Fax: (011) 45 89492530
Elevated serum level of YKL-40 is an independent prognostic factor for poor survival in patients with metastatic melanoma†
Article first published online: 2 FEB 2006
Copyright © 2006 American Cancer Society
Volume 106, Issue 5, pages 1130–1139, 1 March 2006
How to Cite
Schmidt, H., Johansen, J. S., Gehl, J., Geertsen, P. F., Fode, K. and von der Maase, H. (2006), Elevated serum level of YKL-40 is an independent prognostic factor for poor survival in patients with metastatic melanoma. Cancer, 106: 1130–1139. doi: 10.1002/cncr.21678
Quidel (San Diego, CA) provided the YKL-40 ELISA kits.
- Issue published online: 17 FEB 2006
- Article first published online: 2 FEB 2006
- Manuscript Accepted: 29 SEP 2005
- Manuscript Revised: 8 SEP 2005
- Manuscript Received: 27 APR 2005
- Danish Cancer Society, “Direktor Jens Aage Sorensen og Hustru Edith Ingeborg Sorensens Mindefond,” and “Radiumstationens forskningsfond.”
- metastatic melanoma;
YKL-40 is a growth factor for connective tissue cells and stimulates migration of endothelial cells. Cancer cells, macrophages, and neutrophils secrete YKL-40. Its function in cancer is unknown. High serum YKL-40 levels have been associated with a poor prognosis in patients with several solid tumors. The prognostic impact of serum YKL-40 in metastatic melanoma was evaluated.
YKL-40 was measured in serial serum samples from 110 patients with metastatic melanoma obtained immediately before and during treatment and from 245 healthy subjects.
Patients had higher serum YKL-40 values than healthy subjects (P < 0.001). Pretreatment serum YKL-40 was elevated in 45% of the patients and correlated to site of metastases (P = 0.03) and poor performance status (P = 0.002). Multivariate Cox analysis showed that serum YKL-40 (hazard ratio [HR] = 1.9; 95% confidence interval [CI], 1.2–2.8; P = 0.004) and serum lactate dehydrogenase (LDH) (HR = 1.9; 95% CI, 1.2–2.9; P = 0.004) were independent prognostic factors for survival. A combination variable of elevated serum YKL-40 and LDH quadrupled the risk of early death (HR = 4.4; 95% CI, 2.5–7.7; P < 0.001) compared with patients with normal levels. The combination of YKL-40 and LDH had a stronger prognostic impact than the American Joint Committee on Cancer (AJCC) Stage IV classification. Furthermore, serum samples were available from 12 patients during followup. In 9 of 11 patients a significant increase in serum YKL-40 was observed together with disease progression. In one patient with a lasting complete response, serum YKL-40 remained normal.
An elevated serum YKL-40 was an independent prognostic factor for poor survival in patients with metastatic melanoma. When combining serum YKL-40 and LDH, patients could be separated into three prognostic groups based on the number of elevated biomarkers. The findings should be validated in an independent study. Cancer 2006. © 2006 American Cancer Society.
Metastatic melanoma remains a disease with a poor prognosis, although a small number of patients obtain long-term survival with interleukin-2 (IL-2) and interferon-α (IFN) immunotherapy.1–3 An elevated serum level of lactate dehydrogenase (LDH) is an independent prognostic factor for poor survival in metastatic melanoma4 and currently the only biological marker (biomarker) used in the American Joint Committee on Cancer (AJCC) staging for metastatic melanoma.5 New prognostic biomarkers may add to our knowledge and may improve the selection of patients for different kinds of therapies.
YKL-40 is a heparin- and chitin-binding lectin6–8 secreted by activated neutrophils,9 macrophages during late state of differentiation,10, 11 arthritic chondrocytes,6 differentiated vascular smooth muscle cells,8, 12, 13 and fibroblast-like synovial cells.14, 15 The term YKL-40 originates from the one-letter code for its three N-terminal amino acids and the molecular weight (40 kDa).16 YKL-40 is a growth factor for connective tissue cells and acts synergistically with insulin-like growth factor 1 (IGF-1) in stimulating the growth of fibroblasts.17 Mitogen-activated protein kinases (MAP kinase) and PI-3K signaling cascades in fibroblasts is initiated by YKL-40, leading to the phosphorylation of both the extracellular signal-regulated kinase (ERK)-1/2 MAP kinase and protein kinase B (AKT)-mediated signaling cascades, which are associated with the control of mitogenesis.17 This suggests a role of YKL-40 in tissue remodeling and as an antiapoptotic protein. YKL-40 also acts as a chemoattractant for endothelial cells. It stimulates their migration at a level comparable to basic fibroblast growth factor (bFGF),12 and promotes migration and adhesion of vascular smooth muscle cells.13 These in vitro observations indicate a function of YKL-40 in angiogenesis, although in vivo proof of this is yet to be obtained.
Elevated serum levels of YKL-40 are seen in a number of nonmalignant diseases characterized by inflammation and remodeling of the extracellular matrix, such as active rheumatoid arthritis,18–20 severe bacterial infections,21 active inflammatory bowel disease,22 and liver fibrosis.23 The gene for YKL-40 has been sequenced.10 It is expressed by several types of cancer such as colon, breast, ovarian, uterine, prostate, kidney, lung, oligodendroglioma, glioblastoma, and germ cell tumors. An elevated serum level of YKL-40 defined as higher than the 95th percentile of serum YKL-40 in healthy age-adjusted subjects24 was shown as an independent prognostic factor for poor survival in patients with cancer of the breast, colon, ovary, kidney, and lung.25–34
At present, only one study has evaluated the importance of serum YKL-40 for treatment monitoring of cancer patients.27 Radically operated colorectal carcinoma patients with elevated serum YKL-40 6 months after operation had a shorter recurrence-free interval and overall survival than patients with normal serum YKL-40 6 months postoperatively.27
As serum YKL-40 has not previously been evaluated in melanoma, our primary goal was to describe the biomarker in 110 patients with metastatic melanoma and to evaluate the prognostic impact on overall survival from samples collected before treatment. Second, we wanted to examine whether changes in serum levels could be used to predict objective response during treatment and/or progression in the followup period.
MATERIALS AND METHODS
The patients included in the current biomarker study were treated in three Phase II protocols in the period from February 1997 to December 2003. The main inclusion criteria were identical in the three protocols and consisted of biopsy-verified Stage IV disease, a WHO performance status of 2 or better, normal kidney, heart, and liver functions, and the absence of brain metastases. Thirty-three patients received IL-2, IFN, and cisplatin.35 Forty-nine patients received IL-2, IFN ± histamine.36 Twenty-eight patients received electrochemotherapy with intratumoral injections of bleomycin and IL-2.37 The local ethics committees accepted the present project and written informed consent was obtained from each patient. Clinical response in all three protocols was evaluated according to WHO criteria.
The normal range of serum YKL-40 was determined in 245 normal subjects (134 females and 111 males; median age, 49; range, 18–79 yrs). These subjects were all healthy, were not on medication, and had no signs or clinical symptoms of cancer, joint, liver, metabolic or endocrine disease.24
Processing of Blood Samples and YKL-40 Analysis
Blood samples were collected before treatment, after 3 weeks, and after the first treatment course in all three trials. Furthermore, followup samples were obtained from 12 of 17 responding patients. Serum samples were stored at −80°C until analysis for YKL-40, and all samples were analyzed in a blinded fashion. Serum concentrations of YKL-40 were determined by a commercial two-site, sandwich-type enzyme-linked immunosorbent assay (ELISA; Quidel, San Diego, CA)19 using streptavidin-coated microplate wells, a biotinylated-Fab monoclonal capture antibody, and an alkaline phosphatase-labeled polyclonal detection antibody. The sensitivity of the ELISA was 10 μg/L. The intra- and interassay coefficients of variation were < 3.6% and < 3.7%, respectively. A normal reference region was calculated as described by Royston38 on the log-transformed (loge) serum YKL-40 values of the healthy controls adjusting for age and the 95% percentile was chosen as the cut-point.
Calculations were performed using SPSS statistical software (v. 11.5, Chicago, IL). Significant differences between groups with a nonparametric data distribution were analyzed with the Mann-Whitney U-test for two independent groups or Kruskal–Wallis test when three or more independent groups of sampled data were compared. For the comparison of paired nonparametric data we used the Wilcoxon test and the Spearman correlation test was used to test for correlations. Response analysis was performed using Fisher exact or Pearson χ2 test. Clinical response was dichotomized as response (complete [CR] and partial response [PR]) versus no response (no change [NC] or progressive disease [PD]) and correlated with pretreatment serum YKL-40, dichotomized as normal versus elevated.
Univariate and multivariate Cox analyses were performed to investigate the prognostic impact of pretreatment factors in relation to survival. Factors included were WHO performance status (0–1 vs. 2), number of metastatic sites (1–2 vs. ≥ 3), location of metastases (skin and lymph nodes vs. lungs vs. other visceral sites), serum LDH (normal vs. elevated), and serum YKL-40 (normal vs. elevated).
Survival was calculated from the day of treatment start to the endpoint (death or censoring). The median survival time was analyzed by the Kaplan–Meier method. The simultaneous relation of multiple prognostic factors for survival was assessed with Cox proportional hazards model using the enter method. Factors with a P-value less than 0.10 in the univariate analyses were included in the multivariate analysis to identify factors of independent significance. Hazard ratios (HRs) were calculated to estimate the magnitude and the direction of the effect. The assumption of proportional hazards was verified graphically. All data on duration of survival were updated on August 1, 2005.
Patient characteristics are given in Table 1. The median serum level of YKL-40 in pretreatment samples from 110 patients with metastatic melanoma was 95 μg/L (range, 16–1262), which was significantly higher (P < 0.001) compared with the median level of 43 μg/L (range, 20–184) in 245 healthy controls. Individual serum YKL-40 levels of the 245 healthy controls and the 110 metastatic melanoma patients are illustrated in Figure 1. The nonage-corrected upper reference level of serum YKL-40 in the present control material was 124 μg/L and rises slightly with age. Patients with visceral metastases other than lung metastases had significantly higher serum YKL-40 levels compared with patients with lung metastases or skin and lymph node metastases (P = 0.005).
|No. of patients||Percentage (%)|
|Total no. of patients||110||100|
|Median (range)||53 (29–74)|
|Site of metastases|
|WHO performance status|
|No. of metastatic sites|
Forty-five percent of the patients with metastatic melanoma had an elevated serum YKL-40 level above the age-corrected upper 95th percentile of the level in healthy subjects, and 37% had a serum LDH above the upper 95th percentile of the standard for the institution. In patients with a normal LDH level, 27 of 69 patients (39%) had an elevated YKL-40 level, whereas 22 of 41 patients (54%) with an elevated LDH level also had an elevated YKL-40 level. An elevated level of YKL-40 was correlated to visceral metastases other than lung metastases (Spearman rank correlation coefficient r = 0.21; P = 0.03) and to poor performance status (r = 0.29; P = 0.002).
In order to correlate changes in serum YKL-40 during treatment with clinical response, we examined the variation of serum YKL-40. Samples were drawn at fixed times for 3 days before treatment in 11 patients (Fig. 2). A total of 95 samples (four samples missed) were analyzed. The median coefficient of variation was 12% (range, 7–28%). A change of more than 24% (i.e., two times the observed coefficient of variation) was thereafter regarded as a significant change when comparing consecutive samples.
The median survival was 8.5 months (range, 1–102 mos) and 7 patients (6%) were alive (range, 18–102 mos) at the time of analyses. Complete response was observed in 6 patients (5%), partial response in 11 patients (10%), no change in 14 patients (13%), and progressive disease in 76 patients (69%). Three patients (3%) were not assessable for response.
Baseline Serum YKL-40 and Clinical Outcome
The patients were treated with three different IL-2 treatment regimes, which, however, did not result in statistically significant survival differences (Table 2). Thus, pretreatment serum YKL-40 levels were evaluated together in all patients for the survival analyses. Normal versus elevated serum levels of YKL-40 and LDH were analyzed together with number of metastatic sites, location of metastases, and performance status. In the univariate analyses, serum YKL-40, serum LDH, performance status, number of metastatic sites, and location of metastases were all significant prognostic factors for survival (Table 2). Figure 3 shows the survival plot for patients with normal versus elevated pretreatment serum YKL-40.
|Variables||Categories compared||No. of patients||Univariate analyses||Multivariate analysisa|
|HR||95% CI||P||HR||95% CI||P|
|Location of metastases||Skin/lymph nodes vs. lung vs. visceral sites||36 vs. 26 vs. 48||0.002b||0.27b|
|Skin/lymph nodes vs. lung||36 vs. 26||1.6||0.9–2.8||0.09||1.5||0.8–2.6||0.21|
|Skin/lymph nodes vs. visceral sites||36 vs. 48||2.3||1.5–3.7||< 0.001||1.6||0.9–2.7||0.12|
|Treatment||1 vs. 2 vs. 3||33 vs. 49 vs. 28||0.71b|
|1 vs. 2||33 vs. 49||1.2||0.8–1.9||0.41|
|1 vs. 3||33 vs. 28||1.1||0.6–1.9||0.71|
|No. of metastatic sites||1–2 vs. ≥ 3||70 vs. 40||2.0||1.4–3.1||0.001||1.2||0.7–2.0||0.45|
|Performance status||0–1 vs. 2||90 vs. 20||2.5||1.5–4.2||0.001||1.5||0.9–2.6||0.15|
|LDH||Normal vs. elevated||69 vs. 41||2.4||1.6–3.5||< 0.001||1.9||1.2–2.9||0.004|
|YKL-40 (age adjusted)||Normal vs. elevated||61 vs. 49||2.0||1.3–2.9||0.001||1.9||1.2–2.8||0.004|
In the multivariate analysis (Table 2), both an elevated serum YKL-40 (hazards ratio [HR] = 1.9; 95% confidence interval [CI], 1.2–2.8; P = 0.004) and an elevated serum LDH (HR = 1.9; 95% CI, 1.2–2.9 P = 0.004) proved to be independent prognostic factors for poor survival. We also analyzed whether serum YKL-40 added prognostic information to the AJCC Stage IV classification in a Cox regression analysis including these two covariates. Serum YKL-40 was an independent prognostic factor after correction for the Stage IV classification (HR = 1.9; 95% CI, 1.3–2.8; P = 0.002). Thus, serum YKL-40 significantly separated the survival curves for the three Stage IV subcategories. The AJCC staging classification for metastatic melanoma uses a combination of metastatic sites and serum LDH. Stage M1a is patients with skin or lymph node metastases and a normal LDH; Stage M1b is patients with lung metastases as the only visceral site and a normal LDH. Finally, Stage M1c is patients with visceral metastases other than the lung metastases and/or elevated LDH (Fig. 4).
With regard to clinical response, a normal pretreatment serum YKL-40 was seen in 13 of 17 patients (76%), who subsequently responded to treatment, whereas a normal pretreatment level was seen in 47 of 90 patients (52%) not responding to treatment. This difference was not statistically significant (Fisher exact test, P = 0.11).
Baseline Serum YKL-40 in Combination with LDH as a Prognostic Index
Based on the identified independent prognostic factors, we constructed a combination variable. Figure 5A illustrates the survival plot for patients grouped into three prognostic groups with both YKL-40 and LDH normal, one parameter elevated, or both parameters elevated, respectively. The median survival for these three groups was 11.9 months, 7.6 months, and 3.4 months, respectively (P < 0.001). Patients with one elevated biomarker had a 2-fold increase in risk of early death (HR = 2.0; 95% CI, 1.2–3.1; P = 0.003), and patients with both markers elevated had a 4.4-fold risk of early death (HR = 4.4; 95% CI, 2.5–7.6; P < 0.001) compared to patients with no elevated biomarkers in the univariate analysis (Table 3).
|Covariates||No. of patients||Univariate analyses||Multivariate analysisa|
|HR||95% CI||P||HR||95% CI||P|
|AJCC criteria Stage IV|
|M1a vs. M1b vs. M1c||30 vs. 15 vs. 65||0.001b||0.15b|
|M1a vs. M1b||30 vs. 15||1.4||0.7–2.7||0.30||1.5||0.8–2.9||0.21|
|M1a vs. M1c||30 vs. 65||2.4||1.5–3.9||< 0.001||1.7||1.0–2.8||0.06|
|Both normal vs. one elevated vs. both elevated||42 vs. 46 vs. 22||< 0.001b||0.001b|
|Both normal vs. one elevated||42 vs. 46||2.0||1.2–3.1||0.003||1.8||1.1–2.9||0.02|
|Both normal vs. both elevated||42 vs. 22||4.4||2.5–7.6||< 0.001||3.4||1.8–6.6||< 0.001|
In the univariate analyses, there was a significant difference in survival between the M1a and M1c stage (P < 0.001), but not between M1a and M1b (P = 0.30) (Table 3; Fig. 5B). The combination variable of YKL-40 and LDH maintained its independent prognostic impact after correcting for the AJCC Stage IV classification (Table 3).
Serum YKL-40 during Treatment and Clinical Response
During treatment, serum samples were obtained after 3 weeks of treatment and after the first treatment cycle and compared with the pretreatment sample. A significant increase in serum levels of YKL-40 was observed in 60% of the patients after 3 weeks of treatment, and a significant decrease in 12% of the patients. However, no significant correlation was observed between changes in serum YKL-40 and objective response during treatment (Pearson χ2 test, P = 0.40) or after treatment (Pearson χ2 test, P = 0.96).
Serum YKL-40 during Followup and Progressive Disease
Followup samples were obtained from 12 of the 17 patients responding to treatment and compared with pretreatment samples. In 9 of 11 patients, who progressed during followup, a significant increase from pretreatment levels accompanied clinical progression (Fig. 6). In contrast, all nine followup samples from a patient with a lasting complete response were normal.
This is to our knowledge the first report on the novel serum biomarker YKL-40 in patients with metastatic melanoma. Forty-five percent of the patients had elevated serum levels of YKL-40, which were significantly higher than those in healthy persons. As shown in the correlation analyses, patients with an elevated serum YKL-40 had more frequent visceral involvement and poor performance status. Interestingly, an elevated serum YKL-40 was an independent prognostic factor for poor survival with a prognostic impact comparable to serum LDH.
A combination variable of serum YKL-40 and LDH was able to separate three prognostic groups of patients. Patients with initially normal serum levels of YKL-40 and LDH had a median survival of 11.9 months compared with only 3.4 months in patients with both biomarkers elevated, whereas the median survival was 7.6 months in patients with one elevated biomarker. None of the patients with both elevated serum YKL-40 and LDH survived beyond 2 years. Apparently, there was no interaction between serum YKL-40 and LDH, as indicated by the hazard ratio for the combination variable, being similar to the sum of the hazard ratios for the two individual biomarkers.
Applying our data to the currently accepted AJCC Stage IV classification, we observed a significant difference in survival between patients with skin/lymph node metastases (M1a) and patients with visceral metastases other than lung metastases and/or elevated LDH (M1c), but not between patients with M1a and patients with lung metastases (M1b). These results correspond well with the data reported by Balch et al.,39 where a significant difference between M1a and M1b was seen at the 1-year survival but not beyond that timeframe. In our study, serum YKL-40 was an independent prognostic factor after correction for the AJCC Stage IV classification.
We included patients from three Phase II studies, which may be a point of criticism, as the patients had different kinds of treatment. Optimally, samples should be obtained from patients receiving one type of treatment. However, inclusion criteria were similar in the three studies, all patients received IL-2 based immunotherapy, and there were no statistically significant survival differences between the three trials.
The biological function of YKL-40 in cancer diseases is unknown. It has been hypothesized that YKL-40 is a growth factor of cancer cells or protects them from undergoing apoptosis. Cancer progression depends on the interplay between the cancer cells and their microenvironment, particularly cells in the surrounding extracellular matrix. The balance between synthesis and degradation of extracellular matrix components is a key modulator of cancer growth and metastasis.40 The stroma around the periphery of solid cancers have several similarities with granulation tissue such as that found in wound-healing or inflammation.41–43 Recent studies have shown that tumor-associated macrophages and leukocytes play important roles in tumor growth and metastasis, because these cells produce growth and angiogenic factors, chemokines, metalloproteinases, and other extracellular matrix-degrading enzymes.44–46 YKL-40 stimulates migration of endothelial cells at a level comparable to that achieved by bFGF12 and modulates vascular endothelial cell morphology by promoting the formation of branching tubules. YKL-40 may therefore be a positive regulator of angiogenesis surrounding the tumor and could play a role in the growth of primary and metastatic tumors. These hypotheses are supported by the correlations shown between elevated serum YKL40 and poor survival in cancer of the breast, colon, ovary, kidney, and lung.25–34 We found the same association for metastatic melanoma.
In situ hybridization of YKL-40 mRNA expression in biopsies from small cell lung carcinoma shows no YKL-40 mRNA expression in the cancer cells but strong expression in peritumoral macrophages.47 We have also initiated immunohistochemical and in situ hybridization analyses of biopsies from melanoma patients and have demonstrated YKL-40 expression in both melanoma cells and macrophages (Roslind et al., unpubl. obs.).
We examined serum YKL-40 three times a day during a 3-day period in 11 patients before treatment and defined a change of 24% equal to twice the median coefficient of variation to be significant. The objective was to be able to distinguish between normal fluctuations and other changes, such as treatment- or disease-related changes. The majority of patients had an elevation of YKL-40 levels during treatment, but these changes were not significantly correlated with the clinical outcome. Instead, the increase in serum YKL-40 levels may be related to the treatment, as both IL-2 and INF interfere with macrophage activity. This hypothesis has to be explored in further in vitro studies.
Followup blood samples from 12 of the responding patients were collected at 3-month intervals. Serum YKL-40 was never elevated during followup in the patient with a lasting complete response. In contrast, significant increases in serum YKL-40 accompanied recurrence in 9 of 11 patients. In four of the patients these increments were dramatic. Thus, increases in serum levels of YKL-40 in the absence of cytokine treatment or chemotherapy may accompany progression. However, due to the small patient number, this has to be readdressed in another study.
In conclusion, we found an elevated serum YKL-40 to be an independent prognostic factor for poor survival in patients with metastatic melanoma. When combining the levels of serum YKL-40 and LDH, patients could be separated into three prognostic groups based on the number of elevated biomarkers, and this combined variable was independent of the AJCC M stage classification. The potential use of the combination of these two markers as a prognostic index should be validated in an independent study.
The authors thank the staff members of the Departments of Oncology in Aarhus and Herlev for careful management of the patients, and Tonni Loeve Hansen and Debbie Nadelmann, Herlev University Hospital, for expert technical assistance.
- 17The chitinase 3-like protein human cartilage glycoprotein 39 (HC-gp39) stimulates proliferation of human connective-tissue cells and activates both extracellular signal-regulated kinase- and protein kinase B-mediated signalling pathways. Biochem J. 2002; 365: 119–126., , .
- 32High pretreatment serum level of YKL-40 is related to short survival in patients with advanced renal cell carcinoma treated with high-dose continuous intravenous infusion of interleukin-2 [abstract]. Proc ASCO 2003; 22: 1603., , , , .
- 37Electrochemotherapy with low-dose IL-2 in the treatment of disseminated malignant melanoma: clinical and paraclinical evidence of systemic immune response [abstract]. Proc ASCO 2003; 22: 2906., , , , .