Growth/differentiation factor 15 (GDF-15) as new potential serum marker in patients with metastatic colorectal cancer

Abstract

BACKGROUND:

GDF-15 is a protein belonging to the transforming growth factor beta superfamily that has a role in regulating inflammatory and apoptotic pathways. High level GDF-15 in tumor tissues and plasma correlate with an increased risk of recurrence and reduced overall survival.

OBJECTIVE:

The aim of this study was to screen GDF-15 capacity to detecting metastatic CRC and compare it with standard tumor markers CEA and CA19-9.

METHODS:

We collected serum samples from 97 patients with metastatic colorectal cancer and 79 samples from healthy controls. Serum levels of GDF-15, CEA and CA19-9 were measured by immunochemically. A Kaplan-Meier curve was applied for analysis of survival rates, and a log-rank was used for univariate analysis.

RESULTS:

Serum levels of GDF-15 were significantly higher in patients with colorectal cancer compared to healthy controls ( $p<$ 0.001). In addition, serum levels of GDF-15 correlated with extent of liver involvement and patients with higher GDF-15 levels had significantly worse outcome ( $p<$ 0.0001).

CONCLUSIONS:

Our results show GDF-15 as an effective biomarker in patients with metastatic colorectal cancer with the same sensitivity as CEA. In addition, GDF-15 levels strongly correlate with extension of liver involvement in contrast with CEA.

Keywords

GDF-15 colorectal cancer biomarker survival

1. Introduction

Colorectal cancer is the third most common cancer in men (746,000 cases, 10.0% of the total) and the second in women (614,000 cases, 9.2% of the total) worldwide [1]. Patient outcome strongly depends on the stage of tumor at the time of diagnosis and in metastatic disease the prognosis depends on localization and extent of distant metastases [2]. Early detection and diagnosis still present the best chance for successful treatments and improved outcomes. Early detection of metastatic liver disease is important for indication of liver surgery. Therefore, novel biomarkers for earlier cancer detection and early detection of metastatic disease are strongly needed.

GDF-15 also known as non-steroidal anti-inflamm-atory drugs (NSAIDs) activated gene (NAG-1), growth differentiation factor 15 (GDF-15), prostate differentiation factor (PDF), or placental bone morphogenetic protein (PLAB), is a protein belonging to the transforming growth factor beta superfamily that has a role in regulating inflammatory and apoptotic pathways in injured tissues and during disease processes [3]. GDF-15 mRNA is most abundant in the liver, with lower levels seen in some other tissues [4]. GDF-15 over expression is associated with different cancers, including gastric, pancreatic, prostate, colorectal cancer, breast and melanoma [5, 6, 7, 8, 9, 10, 11, 12]. Moreover, GDF-15 is also overexpressed in adenomatous colonic polyps [12].

Many studies have revealed the important role of GDF-15 in CRC. It has been observed that GDF-15 expression levels are markedly enhanced in malignant tissues, established cancer cells and plasma during the transition of numerous localized cancers to invasive and metastatic disease stages as compared with non-malignant tissues, normal cells, and basal GDF-15 concentration in serum [5, 7, 9, 13, 14, 15, 16, 17, 18, 19, 20, 21, 22]. GDF-15 deregulation is involved in the progression of colon cancer, and research showed that the GDF-15 levels in serum gradually increase in the process of conversion from adenomatous polyps to colorectal carcinoma. There is a significant positive relationship between serum GDF-15 levels, clinical stage, presence of metastasis and progression of CRC [20].

Meta-analysis published by Li et al. [23] summarizes the diagnostic and prognostic performance of serum GDF-15. From all studies only 8 were included in the meta-analyses. They summarized that serum GDF-15 levels in CRC patients were higher than those in healthy controls (SMD $=$ 1.08, 95% CI: 0.56–1.59, $P<$ 0.001). For discriminating CRC from healthy controls, the AUC of GDF-15 was 0.816 (95% CI: 0.792–0.838). The sensitivity and specificity were 58.9% (95% CI: 55.0–62.8) and 92.08% (95% CI: 89.2–94.4). Besides, higher GDF-15 expression level was associated with worse overall survival for CRC patients (pooled HR $=$ 2.09, 95% CI: 1.47–2.96). In conclusion, they suggest that serum GDF-15 may be a useful diagnostic and prognostic biomarker for CRC [23]. Only few data are available for metastatic CRC and extension of liver involvement.

In our study, we investigated the serum levels of GDF-15 in the patients with metastatic colorectal cancer compared to the healthy controls. This study compares the levels of GDF-15 to standard biomarkers used in monitoring of colorectal cancer (CEA and CA19-9) and that studies correlation of tumor characteristics and treatment outcomes to the serum levels of GDF-15.

2. Materials and methods:

2.1 Patients and healthy control characteristics

Between November 2011 and May 2013, 97 patients with metastatic CRC and 79 healthy individuals were included in this study. The serum samples from patients were collected before beginning treatment for mCRC ( $n=$ 52) or in the time of progression during palliative treatment for mCRC ( $n=$ 45). From these 77 patients were after resection of primary tumour.

All patients had adequate liver and renal function (transaminases $<$ 2x and creatinine clearance $<$ 1.5x upper normal limit) and signed informed consent. Healthy individuals signed informed consent, underwent healthy examination and had negative colonoscopy.

2.2 Laboratory analyses

Blood for laboratory analyses was collected after overnight fasting via puncture of the cubital vein, simultaneously with blood collection for routine examinations. Routine biochemical parameters were measured in fresh samples. For special parameters, blood was standardly centrifuged for 10 min at 3000 rpm (rotations per minute) and serum was stored at $-$ 80 ${}^{\circ}$ C until analysis. Levels of GDF-15 in serum samples were determined by commercially available ELISA kits (Biovendor-Laboratorni medicina, Czech Republic) according to the manufacturer’s protocols. The sensitivity of the assays was 0.02 ng/ml and intraassay or interassay coefficient of variation (CV) were always less than 10%. Concentrations of CEA a CA 19-9 in serum samples were determined by chemiluminescent immunoanalysis (Architect, Abbott, USA).

2.3 Statistical analysis

Statistical analysis was performed using SAS (SAS Institute Inc., Cary, NC, USA). Basic statistics were calculated for parameters measured in the whole group and in different groups and subgroups, such as mean, standard deviation, variance, median, interquartile range, minimum, maximum. Selected statistical data were also graphically processed, Box & Whisker plot diagrams. Non-parametric analysis of variance two-sided Wilcoxon test was used for comparison of the distribution of the individual parameters in the different groups and subgroups. Due to non-Gaussian distribution of variables Spearman’s correlation coefficient was used to determine the dependency of characters. Statistical significance was determined at the border of alpha $=$ 0.05. Receiver operating characteristic (ROC) curves were generated to assess the diagnostic accuracy of each parameters, the sensitivity and specificity of optimum cut off point were found.

3. Results

3.1 Baseline characteristic of patients and controls

In the cancer group serum 97 patients with generalized colorectal cancer were included (38.1% women and median of age was 64 years). In a majority of patients ( $n=$ 91) histological type was adenocarcinoma, only in 6 patient mucinous adenocarcinoma was diagnosed. Metastatic side was liver in 64 cases (66.0%), lung in 31 cases (32.0%), peritoneal in 14 cases (14.4%) and lymphatic nodules in 22 cases (22.7%). One metastatic side only was diagnosed in 57 patients (58.7%), two in 28 patients (28.9%) and three and more in 12 patients (12.4%). The patient characteristics were published previously [24]. In the control group 79 healthy controls after negative colonoscopy were included (45.6% women and median age 62 years).

In the cancer group serum level of CEA was 308.3 $\pm$ 84.9 ug/l and CA19-9 and was 948.1 $\pm$ 307.0 kIU/l. Both markers were significantly elevated compared to the control group ( $p=$ 0.001; $p=$ 0.006, respectively) [24].

3.2 Serum levels of GDF-15

Serum level of GDF-15 in the cancer group was 7.58 $\pm$ 0.77 ng/ml and was significantly elevated compared to 1.64 $\pm$ 0.08 ng/ml in the healthy control group (Fig. 1C, $p<$ 0.001). In the control group, there was no significant difference in GDP-15 level between men and women ( $p=$ 0.321) and, but there was significant difference in younger than or equal to 55 years ( $N=$ 24) and older than 55 ( $N=$ 55; $p<$ 0.001).

Figure 1.

GDF-15 levels in patients with colorectal cancer and healthy controls ( $p<$ 0.001).

In the cancer group, there was no significant difference between sex, age, part of colon with primary tumor and tumor grade (Table 1). There was indicated difference in presence or absence of the primary tumor ( $p=$ 0.035) and similarly in patients with and without liver metastases ( $p=$ 0.006). Negative association was identified between patients with and without pulmonary metastases ( $p=$ 0.015).

Table 1

Serum GDF-15/MIC-1 levels in different groups

	Serum GDF-15 in	p vallue
	mCRC (ng/ml)
Age ( $n=$ 97), years
$\leqslant$ 55 ( $n=$ 21)	7.47 $\pm$ 1.87	0.940
$>$ 55 ( $n=$ 76)	7.61 $\pm$ 0.85
Sex,
Female ( $n=$ 37)	6.64 $\pm$ 0.94	0.341
Male ( $n=$ 60)	8.16 $\pm$ 1.10
Primary tumor localization,
Right colon ( $n=$ 20) (caecum, ascendens, transversum)	9.39 $\pm$ 2.32
Left colon ( $n=$ 50) (descendens, sigmoideum, rectosigma)	7.25 $\pm$ 0.92	0.307
Rectum ( $n=$ 27)	7.02 $\pm$ 1.49	0.372
Degree of differentiation,
Well or moderate ( $n=$ 75)	7.26 $\pm$ 0.82	0.436
Poorly ( $n=$ 21)	8.70 $\pm$ 1.99
Primary tumor,
Presence ( $n=$ 20)	10.76 $+$ 2.19	0.035 ${}^{*}$
Absence ( $n=$ 77)	3.93 $+$ 0.33
Liver metastases,
Presence ( $n=$ 64)	8.93 $+$ 1.03	0.006 ${}^{*}$
Absence ( $n=$ 33)	4.26 $+$ 0.42
Lung metastases,
Presence ( $n=$ 31)	4.87 $+$ 0.77	0.015 ${}^{*}$
Absence ( $n=$ 66)	8.86 $+$ 1.04

${}^{*}$ Statistically significant.

The group of patients with liver metastases only was divided into patients with the sum of the longest dimension of liver metastases smaller than 100 millimeters ( $N=$ 20) and larger than or equal to 100 millimeters ( $N=$ 29). The difference in GDF-15 levels was statistically significant ( $p=$ 0.002), 5.05 $\pm$ 0.92 mg/ml and 13.36 $\pm$ 1.97 mg/ml, respectively (Fig. 2).

Figure 2.

The serum levels of GDF-15 in patients without pulmonary and with liver metastasis. Sum of liver metastases larger than or equal to 100 mm, smaller than 100 mm and healthy controls.

3.3 Sensitivity and specificity of GDF-15, CEA and CA19-9

ROC curves analysis showed that serum GDF-15 with an AUC of 0.921 (Fig. 3A) and serum level cut-off 3.5 ng/ml has the sensitivity and specificity to distinguish colorectal cancer from healthy controls are 65.98% and 97.47%, respectively ( $p<$ 0.001). GDF-15 has similar sensitivity and specificity compared to CEA (72.16% and 97.47%, respectively) and better sensitivity with same specificity compared to CA-19-9 (46.39% and 96.25%, respectively).

Figure 3.

ROC curve analysis of GDF-15. GDF-15 (AUC $=$ 0.921) with serum level cut-off value $>$ 3.5 ng/ml, the sensitivity and specificity of TFF1 to distinguish colorectal cancer from healthy controls are 71.13% and 98.73%, respectively ( $p<$ 0.001).

3.4 The level of GDF-15 as prognostic factor

The serum level of GDF-15 appears to have a prognostic value. The cut-off for serum GDF-15 level was determined as 3.5 ng/ml. Patients were divided into the group with negative serum levels of GDF-15 ( $\leqslant$ 3.5 ng/ml, $N=$ 33), slight increase ( $>$ 3.5 and $\leqslant$ 7.0 ng/ml, $N=$ 34) and large increase ( $>$ 7.0 ng/ml, $N=$ 30). From all 97 patients overall survival (OS, time from first sample collection to death) was calculated. No statistically significant difference ( $p=$ 0.0530) between group with negative GDF-15 levels (mOS 21.3 months) and group with slight increase (mOS 14.6) was observed. But there was statistically significant difference between negative (mOS 21.3 months) and the group with large increase of GDF-15 serum level (mOS 6.4 months, $p<$ 0.001; Fig. 4). Similar difference was observed between the group with slight increase and large increase of GDF-15 serum level ( $p=$ 0.003; Fig. 4). We performed the similar calculation with CEA and CA19-9, but there was no significant difference found.

Figure 4.

Overall survival according to GDF-15 levels. Kaplan-Meier curves of patients with GDF-15 level $\leqslant$ 3.5 ng/ml (mOS 20.6 months) compared to the patients with GGF-15 level $>$ 3.5– $\leqslant$ 7.0 ng/ml (mOS 14.6 months, $p=$ 0.053) and $>$ 7.0 ng/ml (mOS 6.4 months, $p<$ 0.001).

4. Discussion

Early diagnosis is crucial for successful treatment even in metastatic colorectal cancer with liver involvement. Therefore, new biological markers for early detection (more sensitive and specific than CEA and CA19-9) and predictors of prognosis for colorectal cancer are urgently needed in clinical practice.

Many studies have evaluated the application of GDF-15 in the diagnosis and prognosis of CRC; however, there are some inconsistent results which have been reported. GDF-15 expression level is usually low in resting cells but may be substantially increased following response to diverse cellular stress signals, such as hypoxia, inflammation, light exposure, acute tissue injury and during cancer progression [25, 26, 27].

The meta-analysis published by Li et al. [23] shown results of GDF-15 that achieved a sensitivity of 58.9%, specificity of 92.08% and AUC of 0.816 (similar results as in our analysis), indicating that GDF-15 has a good diagnostic performance for CRC. Moreover, they suggested that GDF-15 expression level could be a prognostic biomarker in CRC patients. Compared with patients with low GDF-15 expression level, patients with an increased level of GDF-15 had a 2.09-fold higher risk of death. These results suggested that GDF-15 could serve as a diagnostic and prognostic biomarker in CRC [23].

Earlier studies identified serum GDF-15 as a new marker for colon cancer, but they are comparing only CRC vs controls without information about stage [28], only few blood samples were available [19, 20, 23] or only limited number of patients in stage IV were examined or assess stage III and IV as one group [28]. Direct comparison with data from other studies is very difficult due to small number of mCRC patients in earlier studies and only limited information about results of this patients separately. Wallin et al. [20] published levels of GDF-15 in mCRC patients (146.5 pg/ml), but only in 8 cases and it was measured in plasma. All studies with patients in stage IV do not present any data about side of distant metastasis and none compare GDF-15 with standard tumor markers (CEA and CA19-9) was not investigated.

We found that serum GDF-15 levels were significantly higher in patients compared to controls and it correlate with the extent of colon or rectum involvement (non-resected primary tumor or local recurrence), also with the extent of liver involvement only. This finding is promising because these patients could be candidates for a radical resection of liver metastases, the only modality than can complete cure them.

The GDF-15 has the same sensitivity and specificity as CEA. These results suggested that serum GDF-15 could be a useful biomarker. The entry level of GDF-15 appears to be a prognostic factor which strongly correlates with overall survival. This can be partially explained by correlation of GDF-15 serum levels with liver involvement (especially in patient without pulmonary metastases). The patients with sum of liver metastases larger than 100 mm had significantly higher levels of GDF-15 and their liver reserve is smaller. Another partial explanation could be, that patients with non-resected primary tumor, who have higher levels of GDF-15, were not operated because of their worse performance status and that may be the reason for a worse outcome of the treatment. Further investigation and data are needed to clarify these results.

5. Conclusions

Our data indicated that serum GDF-15 can serve as an effective serum tumor marker for the detection of distant metastasis with the same sensitivity as CEA and better sensitivity than CA19-9. Our data have shown that GDF-15 strongly correlates with extension of liver involvement in patients without pulmonary metastases, who are the candidates for a curative liver resection. The entry level of GDF-15 appears to be prognostic factor that correlates with overall survival.

Footnotes

Acknowledgments

The study was supported by research projects TIP ČR FR-TI3/666, Progres Q25/LF1 and MH CZ DRO VFN 64165. The authors are thankful to nurses from the Clinical Departments and to laboratory staff for technical assistance.

References

Ferlay

et al., Estimates of worldwide burden of cancer in 2008: GLOBOCAN 2008, Int J Cancer 127 (2010), 2893–2917.

O’Connell

J.B.

et al., Colon cancer survival rates with the new American Joint Committee on Cancer sixth edition staging, J Natl Cancer Inst 96 (2004), 1420–1425.

Zimmers

et al., Growth differentiation factor-15/ macrophage inhibitory cytokine-1 induction after kidney and lung injury, Shock 23 (2005), 543–548.

Hsiao

et al., Characterization of growth-differentiation factor 15, a transforming growth factor beta superfamily member induced following liver injury, Mol Cell Biol 20 (2000), 3742–3751.

Welsh

J.B.

et al., Large-scale delineation of secreted protein biomarkers overexpressed in cancer tissue and serum, Proc Natl Acad Sci USA 100 (2003), 3410–3415.

Baek

K.E.

et al., Upregulation and secretion of macrophage inhibitory cytokine-1 (MIC-1) in gastric cancers, Clin Chim Acta 401 (2009), 128–133.

Koopmann

et al., Serum macrophage inhibitory cytokine 1 as a marker of pancreatic and other periampullary cancers, Clin Cancer Res 10 (2004), 2386–2392.

Golkar

et al., Resveratrol inhibits pancreatic cancer cell proliferation through transcriptional induction of macrophage inhibitory cytokine-1, J Surg Res 138 (2007), 163–169.

Karan

et al., Dysregulated expression of MIC-1/PDF in human prostate tumor cells, Biochem Biophys Res Commun 305 (2003), 598–604.

10.

Bauskin

A.R.

et al., Role of macrophage inhibitory cytokine-1 in tumorigenesis and diagnosis of cancer, Cancer Res 66 (2006), 4983–4986.

11.

Weide

et al., High GDF-15 Serum Levels Independently Correlate with Poorer Overall Survival of Patients with Tumor-Free Stage III and Unresectable Stage IV Melanoma, J Invest Dermatol 136 (2016), 2444–2452.

12.

Buckhaults

et al., Secreted and cell surface genes expressed in benign and malignant colorectal tumors, Cancer Res 61 (2001), 6996–7001.

13.

Lee

D.H.

et al., Macrophage inhibitory cytokine-1 induces the invasiveness of gastric cancer cells by up-regulating the urokinase-type plasminogen activator system, Cancer Res 63 (2003), 4648–4655.

14.

Wollmann

et al., The macrophage inhibitory cytokine integrates AKT/PKB and MAP kinase signaling pathways in breast cancer cells, Carcinogenesis 26 (2005), 900–907.

15.

Brown

D.A.

et al., Macrophage inhibitory cytokine 1: a new prognostic marker in prostate cancer, Clin Cancer Res 15 (2009), 6658–6664.

16.

Nakamura

et al., Quantitative analysis of macrophage inhibitory cytokine-1 (MIC-1) gene expression in human prostatic tissues, Br J Cancer 88 (2003), 1101–1104.

17.

de Wit

N.J.

et al., Analysis of differential gene expression in human melanocytic tumour lesions by custom made oligonucleotide arrays, Br J Cancer 92 (2005), 2249–2261.

18.

Barderas

et al., In-depth characterization of the secretome of colorectal cancer metastatic cells identifies key proteins in cell adhesion, migration, and invasion, Mol Cell Proteomics 12 (2013), 1602–1620.

19.

Brown

D.A.

et al., MIC-1 serum level and genotype: associations with progress and prognosis of colorectal carcinoma, Clin Cancer Res 9 (2003), 2642–2650.

20.

Wallin

et al. Growth differentiation factor 15: a prognostic marker for recurrence in colorectal cancer, Br J Cancer 104 (2011), 1619–1627.

21.

Mehta

R.S.

et al., Association Between Plasma Levels of Macrophage Inhibitory Cytokine-1 Before Diagnosis of Colorectal Cancer and Mortality, Gastroenterology 149(3) (2015), 614–622.

22.

Wang

et al., Macrophage inhibitory cytokine 1 (MIC-1/GDF15) as a novel diagnostic serum biomarker in pancreatic ductal adenocarcinoma, BMC Cancer 14 (2014), 578–588.

23.

et al., Growth differentiation factor 15 is a promising diagnostic and prognostic biomarker in colorectal cancer, J Cell Mol Med 20(8) (2016), 1420–1426.

24.

Vocka

et al. Trefoil factor family (TFF) proteins as potential serum biomarkers in patients with metastatic colorectal cancer, Neoplasma 62(3) (2015), 470–477.

25.

Ding

et al., Identification of macrophage inhibitory cytokine-1 in adipose tissue and its secretion as an adipokine by human adipocytes, Endocrinology 150 (2009), 1688–1696.

26.

Schober

et al., Expression of growth differentiation factor-15/macrophage inhibitory cytokine-1 (GDF-15/MIC-1) in the perinatal, adult, and injured rat brain, J Comp Neurol 439 (2001), 32–45.

27.

Ago

et al. Upregulation of Nox4 by hypertrophic stimuli promotes apoptosis and mitochondrial dysfunction in cardiac myocytes, Circ Res 106 (2010), 1253–1264.

28.

Mehta

R.S.

et al., A prospective study of macrophage inhibitory cytokine-1 (MIC-1/GDF15) and risk of colorectal cancer, J Natl Cancer Inst 106 (2014), dju016.