inflammation has been reported to play a role in gliomagenesis, as well as in the invasiveness and proliferation of glioma (7-9). In a healthy central nervous system (CNS), the only cells with immune functions in the parenchyma are microglia and astrocytes. However, gliomas cause a breakdown of the blood brain barrier, so circulating immune cells, not normally found in the CNS, can gain access to tumour areas, and even the tumour cells themselves secrete a number of cytokines that modulate the activity of stromal cells (7). The glioma microenvironment, which is strongly immunosuppressive, is an indispensable participant in the neoplastic process (9). In addition, recent studies have demonstrated a primary role for cancer stem cells in tumour behaviour that not only includes control of tumour proliferation but can also initiate a microenvironment associated with immunosuppression (10).
Thus, inflammation and host immune response-related markers may be relevant as biological markers of glioma progression, although they have not been investigated to a great extent and not with respect to their relevance for the survival of patients with glioma.
There are two commonly used methods for detecting the acute phase response – the erythrocyte sedimentation rate (ESR), a marker of increased red blood cell aggregation, and the more specific measurement of C-reactive protein (CRP) concentration (11). The ESR has been in use since 1921, initially for monitoring clinical progress in patients with tuberculosis (12). It has also been shown to have prognostic value in the management of specific neoplastic disease (13). Increased ESR values have, for example, been found to correlate with overall poor prognosis in Hodgkin's disease, breast, gastric and colorectal, prostate and renal cell carcinoma, as well as in chronic lymphocytic leukaemia (14).
Another method for assessing the acute phase response is to measure concentration of CRP, first described in 1930 by Tillet and Francis (15). CRP belongs to the pentraxin family of calcium-dependent, ligand-binding plasma proteins. It is produced by hepatocytes in response to elevated cytokine levels following an inflammatory stimulus (16). De novo synthesis of CRP in liver starts very rapidly after a single stimulus, with serum concentrations subsequently rising above 5 mg/l in a few hours (17), and falling rapidly when the stimulus for increased production ceases (16). Two hypotheses for the association between circulating concentrations of CRP and cancer have been proposed in the review of Heikkilä et al. (18). The first hypothesis states that elevated CRP levels are a consequence of a pre-existing and malignant state, whereas the second proposes that chronic inflammation-related elevation of CRP could cause carcinogenesis (18, 19). The use of CRP as a tumour biomarker has gained interest and has been reported to have a prognostic impact in patients with various malignancies, including oesophageal cancer (20), lung cancer (21), breast cancer (22), renal cell carcinoma (23), hepatocellular carcinoma (24-26), endometrial cancer (27), melanoma (28), diffuse large B-cell lymphoma (29), Hodgkin's lymphoma (30), non-Hodgkin's lymphoma (31), extranodal natural killer/T-cell lymphoma (32), urothelial carcinoma (33), penile cancer (34), and castration-resistant prostate cancer (35).
We hypothesize that the prognosis of glioma can be improved by considering additional independent biological prognostic markers. We first aimed to evaluate the impact of the pre-treatment ESR and CRP levels on the prognosis of patients with glioma. Secondly, in an attempt to obtain insight into their possible role in various carcinogenesis-related processes, we also sought correlations of both inflammation parameters with some of the glioma biomarkers determined in the same patient population: invasion and angiogenesis-related biomarker cathepsin B, macrophage and glial migration marker cluster of differentiation (CD) 68, and the neural and glioma stem cell marker nestin. These were the subjects of our previous investigations and were all found to be highly prognostic of survival of patients with glioma.
Patients and Methods
Patients and tumor characteristics. This retrospective study included patients with primary brain tumours who underwent surgery at the Department of Neurosurgery, Clinical Centre Maribor, between 1986 and 1999 for the ESR group (94 patients), and between 1994 and 2009 for the CRP group (165 patients). All patients underwent complete tumor resection. The tumor samples were obtained from different locations in the tumour. The histological slides were analyzed and classified according to the WHO classification of brain tumours (36). Our institutional databases were used to obtain patient and tumour characteristics, such as age, sex, date and type of initial operation, and histological type. Pathologies included 20 low-grade glioma (LGG; WHO grade I, II) and 74 high-grade glioma (HGG; WHO grade III, IV) in the ESR group, and 23 LGG and 142 HGG tumours in the CRP group (Table I). Pre-operatively, administration of dexamethasone was started with a daily dose of 4×4 mg or 3×8 mg. Following surgery, patients with Karnofsky performance score above 50 were seen by an oncologist to decide on adjuvant (radiotherapy, temozolomide) therapy (Table II). The serum marker data included values of ESR and CRP, while protein biomarkers cathepsin B, CD68 and nestin were determined by immunohistochemistry (IHC) as described below, together with details of the follow-up data.
Markers of inflammation. A retrospective study of 94 patients was performed, including information about their ESR values, and of the further 165 patients with their CRP values, measured prior to neurosurgical intervention.
Pre-treatment ESR and CRP values were measured in peripheral venous blood samples before surgery. The blood was collected into K2 EDTA. For ESR, a sample of each specimen was diluted in 1/4 volume of 109 mmol/l trisodium citrate and placed in a 200 mm tube in a vertical position. At the end of the first hour, the distance from the meniscus to the top of the column of the erythrocytes was recorded as the ESR (in mm/h). The preoperative ESR value was categorized as low (0-15 mm/h) or high (>15 mm/h).
CRP was measured using turbidimetry on a Hitachi 912 analyzer (Boehringer Mannheim, Mannheim, Germany) according to the manufacturer's instructions. The preoperative CRP value was categorized as low (0-5 mg/l) or high (>5 mg/l).
IHC analysis. IHC for cathepsin B was performed using the avidin-biotin-peroxidase complex (ABC) technique (37, 38). Immunostaining was performed with an automatic immunostainer TM 500 (DAKO-Bio Tek Solutions, Santa Barbara, CA, USA). Slides were incubated with monoclonal antibodies against cathepsin B (clone 3E1, KRKA, d.d. Novo mesto, Slovenia) for 25 min at room temperature. They were then washed three times for 5 min in diaminobenzidine solution at room temperature. Tissue sections were counterstained with Mayer's haematoxylin. Immunostaining for cathepsin B was scored separately for the tumour cells and the endothelial cells. Twenty representative fields were counted for each slide. The frequency of immunoreactivity for cathepsin B in tissue sections was evaluated as negative when no positively stained cells were observed within the tumour, weak (1+) when fewer than 30% of the tumour cells were positively stained, moderate (2+) when 30-60% of the tumour cells were positively stained, and strong (3+) when more than 60% of tumour cells stained positively. The intensity of staining was evaluated as 0, 1+, 2+ and 3+ for no staining, weak, medium and strong staining, respectively. The IHC score was recorded as the sum of the frequency and intensity score for tumor cells, endothelial cells, and for both types of cells together. The results of cathepsin B staining in tumour cells and in endothelial cells were subdivided into two groups, one with weaker staining, scored 0 to 4+, and the other with more intense staining, scored 5+ to 6+. The results of staining in both types of cells together, were subdivided into groups with scores of 0 to 8+ and 9+ to 12+ as weak and strong positive staining, respectively.
Histological characteristics of patients by glioma subtype.
IHC staining for nestin and for CD68 was performed using the standard technique according to the protocol of the Department of Pathology at the Maribor Clinical Centre (39-41). The slides were incubated overnight at 4°C with primary antibodies raised against nestin (Ab #4350; Karolinska Institute, Stockholm, Sweden) or monoclonal antibodies against CD68, a macrophage and microglia marker (MO876; DAKO, Glostrup, Denmark). After washing in Tris-buffered saline (TBS; Sigma, Munich, Germany) immunoperoxidase staining was performed by an EnVisionAb (complex) method using the Envision kit (DAKO) (39). After rinsing in tap water, the sections were counterstained with Mayer's haematoxylin and mounted. Results were scored as described above for cathepsin B. The results of nestin staining in tumour cells and in endothelial cells were sub-divided into two groups, with weaker staining scoring 0 to 3 and more intense staining scoring 4 to 6. The results of staining in both cell types together were sub-divided into groups with scores of 0 to 6+ and 7+ to 12+ as weak and strong positive staining. The results of CD68 staining in tumour cells were sub-divided into two groups, with weaker staining scored 0 to 3+, and with more intense staining scored 4+ to 6+ (38, 40, 41).
Statistical analysis. Variables used in the analysis included ESR, CRP, cathepsin B, nestin and CD68 IHC scores (in tumour cells and in endothelial cells, as well as the total IHC scores). Descriptive statistical methods (e.g. median value, frequency tables), independent Student t-test, and simple linear correlation were used.
Overall survival probabilities were calculated by the method of Kaplan and Meier (42). The subgroups were compared using the log-rank test. The survival time was determined as the interval between initial operation and the patient's death or July 2012 for those surviving to this date. Statistical analysis was performed using the program Statistica for Windows 6 (StatSoft, Inc., Tulsa, OK, USA). A two-tailed p-value less than 0.05 was considered significant in all tests.
Results
ESR in patients with glioma. All patients' characteristics are given in Table I, together with the glioma subtype. The group of patients with evaluated ESR (55 men and 39 women; median age=51 years, range 12 to 75 years) consisted of 74 (78.4%) diagnosed with HGG, including 61 with GBM. Karnofsky indices above 50 were found in 63 patients with HGG, and in 51 with GBM. Follow-up ranged from 0 to 294 months, with a mean and median of 28 and 7 months. By the time of final data acquisition, 88 (93.6%) patients had died.
Patients with adjuvant therapy as prescribed by oncologist.
Survival data of patients in subgroups from the erythrocyte sedimentation rate (ESR) group*.
The mean and median preoperative ESR values in patients overall were 14.7 and 10 mm/h. A mean ESR level of 15 mm was selected as the cut-off point. Higher levels were measured in 29 (30.9%) patients, whereas no elevation was noted in 65 (69.1%) patients. A significant correlation between ESR and survival time was found in patients overall (p=0.02), in the HGG group (p=0.03) and in the GBM group (p=0.045).
Survival in the ESR group was analyzed separately for patients overall, for HGG and for GBM (Table III). In the HGG group, significantly (p=0.03) shorter survival was observed in 27 patients with a high ESR (median=3 months) than in other 47 patients with a low ESR (median=7 months). ESR levels above 15 mm/h were significantly related to poor prognosis in patients overall, in both the HGG and the GBM groups (log-rank p=0.001, p<0.01, and p=0.04, respectively) (Table III and Figure 1).
Correlations between erythrocyte sedimentation rate (ESR) and biomarkers in malignant glioma.
Survival data of patients in subgroups of patients from the C-reactive protein (CRP) group*.
Correlation between C-reactive protein (CRP) and glioma markers in malignant glioma.
Pre-treatment ESR correlated with previously established prognostic biomarkers in the same group of patients (38, 40). In the GBM group, cathepsin B immunostaining in endothelial cells, nestin staining of tumour cells, and total nestin score correlated with ESR levels (p=0.04, p<0.05, and p=0.04, respectively) (Table IV).
CRP. The group of patients with evaluated CRP included 165 patients who underwent surgery for HGG (n=142; 86.1%) or LGG (n=23; 13.9%) (Table I). The median age was 56 (range 5-83) years and the majority of patients were male (n=105; 63.6%). Karnofsky performance scores above 50 were found in 111 patients from the HGG subgroup and 85 patients from the GBM subgroup. The follow-up ranged from 0-154 months with a mean and median duration of 21.7 and 10 months and, by the time of final data acquisition, 144 (87.3%) patients had died.
The preoperative CRP concentration was within the normal range (0-5 mg/l) in 132 patients and elevated in 33 patients (20.0%). High CRP levels were observed in 30 patients out of 142 from the HGG group, with mean and median values of 18.1 and 11 mg/l (range=6-54 mg/l). In the subgroup of 112 patients with HGG and low CRP the mean and median values of serum CRP was 2.2/2 mg/l (range=0-5 mg/l). Serum CRP levels correlated significantly with survival time in patients overall (p=0.02) and in the HGG group (p=0.03). The univariate survival analysis in the CRP group, including data of HGG and GBM subgroups, is presented in Table V. Patients with HGG in the low CRP group (112 patients) survived significantly (p=0.03) longer (median=9 months) than those in the high CRP group (30 patients) (median=5 months). A serum CRP level above 5 mg/l before treatment was identified as a predictor of poor survival in patients overall (p<0.01), in the HGG group (p=0.02) and in the GBM group (p=0.04) (Figure 2).
The correlation of pre-treatment CRP with biomarkers is presented in Table VI. In the HGG subgroup, a significant correlation was found between the CRP and IHC score for cathepsin B in endothelial cells (p=0.01) and total cathepsin B score (p=0.03). Both correlations were confirmed in the GBM subgroup (p=0.01 and 0.02, respectively). Additionally, a significant correlation between CRP level and CD68 IHC score was observed in patients with GBM (p=0.02).
Discussion
The clinical usefulness of ESR and serum CRP levels for predicting survival of patients with glioma has been established for the first time, to the best of our knowledge.
ESR. Survival analysis in our cohort of patients revealed that pre-treatment ESR has a strong prognostic impact on survival of subgroups of patients with glioma. In both the HGG and GBM groups, a high ESR value was a significant prognosticator of shorter survival. Mirzayan et al. stated that there is very little information to be gained from the variation of ESR in patients with brain tumours (13), although they only followed levels after intracranial surgery and did not relate those to prognostic relevance. The mechanism underlying an-elevated ESR level is not clear, but it is reasonable to speculate that it is related to the systemic response to the malignant tumour burden, as discussed below for CRP.
CRP. In our study, elevated serum CRP levels were associated with reduced survival overall in patients with glioma. A high pre-surgical CRP level was associated with poor survival in the group of patients with more malignant glioma – the HGG and GBM subgroups. Recently, Reynés et al. reported that inflammation markers, including CRP, were significantly elevated in patients with GBM (43). Forty patients and 60 healthy subjects were included in their study. In contrast to our study, they did not find an association between CRP levels and survival or progression-free survival, due possibly to the small population of patients with malignant glioma included in their study. Although the mechanisms underlying the relationship between high pre-treatment CRP values and shorter survival are not clear, we propose that the pre-operative serum CRP level can be used to predict survival, at least in patients with more malignant disease, who require aggressive management.The origin of plasma CRP in patients with tumour is not clear, since it is reported to be produced only by hepatocytes (16). Heikkilä et al. reviewed the literature for evidence of the role of circulating CRP in the diagnosis and aetiology of cancer, and provided no strong evidence for a causal role of CRP in malignancy (18). Allin et al. reported that polymorphisms in the CRP gene associated with increased circulating levels of CRP are not associated with increased risk of cancer (44). In their prospective study, however, which included approximately 10,000 apparently healthy individuals from the general population observed for up to 16 years, increasing levels of CRP were found to be associated with increasing risk of incidence of lung cancer and colorectal cancer, but not of breast or prostate cancer (19, 45). Epidemiological studies thus suggest that for several types of solid cancer, elevated circulating levels of CRP are associated with poor prognosis. Based on these data, it can be speculated that CRP reflects the systemic response to inflammation, as a response to cancer progression. Tumour-associated inflammation is characterized by infiltration of innate immune cells, such as microglia/macrophages, and the production of cytokines and chemokines (46, 47). Several studies have suggested a strong relation between chronic inflammation and consequent development of cancer (5, 18, 48).
A: Prognostic significance of erythrocyte sedimentation rate (ESR) in patients with primary brain tumours. Patients were classified into two categories of ESR, low (0-15 mm/h) and high (>15 mm/h), according to their pre-treatment value. Survival interval was determined as the interval between the date of the initial operation and date of patient's death or determined endpoint (July 2012) for those surviving to this date. Prognostic significance of pre-treatment ESR for the survival of patients with high-grade glioma (B) and for patients with glioblastoma multiforme (C).
A: Prognostic significance of C-reactive protein (CRP) in patients with primary brain tumours. Patients were classified into two categories of low (0-5 mg/l) and high (>5 mg/l) CRP level, according to their pre-treatment value. Survival interval was determined as the interval between the date of the initial operation and date of patient's death or determined endpoint (for those alive July 2012). Statistical analysis was performed according to Kaplan and Meier (42). Prognostic significance of pre-treatment ESR for the survival of patients with high-grade glioma (B) and for the patients with glioblastoma multiforme (C).
Inflammation can, on the other hand, by itself promote tumorigenesis via genetic damage caused by increased levels of reactive oxygen and nitrogen species or activation-induced cytidine deaminase, a mutagenic enzyme. High levels of CRP and ESR may, in such cases relate to an immunogenic aetiology of carcinogenesis. The cellular origin of increasingly secreted plasma CRP is currently not known and to shed light on this issue, it could be advantageous to measure CRP level in glioma tissues.
Correlations. Clues to the origin of CRP elevation and cause of ESR elevation in the blood of malignant glioma patients could come from correlations with other prognostic markers, as measured by IHC in the same patients' tumour tissues (38). Various proteolytic enzymes have been shown to be good prognostic factors in many types of tumors, due to their specific roles in individual steps of tumor progression (49). In gliomas, the cysteine proteases cathepsins B, L and S, the urokinase-plasminogen serine protease system and several metalloproteinases have been reported to be up-regulated, and some of these have been shown to have a prognostic impact on patient survival (50, 51). These may be related to the acquisition by the tumour of invasive properties and angiogenesis, both highly relevant for glioma progression. Our previous studies demonstrated that protein levels of cathepsin B in tumour and endothelial cells are significantly higher in GBM than in LGG, and has a high prognostic impact. Here, we have shown in the GBM group of patients that IHC scores for cathepsin B in endothelial cells correlate significantly with ESR. In the highly malignant HGG and GBM subgroups, cathepsin B expression in endothelial cells and total cathepsin B expression correlated significantly with CRP.
CD86 is a known macrophage marker, presumably also marking malignant invasive GBM cells. As reported previously (41), it correlates highly with prognosis in all subgroups of glioma, indicating tumour aggressiveness, presumably assisted by tumour-associated macrophages. Here, we found a significant correlation between CRP level and CD68 IHC score in patients with GBM. This leads to the conclusion that CRP is associated with tumour invasion, reflected in the high levels of biomarker cathepsin B and indicating more angiogenic tumours, possibly assisted by infiltrating CD68-expressing macrophages with their immunosupportive activity.
In our previous study we compared markers cathepsins B, L, musashi and nestin. We found the highest prognostic relevance of the GBM stem cell marker nestin, as expressed in tumor cells and in endothelial cells of primary gliomas (40), which may be used to predict for risk of death in patients with malignant primary tumours of the CNS. Here, we found that nestin staining of tumour cells, as well as total nestin score, correlates significantly with ESR levels. It is not possible at present to speculate on the biological relationship between stem cell markers and the ESR.
In conclusion, our study strongly suggests that pre-treatment ESR values and serum CRP levels are both significant and independent prognostic parameters in glioma patients. Both biomarkers might complement the prognostic value of cathepsin B, CD68 and nestin and possibly some of other established glioma markers, not investigated in this report.
Acknowledgements
This work was supported by the GLIOMA- Interreg (Slovenian - Italian Cooperation 2007-2013) project. We thank Dr. Irena Strojnik for technical assistance in preparing the manuscript and Mrs. Eva Fliser, medical biochemistry specialist, for providing us with the technical details of CRP analyses. We also thank Roger Pain for reviewing the manuscript.
- Received November 3, 2013.
- Revision received November 17, 2013.
- Accepted November 19, 2013.
- Copyright© 2014 International Institute of Anticancer Research (Dr. John G. Delinassios), All rights reserved
