Abstract
Background: Bone metastases develop in several malignancies (multiple myeloma, breast, prostate and lung carcinoma) and cause several complications. The aim of this study was to search for new biomarkers to use in monitoring of bone metastatic disease with the use of xMAP technology. Patients and Methods: We assessed 62 oncological patients: 23 with no bone metastases, 28 with metastatic disease not having undergone therapy and 11 with metastatic disease treated by denosumab. Serum levels of dickkopf-related protein 1 (DKK1), growth differentiation factor-15 (GDF15), neuron-specific enolase (NSE), osteoprotegerin (OPG), osteonectin, periostin, tartrate-resistant acid phosphatase (TRAP5), tumor necrosis factor related weak inducer of apoptosis (TWEAK), chitinase-3-like protein 1 (YKL40), carboxy-terminal telopeptide (CTX) and procollagen type 1 N-terminal propeptide (PINP) were measured in each sample. Results: The following biomarkers were observed to have significantly higher levels in the groups of patients with metastases in comparison to metastasis-free patients: GDF15 (p<0.0001), osteonectin (p=0.0311), TRAP5 (p<0.0046), TWEAK (p<0.0343) and YKL40 (p<0.0034). The changes in DKK1, NSE, OPG and periostin were not significant. Conclusion: We identified five new biomarkers: GDF15, osteonectin, TRAP5, TWEAK, and YKL40 as being promising markers for monitoring bone metastases.
- Cancer
- bone metastasis
- scintigraphy
- biomarkers
- multiplex technology
Bone metastases develop in several malignancies, most commonly in multiple myeloma, breast carcinoma, carcinoma of lung and prostate, and cause several complications: pain, pathological fractures, hypercalcemia and others, all of which reduce a patient's quality of life. The progression of bone metastasis is associated with a poor prognosis and short survival. The development of metastatic disease is a complex process. It was found that some genes expressed by tumor cell lines were common for metastases to all sites, whereas others were selectively expressed if the cell lines had a predilection to grow in a given tissue. The vicious circle related to bone metastasis develops when factors secreted by or expressed on tumor cells (e.g. parathyroid hormone-related peptide) activate osteoblasts and osteoclasts in the bone microenvironment to produce cytokines; bone remodeling and osteolysis cause release of growth factors, which then stimulate tumor cell growth, motility and further release of e.g. parathyroid hormone-related peptide (1, 2). New molecules are still being shown to have a role in this process and are also potential candidate biomarkers.
Current diagnostics of bone metastatic disease is based on imaging methods, which are not satisfactory for early detection or regular treatment monitoring. Today, there is accessible and effective biological treatment for patients but early detection is required. New biomarkers, in combination with traditional markers of bone turnover, could help to improve the strategy for managing bone-metastatic disease and this is why this is a hotspot for further research. xMAP multiplex technology, enables the simultaneous measurement of large numbers of circulating biomarkers in a small sample volume and could be a useful tool for the evaluation of biomarkers in oncology, as we have shown in previous studies carried out in our laboratory on different oncology topics (3-5). New immunoanalytic technologies could open the door to regular patient monitoring using only blood sampling, with no doses of radiation, contrary to currently used monitoring by imaging methods.
The present study focused on testing the Human Cancer/Metastasis Biomarker Panel 1, which is a commercially available multiplex panel for the measurement of the marker serum levels using xMAP technology. This panel contains nine biomarkers that have been identified for a variety of tumor types: Dickkopf-related protein 1 (DKK1), growth differentiation factor-15 (GDF15), neuron-specific enolase (NSE), osteoprotegerin (OPG), osteonectin, periostin, tartrate-resistant acid phosphatase (TRAP5), tumor necrosis factor related weak inducer of apoptosis (TWEAK), chitinase-3-like protein 1 (YKL40). We tested this panel for the detection of tumor-induced bone disease (bone metastases) and compared it to serum bone markers now routinely used in the monitoring of several bone diseases: carboxy-terminal telopeptide (CTX) and procollagen type 1 N-terminal propeptide (PINP). We evalutated the relationship of the levels of circulating bone markers to the results of whole-body skeletal scintigraphy.
Patients and Methods
Group of patients. The multiplex panel of biomarkers was studied in a cohort of 62 oncological patients with solid tumors undergoing whole-body skeletal scintigraphy using technetium (99mTc). The assessed group included patients with the following diagnosis: prostate cancer, breast cancer, lung cancer and colorectal cancer (CRC). According to the results of scintigraphy, the cohort was divided into groups based on the occurrence of bone metastases. In our study, group 0 was a control group of 23 patients with no signs of bone metastases (age: median=68 years, minimum=48 years, maximum=85years). We enrolled two groups of patients with bone metastases: group 1 represents 28 patients with no therapy (age: median=65 years, minimum=35 years, maximum=83 years); group 2 represents 11 patients who were treated before scintigraphy with denosumab (median age=64 years, minimum=52 years, maximum= 83 years). The Kruskal–Wallis test showed the groups of patients did not differ significantly in age. The study was approved by the local Ethical Committee on 13th July 2011 and all the patients signed an informed consent form before participating in this study.
Plasma samples. Peripheral blood was drawn before scintigraphy using VACUETTE® Z Serum Sep tubes (Greiner Bio-One, Kremsmünster, Austria) and allowed to clot. Sera were separated by centrifugation at 1,700 × g for 10 min and all specimens were immediately aliquoted and frozen. Samples were stored at −80°C. No more than one freeze-thaw cycle was allowed before analysis.
Analytical methods used. Serum levels of DKK1, GDF15, NSE, OPG, osteonectin, periostin, TRAP5, TWEAK and YKL40 were assayed using a Human Cancer Metastasis Biomarker Magnetic Bead Panel (Merck Millipore, Darmstadt, Germany). CTX levels were measured by the chemiluminescent immunoassay β-CrossLaps Roche (F. Hoffmann–La Roche, Basel, Switzerland) and PINP levels were measured using the radioimmunoassay UniQ PINP RIA (Orion Diagnostica, Espoo, Finland). Analytical ranges of the used assays: DKK1 (0.007-30), YKL40 (0,048-20), TWEAK (0.012-50), TRAP5 (0.019-80), periostin (0.195-800), osteonectin (0.976-4.0), OPG (0.007-30), NSE (0.036-150), GDF15 (0.002-10), CTX (0.01-6.00) all in ng/ml and PINP (5-250 μg/l).
Statistical methods. Descriptive statistics, i.e. the median, the 25th and the 75th percentiles, were calculated for all markers. The Kruskal–Wallis test was used to make a comparison of marker levels between all three groups. Significance was set for p-values lower than 0.05. If the Kruskal–Wallis test was positive (p<0.05) then post hoc analysis was performed: a test for pairwise comparison of subgroups according to Conover (6). Furthermore, for markers where a difference between groups was found, receiver operating characteristic (ROC) curves were drawn and the area under the curve (AUC) was calculated. Because the multiplex is not a routine method, mean coefficients of variation (CV%; percentage relative standard deviation) were calculated for all multiplexed analytes. Results outside of the calibration curve ranges were stated as the value of the lowest or highest calibration point.
Results
A summary of descriptive statistics are shown in Table I. The differences among groups, AUC values and 95th percentiles are shown in Table II. Regarding the biological activity markers included in the multiplex panel, only GDF15, osteonectin, TRAP5, TWEAK and YKL40 differed significantly between groups. Both classic routine bone markers exhibited differences between groups. Osteonectin levels only differed significantly between group 0 and 2. All differing markers had higher levels in group 1 compared to 0, with the exception of TWEAK, which had lower levels in group 1 compared to 0, but without significant differences from group 2. All studied markers can be ordered by their decreasing ability to distinguish the occurrence of bone metastasis according to decreasing AUC (ROC analysis, group 0 versus group 1) as follows: PINP, GDF15, TRAP5, TWEAK, YKL40, osteonectin and CTX.
Mean CVs for doublets in multiplex analysis were below 6.5% for all analytes; the best was for periostin (3.31%) and the worst for TWEAK (6.11%).
Discussion
Multiplex analytic technology xMAP, used in our study, is an open-architecture system based on color-coded magnetic microspheres coated by antibodies to perform quantitative sandwich immunoanalysis of proteins. Reactions are then read in a Magpix analyzer, a compact fluorescence-based LED detection system, capable of simultaneously measuring up to 50 targets of interest in a single microtiter well. Multiplexing of assays enables similar and standard analytic conditions for all proteins measured in one well. This feature of multiplexing is most valid for comparison of novel biomarkers without bias. Multiplexing reduces radically sample volume, labor -time and costs over traditional methods such as ELISA or RIA.
In our study, the most promising new biomarker from the multiplex panel was shown to be GFD15. It is the only known secreted p53-regulated cytokine and therefore can serve as a biomarker for p53 activation both in vitro and in vivo. GDF15 may play antitumoral role during the early stages of cancer but conversely, it can promote invasiveness and metastatic behavior at advanced stages, and it is involved in the epithelial–mesenchymal transition in tumors (7). Roles for GDF15 have recently been identified in modulating osteoclast differentiation and for therapy for bone metastases from prostate cancer (8). In our study we found higher levels in patients with bone metastases compared to the controls. We found no significant change in GDF15 levels in group 2, patients who had undergone biological treatment for metastases. In the literature, high plasma levels of GDF15 before a diagnosis of CRC are associated with greater CRC-specific mortality (9). GDF15 could be considered a novel prognostic marker for CRC for clinical practice. Serum GDF15 levels are significantly elevated in most cases of pancreatic ductal adenocarcinoma (PDAC), including those negative for cancer antigen 19-9 and with early-stage disease, and thus may serve as a novel diagnostic marker for early diagnosis and postoperative monitoring of PDAC (10). Westhrin et al. described the role of GDF15 in osteoclast differentiation and showed an association of high serum GDF15 level with bone disease in multiple myeloma. In their study, the same analytical technology was used (11).
A further promising marker for bone metastasis in the multiplex panel is TRAP5. It is one of the most abundant enzymes in osteoclasts and plays a role in bone resorption. Elevated TRAP levels in serum can be caused by Paget's disease, hemodialysis, primary hyperparathyroidism, metastatic malignancies involving bone resorption, multiple myeloma and bilateral ovariectomy (12). In our study, we found elevated levels in patients with bone metastases without treatment (group 1) compared to the control group. Sarvari et al. found that serum TRAP5 was not markedly elevated in those with limited bone metastases but was strongly elevated in those with extensive bone metastases (13). Our results are in accordance with those of Tang et al., who found higher TRAP5 values in patients with bone metastases compared to patients with no bone metastases (14). In our study, TRAP5 was the only multiplex marker showing a significant difference in patients undergoing biological treatment (denosumab; group 2) compared to untreated patients (group 1). We observed a decrease in patients after denosumab treatment.
Chitinase-3-like protein 1 commonly known as YKL40 molecule or CHI3L1, is a member of the mammal chitinase family and it is a matrix protein of specific granules in human neutrophils. Furthermore YKL40 is secreted by chondrocytes, synovial cells, and macrophages. YKL40 is suspected to have a role in remodeling or degradation of the extracellular matrix (15). In our study, we found higher serum levels of YKL40 in the bone-metastatic groups compared to the controls. This is in accordance with the literature findings where high pre-treatment serum levels of YKL40 are described as being associated with a poor outcome of metastatic prostate cancer and non-small cell lung cancer (NSCLC), and a marker for early death in prostate cancer and so could serve as a new prognostic biomarker in patients (16-18).
TWEAK mediates different biological effects such as proliferation, angiogenesis, and osteoclastogenesis. These effects help tumor cells to survive and spread. Williams et al. highlight that TWEAK plays a role in the progression of multiple myeloma and may facilitate bone destruction (19). In our study, we recorded higher levels in patients with solid tumor and bone-metastatic disease compared to patients without metastases in bones. This could show a role of TWEAK in solid tumor spread into bones. TWEAK may be a potential molecule for further studies as a biomarker for stratification of patients.
Osteonectin, which is also known as secreted protein, acidic and rich in cysteine (SPARC) and basement-membrane protein 40 (BM40), has a major role in bone mineralization and collagen binding. Osteonectin affects cell–matrix interactions by accelerating the production and increasing the activity of matrix metalloproteinases, which helps cancer cells to enter the bone, along with promoting angiogenesis, proliferation and migration. Overexpression of osteonectin is reported in many human cancer types such as of the breast, prostate and colon (20). We only found higher levels of osteonectin in the group of treated patients with bone metastases. While we tested the novel analytical multiplex panel, we admit to a higher level of analytical imprecision for osteonectin in serum. This imprecision is caused by calibration curve ranges. Many sample results for osteonectin were at the very low end of the sigmoidal calibration curve i.e. in a flat area of calibration, which causes poor resolution of such samples.
OPG is known for its role in bone metabolism and is also related to the activity of bone metastasis (21). Contrary to our study, serum OPG levels are described in the literature as being increased in patients with bone-metastatic disease in prostate cancer compared with patients with organ-confined disease. Serum OPG increases may indicate disease progression or relapse in prostate cancer (22, 23).
Another molecule measured in our study was the small secretary protein DKK1. It is known to be involved in proliferation, differentiation, survival, apoptosis, and cell motility. Serum levels of DKK1 increase in patients with different cancer types and higher expression has been reported to be associated with, worse prognosis, spread of metastases, and poorer survival. Liang et al. reported that serum DKK1 is even a potential biomarker for screening of gastrointestinal cancer (24). In one of the latest studies performed by Rachner et al., multivariate analyses defined a high serum level of DKK1 as an independent prognostic marker of poor survival in prostate cancer (25). We found no relationships between DKK1 level and bone-metastatic disease in our study.
NSE is a routinely used tumor marker of small cell lung cancer. It is a universal marker of the neuroendocrine nature of cells within a tumor. It has been described that high levels of NSE signal a worse prognosis e.g. for NSCLC or prostate cancer, but we do not suppose there to be a close connection to the spread of bone tumor. We did not find any significant differences in the level of NSE between groups.
Another molecule for which no changes in correlation to bone metastases were found in our study is periostin. Periostin serves as a cell-adhesion molecule, that might promote invasion in a tumor microenvironment. High expression of periostin is reported in most solid tumors, including breast, colon, head and neck, pancreatic, lung, papillary thyroid, ovarian, gastric and liver carcinoma, as well as neuroblastoma (26). Our observations were contradictory to results presented by Sasaki et al., who described serum periostin levels to be elevated in patients with bone metastases from breast cancer compared to patients without evidence of bone metastasis (27).
PINP and CTX as routine markers of turnover of collagen I in bone remodelation are widely discussed as markers of bone metastasis. One study, performed by Lumachi et al., showed that CTX and PINP levels are useful in monitoring of patients with NSCLC to aid early detection of bone metastases (28). However, these markers are still not in routine use in oncology for bone metastasis detection.
Treatment, monitoring and tailoring are the most desired future functions of markers in bone-metastatic disease management. For example Lester et al. evaluated a bone marker-directed schedule of treatment with zolendronic acid based on levels of the bone resorption marker urinary N-telopeptide of type I collagen. Their experience suggests that a tailored approach to bisphosphonate therapy may be a more cost-effective approach than the currently licensed and recommended fixed schedule of intravenous treatment (29). Treatment monitoring would enable improvement of cost-effectiveness, the reduction of doses through individual tailoring, and hopefully a decrease in the severity of adverse therapy effects. The necessity for proper choice of individualized patient treatment by novel targeted antibody drugs is currently a hot topic. However, one problem is that bone markers are mainly studied separately, e.g. amino-terminal collagen crosslinks (NTX) levels in the BISMARK study (30) and so their comparison is very complicated.
In conclusion, using the novel multiplex panel, we identified five new biomarkers: GDF15, TRAP5, TWEAK, and YKL40 which are promising markers for bone metastasis monitoring. The best marker of bone metastases in our study was PINP, but surprisingly the five novel markers performed better than CTX. We are continuing with a study of novel markers in larger cohorts with the aim of providing multiparametric evaluation in conjunction with multiplex analysis of biomarkers for bone-metastatic disease.
Acknowledgements
This study was supported by the research project of Ministry of Health of The Czech Republic IGA NT 13655-4.
- Received January 25, 2016.
- Revision received March 8, 2016.
- Accepted March 18, 2016.
- Copyright© 2016 International Institute of Anticancer Research (Dr. John G. Delinassios), All rights reserved