Life Sciences 130 (2015) 18–24

Contents lists available at ScienceDirect

Life Sciences journal homepage: www.elsevier.com/locate/lifescie

Evaluation of serum midkine as a biomarker in differentiated thyroid cancer Zhaowei Meng a,⁎, Jian Tan a,⁎⁎, Guizhi Zhang a, Weijun Tian b, Qiang Fu b, Weidong Li b, Xianghui He b, Shuanghu Wu b, Zhiqiang Yang b, Xiaoyu Liang b, Liyan Dong b, Qing Zhang c, Li Liu c, Yujie Zhang d, Ke Xu e, Boning Liu e, Ning Li a, Xue Li a, Qiang Jia a, Yajing He a, Sheng Wang a, Renfei Wang a, Wei Zheng a, Xinghua Song a,f, Jianping Zhang a, Tianpeng Hu a, Na Liu a, Arun Upadhyaya a a

Department of Nuclear Medicine, Tianjin Medical University General Hospital, Tianjin, PR China Department of General Surgery, Tianjin Medical University General Hospital, Tianjin, PR China c Department of Health Management, Tianjin Medical University General Hospital, Tianjin, PR China d Department of Pathology, Tianjin Medical University General Hospital, Tianjin, PR China e Tianjin Key Laboratory of Lung Cancer Metastasis and Tumor Microenviroment, Tianjin Lung Cancer Institute, Tianjin Medical University General Hospital, Tianjin, PR China f Department of Nuclear Medicine, Second Affiliated Hospital of Zhejiang Medical University, Hangzhou, PR China b

a r t i c l e

i n f o

Article history: Received 10 October 2014 Received in revised form 18 February 2015 Accepted 28 February 2015 Available online 25 March 2015 Keywords: Midkine (MK) Differentiated thyroid cancer (DTC) Thyroglobulin (Tg) 131 I treatment Receiver operating characteristic (ROC) curves Kaplan–Meier analysis method

a b s t r a c t Aims: Midkine is a multifunctional cytokine identified to be a promising cancer biomarker. We aimed to prospectively investigate serum midkine as a diagnostic and prognostic biomarker in differentiated thyroid cancer (DTC). Main methods: 162 patients with thyroid nodules participated in the surgical cohort (post-surgical pathology proved 70 cases with DTC and 92 cases with benign thyroid nodules), 75 healthy subjects served as control. Diagnostic values of pre-surgical midkine and thyroglobulin for DTC were conducted by receiver operating characteristic (ROC) curves. 214 DTC patients participated in the 131I treatment cohort. Prognostic values of pre-131I-ablative midkine and thyroglobulin to predict 131I-avid metastases were performed by ROC curves. Metastasis-free survival was analyzed by the Kaplan–Meier method. Key findings: Much better diagnostic capability of midkine than thyroglobulin was shown to differentiate DTC from benign thyroid nodules, with cut-off midkine value of 323.12 pg/ml and diagnostic accuracy of 75.31%. Nearly similar diagnostic capabilities of midkine and thyroglobulin were shown to distinguish DTC from normal participants. Pre-131I-ablative thyroglobulin demonstrated perfect ability to predict metastases, with cut-off value and diagnostic accuracy of 19.50 ng/ml and 96.73%. Midkine also performed well with a cut-off value and diagnostic accuracy of 504.71 pg/ml and 89.25%. DTC patients with midkine or thyroglobulin levels higher than those of thresholds (500 pg/ml or 20 ng/ml) showed a significantly worse 131I-avid metastasis-free survival by the Kaplan–Meier method (P b 0.01). Significance: Our results show that midkine is as good as or even better than thyroglobulin to screen patients with thyroid nodules for DTC before surgery, and to predict whether metastases exist before the first 131I ablative therapy. © 2015 Elsevier Inc. All rights reserved.

1. Introduction The incidence of thyroid cancer has been increasing rapidly worldwide [35]. The follow-up strategy for differentiated thyroid cancer (DTC) combines the use of 131I whole body scan and serum thyroglobulin (Tg) determination [2,28,29]. Studies have shown a good value of Tg ⁎ Correspondence to: Z. Meng, Department of Nuclear Medicine, Tianjin Medical University General Hospital, Anshan Road No. 154, Heping District, Tianjin 300052, PR China. Tel.: +86 15620882272, +86 18622035159; fax: +86 22 27813550. ⁎⁎ Correspondence to: J. Tan, Department of Nuclear Medicine, Tianjin Medical University General Hospital, Anshan Road No. 154, Heping District, Tianjin 300052, PR China. Tel.: +86 22 60362868; fax: +86 22 27813550. E-mail address: [email protected] (Z. Meng), [email protected] (J. Tan).

http://dx.doi.org/10.1016/j.lfs.2015.02.028 0024-3205/© 2015 Elsevier Inc. All rights reserved.

for prognostic prediction before the first 131I ablative therapy and for monitoring the whole process of disease progression [1,19,36]. However, some studies did demonstrate undetectable pre-operative Tg in patients with DTC [3], and low post-operative non-stimulated Tg in patients with 131I-avid metastases [31]. Moreover, pre-operative Tg measurement is generally considered to be an insensitive and nonspecific diagnostic test for thyroid cancer [2,4,28,29]. Thus, it is necessary to identify a new serologic biomarker with both sufficient sensitivity and specificity to detect DTC before surgery and to predict prognosis. Midkine (MK) is a heparin-binding growth factor, initially found as the product of a retinoic acid induced gene [13,23]. During the past 25 years or so, there has been increasing evidence indicating that MK played an essential role in carcinogenesis in various types of

Z. Meng et al. / Life Sciences 130 (2015) 18–24

malignancies [10,12,24,25,32]. Moreover, since MK is a secretory factor which is measurable, many studies reported serum MK as a promising tumor marker and MK levels were closely correlated with cancer patients' outcome [7–9,17,18,21,22,26,27,30,32,34,38]. Regarding MK and thyroid cancer, our recent investigation discovered that MK immunohistochemistry could potentially be used for differential diagnosis between papillary thyroid cancer (PTC) and multi-nodular goiter, and for the prediction of synchronous metastases [37]. However, the diagnostic and prognostic values of serum MK for DTC, particularly before surgery and before 131I ablative therapy, have not yet been comprehensively studied. In this study, we aimed to systematically evaluate the diagnostic value of serum MK for DTC before thyroidectomy. We also intended to evaluate the prognostic value of serum MK before the first 131I ablative therapy. Comparison between MK and Tg was to be performed. Parameters would be compiled and statistically analyzed to determine sensitivity, specificity, accuracy, positive predictive value (PPV) and negative predictive value (NPV) for differential diagnosis or prognosis. Metastasis-free survival was analyzed by the Kaplan–Meier method. 2. Patients and methods 2.1. Patients From January 2011 to January 2012, the surgical cohort was enrolled as a prospective investigation in the surgical department of our hospital. In this cohort, 162 cases with thyroid nodules were recruited. In this study, we intended to determine the diagnostic value of pre-surgical serum MK for DTC before thyroidectomy, and to compare it with Tg. All the participants with DTC received total thyroidectomy, and neck lymph node resection if metastasis was suspected. Participants with benign thyroid nodules received either lobectomy or nodule resection. Surgery and hisptopathology were conducted in our hospital. As a result, 70 patients were diagnosed with DTC (68 cases of PTC and 2 cases of follicular thyroid cancer [FTC]) and 92 patients were diagnosed with benign thyroid nodules (72 cases of multinodular goiter and 20 cases of thyroid adenoma). Serum MK and other parameters were tested before operation. From January 2011 to May 2013, the 131I treatment cohort was enrolled as a prospective investigation in the nuclear medicine department of our hospital. This cohort comprised of 214 patients diagnosed with DTC (including the 70 cases which participated in the prior surgical cohort and 144 cases who did not enter that cohort, 202 cases of PTC and 12 cases of FTC). All the patients were given 131I for thyroid ablation (approximately 100 mCi), and later for eradication of metastases (approximately 100 mCi to 200 mCi each time, according to the metastatic situations). Five days after 131I administration, a full body scan was performed by using a dual-detector SPECT/CT equipped with high-energy

19

collimators (Discovery VH; General Electric Medical Systems, Milwaukee, Wisconsin, USA). Repeated 131I therapies were given with intervals of approximately 3 to 4 months, if obvious residue thyroid tissue existed or metastases were detected. Clinical assessments (including serum MK and other parameter measurements, evaluation of thyroid remnant ablation completeness and whether metastases existed) were done during each 131I therapy along with the readings of the 131I whole body scans. The criteria of successful ablation were defined as negative serum Tg levels and no uptake of residue thyroid tissue on the whole body scan. Metastases were defined as elevated serum Tg levels and 131I-avid lesions on the whole body scan and/or lesions discovered by ultrasound or CT. All DTC patients were followed for at least 12 months after the first ablation. Later, in order to do the analyses, the 131I treatment cohort was divided into two subgroups accordingly (subgroup 1 had successful ablation and no metastases, and subgroup 2 had 131I-avid metastases). All management procedures were performed in accordance with the American and European guidelines [2,29]. From June 2013 to September 2013, the normal control cohort was enrolled from the health management department of our hospital. This cohort included 75 healthy participants with no diseases reported, who had come to receive their annual health checkup. Serum MK and other parameters were tested during the health checkup. The Institutional Review Board of Tianjin Medical University General Hospital approved the ethical and methodological aspects of this investigation. All participants provided their written informed consent to participate in this study. 2.2. Serum parameter evaluation Free triiodothyronine (FT3, reference 3.50–6.50 pmol/L), free thyroxine (FT4, reference 11.50–23.50 pmol/L) and thyroid stimulating hormone (TSH, reference 0.30–5.00 μIU/mL, maximum 150.00 μIU/mL) assays were performed on a fully automated ADVIA Centaur analyzer (Siemens Healthcare Diagnostics, New York, USA). These assays were based on the chemiluminescent reaction principle. Tg (reference 0– 55.00 ng/mL, maximum 300.00 ng/mL) and Tg antibody (TgAb, reference 0–40.00 IU/mL, maximum 3000.00 IU/mL) were also assessed by a chemiluminescent reaction on a fully automated IMMULITE 2000 analyzer (Siemens Healthcare Diagnostics, Los Angeles, USA). 2.3. Serum MK measurement Blood samples were obtained by venipuncture and immediately centrifuged at 3000 ×g for 10 min. Serum samples were collected, aliquoted, and stored frozen at −80 °C until use. Sample MK level measurement was performed by enzyme-linked immunosorbent assay (ELISA) method. We used a commercial kit (DuoSet ELISA, R&D systems Inc., Minneapolis, MN 55413, USA). The assay was conducted according

Table 1 Data comparison in the surgical and the control cohorts. Group⁎ (case number)

Age (years old)

MK⁎⁎ (pg/ml)

Tg⁎⁎ (ng/ml)

TgAb⁎⁎ (IU/mL)

FT3⁎⁎ (pmol/L)

FT4⁎⁎ (pmol/L)

TSH⁎⁎ (μIU/mL)

1 (75) 2 (70) 3 (92)

48.63 ± 17.25 49.34 ± 13.68 48.13 ± 12.97

255.01 ± 126.78 838.90 ± 916.93 266.86 ± 133.12

8.44 ± 17.76 48.39 ± 81.00 17.44 ± 24.22

27.00 ± 39.89 29.98 ± 87.59 28.59 ± 40.37

4.54 ± 0.52 4.45 ± 0.54 4.50 ± 0.57

15.74 ± 1.79 15.80 ± 1.73 16.01 ± 1.98

3.71 ± 1.82 3.68 ± 1.65 3.60 ± 2.25

T value(1):(2)# P value(1):(2)# T value(1):(3)# P value(1):(3)# T value(2):(3)# P value(2):(3)#

0.276 0.783 0.212 0.832 0.576 0.566

5.460 b0.001 −0.584 0.560 5.908 b0.001

4.167 b0.001 −2.684 0.008 3.470 0.001

0.266 0.790 −0.254 0.799 0.134 0.893

−1.018 0.310 0.457 0.648 −0.574 0.567

0.223 0.824 −0.924 0.357 −0.696 0.487

−0.114 0.909 0.342 0.733 0.241 0.810

⁎ Group 1 = normal control group, group 2 = thyroid cancer group, group 3 = benign thyroid nodule group. ⁎⁎ MK = midkine, Tg = thyroglobulin, TgAb = thyroglobulin antibody, FT3 = free triiodothyronine, FT4 = free thyroxine, TSH = thyroid stimulating hormone; # Analyzed by independent samples T test.

20

Z. Meng et al. / Life Sciences 130 (2015) 18–24

Fig. 1. Scatter graphs were plotted to demonstrate the inter-group differences of pre-surgical serum midkine (A) and thyroglobulin (B) in patients as well as in control subjects.

Fig. 2. Diagnostic and prognostic values of serum midkine in differentiated thyroid cancer. Receiver operating characteristic curves were drawn to assess diagnostic capabilities of pre-surgical midkine and pre-surgical thyroglobulin to distinguish between patients with differentiated thyroid cancer and patients with benign thyroid nodules (A), or between patients with differentiated thyroid cancer and normal control participants (B). A receiver operating characteristic curve was drawn to determine diagnostic capabilities of pre-131I-ablative midkine and pre-131I-ablative thyroglobulin to discern whether or not metastases existed in patients with differentiated thyroid cancer (C). Metastasis-free survival curves were drawn in differentiated thyroid cancer patients with high or low concentrations of pre-131I-ablative midkine (D) and pre-131I-ablative thyroglobulin (E). Patients with midkine level of b500 pg/ml or ≥500 pg/ml had 94.97 % or 29.09 % twelve-month 131I-avid metastasis-free survival rates, Log-Rank test showed a significant difference between the two groups (P b 0.01). Patients with thyroglobulin of b20 ng/ml or ≥20 ng/ml had 99.38 % or 11.54 % 131I-avid metastasis-free survival rates, Log-Rank test demonstrated a significant difference between the two groups (P b 0.01).

Z. Meng et al. / Life Sciences 130 (2015) 18–24

to the manufacturer's instructions and values were reported as pg/ml. Briefly, 100 μl of each serum sample or standard was incubated in a 96-well microplate for 2 h at room temperature. The microplate was pre-coated with goat anti-human MK antibody, and then blocked by 1% bovine serum albumin. After washing three times, biotinylated goat anti-human MK antibody was added and incubated with captured MK for 2 h at room temperature. After another three washes, 100 μl aliquots of streptavidin-conjugated horseradish-peroxidase were added and allowed to react for 30 min in a dark place. After plate washing, substrate solutions (1:1 mixture of H2O2 and tetramethylbenzidine) were added to the wells (100 μl per well) for a 20-minute reaction. Finally, 1 mol/L H2SO4 (stop solution) was added (50 μl per well), and the optical densities of the wells were measured at 450 nm with a Multiskan MS Plate Reader (Labsystems, Helsinki, Finland). After creating a standard curve by a four parameter logistic curve-fit method, concentrations of the samples were determined. All specimens were tested blinded and in duplicate. 2.4. Statistical analysis All data were presented as mean ± standard deviation. Statistics were performed with SPSS 17.0 (SPSS Incorporated, Chicago, Illinois, USA). Differences of indices between the two groups were analyzed by independent samples T test. The χ2 test was used to check whether sex had a significant influence on the inter-group differences. Pearson and Spearman bivariate correlations were made among the variables. Receiver operating characteristic (ROC) curves were drawn and diagnostic efficacies were determined by comparing the areas under the curves. Then, optimal cut-off values were selected, and sensitivity, specificity, diagnostic accuracy, PPV and NPV for differential diagnosis or prognosis were assessed, respectively. Metastasis-free survival was plotted using the Kaplan–Meier method and analyzed by the LogRank test. A P value not exceeding 0.05 was considered as statistically significant. We performed the statistics. 3. Results 3.1. MK showed much better performance for differential diagnosis before surgery than Tg To investigate the role of MK as a tumor marker for DTC, serum MK levels were first analyzed by ELISA in the surgical and the control cohorts. We compared MK with Tg, a known marker for thyroid cancer, which is not recommended before surgery nevertheless. Pre-surgical MK levels and pre-surgical Tg levels were found to be significantly higher in the thyroid cancer patients than in the normal control subjects, as well as that in the benign thyroid nodule patients (Table 1). Scatter graphs of MK and Tg were plotted to demonstrate the intergroup differences as well (Fig. 1). All the other parameters showed no significant differences among the groups. Gender did not have any substantial impact on diagnosing DTC from normal control cases, nor on the differential diagnosis between malignant and benign thyroid nodules (Supplementary Table 1). Pearson and Spearman bivariate correlations were carried out. Supplementary Table 2 disclosed that MK and Tg were closely and positively correlated with each other (P b 0.01). Supplementary Table 2 also showed that MK and age were also positively correlated with each other (P b 0.05). Correlation coefficients among other factors were not significant. Diagnostic efficacy of pre-surgical MK, in comparison with presurgical Tg, was carried out in two separate analyses. In the first analysis, differentiation between thyroid cancer and benign thyroid nodule was conducted. A ROC curve was drawn in Fig. 2A. Diagnostic capability of MK was demonstrated to be much better than Tg. Table 2 indicated that the area under the curve of MK was much higher than that of Tg. Furthermore when optimal cut-off values were determined (323.12 pg/ml for

21

Table 2 Differentiation between thyroid cancer and benign thyroid nodule according to the ROC curve-related data. MK⁎ Area under the curve Optimal cut-off value Sensitivity(%) Specificity(%) Accuracy(%) PPV* (%) NPV* (%)

0.834 323.12 pg/ml 75.70% 75.00% 75.31% 69.74% 80.23%

Tg⁎ 0.647 13.48 ng/ml 57.10% 63.00% 60.49% 54.05% 65.91%

⁎ MK = midkine, Tg = thyroglobulin, PPV = positive predictive value, NPV = negative predictive value.

MK and 13.48 ng/ml for Tg), the sensitivity, specificity, diagnostic accuracy, PPV, and NPV of MK were all higher than those of Tg. In the second analysis, differentiation between thyroid cancer and normal control participants was carried out. Table 3 showed nearly the same diagnostic capability of MK and Tg, though the ROC curves displayed that the area under the curve of MK was higher than that of Tg (Fig. 2B). 3.2. Both MK and Tg had good prognostic values to predict metastasis-free survival, although the latter is better One month after surgery, the blood parameters of the participants in the surgical cohort were measured again. Serum thyroid hormones obviously decreased while TSH increased (Table 4). MK and Tg also decreased obviously. All the patients of the surgical cohort entered the 131 I treatment cohort. Altogether, we recruited 214 DTC patients in the 131 I treatment cohort, who were continuously followed up for at least 12 months to appraise the completeness of 131I ablation and whether metastases existed. Based on the follow-up information until the end of 12 months, we divided the 131I treatment cohort into two subgroups, namely, subgroup 1 patients who had successful ablation and no metastases (167 cases), and subgroup 2 patients who had 131I-avid metastases (47 cases, with lymph node, pulmonary or skeletal metastases). Comparisons of these two groups were performed. Age, MK, Tg, FT3 and FT4 in subgroup 2 were significantly higher than those in subgroup 1 (Table 5). Scatter graphs of MK and Tg were plotted to demonstrate the inter-subgroup differences as well (Fig. 3). Gender did not influence the prediction of whether or not 131I-avid metastases existed (Supplementary Table 3). Pearson and Spearman bivariate correlations were done as well (Supplementary Table 4). Significant positive correlations were also identified between MK and Tg (P b 0.01). Besides, Pearson coefficients showed a positive correlation between MK and age, FT3 or FT4, while Spearman coefficients showed a negative correlation between MK and TgAb. Prognostic capability of pre-ablative MK, in comparison with preablative Tg, was done by the ROC curve (Fig. 2C). Pre-ablative Tg demonstrated perfect ability for prediction of metastases with an area under the curve of 0.995 and diagnostic accuracy of 96.73 %, after an

Table 3 Distinction between thyroid cancer and normal control according to ROC curve-related data. MK⁎ Area under the curve Optimal cut-off value Sensitivity(%) Specificity(%) Accuracy(%) PPV⁎ (%) NPV⁎ (%)

0.853 323.48 pg/ml 75.70% 80.00% 77.93% 77.94% 77.92%

Tg⁎ 0.771 5.54 ng/ml 70.00% 85.30% 77.93% 81.67% 75.29%

⁎ MK = midkine, Tg = thyroglobulin, PPV = positive predictive value, NPV = negative predictive value.

22

Z. Meng et al. / Life Sciences 130 (2015) 18–24

Table 4 Changes of serum factors after surgery in thyroid cancer patients (70 cases). MK⁎ (pg/ml)

Tg⁎ (ng/ml)

Before surgery After surgery

838.90 ± 916.93 579.68 ± 776.53

48.39 ± 81.00 36.53 ± 80.00

T value⁎⁎ P value⁎⁎

1.805 0.073

0.872 0.385

TgAb⁎ (IU/mL) 29.98 ± 87.59 38.31 ± 111.34 −0.492 0.624

FT3⁎ (pmol/L)

FT4⁎ (pmol/L)

TSH⁎ (μIU/mL)

4.45 ± 0.54 1.97 ± 0.66

15.80 ± 1.73 5.08 ± 1.88

3.68 ± 1.65 89.09 ± 40.20

24.304 b0.001

35.063 b0.001

−17.759 b0.001

⁎ MK = midkine, Tg = thyroglobulin, TgAb = thyroglobulin antibody, FT3 = free triiodothyronine, FT4 = free thyroxine, TSH = thyroid stimulating hormone; ⁎⁎ Analyzed by independent samples T test.

optimal cut-off value of 19.50 ng/ml was set. Pre-ablative MK also performed very well with an area under the curve of 0.876 and diagnostic accuracy of 89.25 %, after an optimal cut-off value of 504.71 pg/ml was set (Table 6). Tg was superior to MK for prediction of metastases. We changed the above cut-off values to more practical levels of 500 pg/ml for MK and 20 ng/ml for Tg. Then, two groups of patients with different MK concentrations and different Tg concentrations were compared by the Kaplan–Meier method for the twelve-month 131 I-avid metastasis-free survival (Fig. 2D and E). DTC patients with a pre-ablative MK level of higher than 500 pg/ml showed a significantly worse 131I-avid metastasis-free survival (P b 0.01), only 29.09 % of them were found to have no metastases. On the contrary, DTC patients with a pre-ablative MK level of lower than 500 pg/ml showed a significantly better survival (P b 0.01), 94.97 % of them were found to have no metastases. Similarly, DTC patients with a pre-ablative Tg level of higher than 20 ng/ml showed a significantly worse 131I-avid metastasis-free survival than otherwise (P b 0.01), 11.54 % of the former and 99.38 % of the latter were found to be free of metastases. 4. Discussion MK is a pleiotropic growth factor primarily expressed during embryogenesis, but down-regulated to very low levels in healthy adults. After compiling available publications [7,9,15,17,26,27,30,38], Jones [10] reported that the mean level of MK was around 253 pg/ml in healthy human subjects. Consistently, our control participants had a mean MK level of 255.01 pg/ml. MK signaling is largely mediated by cell surface receptors and membrane proteins. Among the signaling systems downstream from the receptors, multiple kinases in mitogen-activated protein kinase (MAPK) and phosphoinositide 3-kinase (PI3K)/Akt cascades appear to play central roles [11,24,25]. Overexpression of MK is a typical feature of cancer, regardless of tumor tissue type. This phenomenon is reminiscent of the p53 gene mutation [12]. In contrast to p53, the blood MK level can be monitored. MK is secreted by the cells that produce it, and then reaches the circulation. There is not any reported instance in which a tumor type with elevated tissue MK does not show increased circulating MK. There are generally four important features common to the cancer studies of circulating MK [10,24,25]: 1) MK levels are significantly elevated in

cancer compared to healthy normal samples, 2) MK levels are significantly elevated in cancer when compared to non-malignant diseases of the same organ or tissues, 3) MK levels positively correlated with increasing severity and malignancy of cancer, and 4) when tumor lesions are surgically resected, circulating MK levels usually decrease afterwards, and would often increase again if recurrence or metastases happen. Therefore, MK has generally become a convenient, accessible, noninvasive and inexpensive biomarker for cancer. The first diagnostic tests that quantify MK are just now receiving regulatory clearance and entering the clinic [10]. Many studies investigated the diagnostic thresholds to discriminate malignancy from benign disease. Zhu et al. [38] evaluated MK as a diagnostic biomarker in hepatocellular carcinoma and compared MK with alpha-fetoprotein. The authors found that the sensitivity of serum MK (cut-off value of 654 pg/ml) for hepatocellular carcinoma diagnosis was much higher than alpha-fetoprotein (86.9 % versus 51.9 %) with similar specificities (83.9 % versus 86.3 %). ROC curve analyses also showed that serum MK had a better performance in distinguishing early-stage hepatocellular carcinomas as well as small hepatocellular carcinomas. Even in very early-stage hepatocellular carcinomas, MK showed an obviously higher sensitivity compared with AFP (80% vs. 40%). Lucas et al. [21] proposed a cut-off value of 500 pg/ml to differentiate pediatric patients with embryonal tumor from non-cancerous controls, which led to a sensitivity of 65.5 % and a specificity of 86.2 %. Krzystek-Korpacka et al. [18] found an optimal cut-off value of 563 pg/ml by ROC analysis to distinguish patients with esophageal squamous cell carcinoma from healthy controls, which showed a sensitivity of 85.0 % and specificity of 90.5 % with an area under the curve value of 0.831. Shimada et al. [34] defined a cut-off value of 300 pg/ml when studying another cohort of patients with esophageal squamous cell cancer. They reported a sensitivity of 61.0 % and a specificity of 96.3 %. In the current study, we set an optimal cut-off value of 323.12 pg/ml for MK to differentiate maligant and benign thyroid nodules, which produced a sensitivity, specificity and diagnostic accuracy of 75.70%, 75.00% and 75.31%, respectively. These data were much better than the results from Tg (Table 2). In the analysis to differentiate DTC from normal participants without any thyroid diseases, a cut-off value of 323.48 pg/ml for MK was determined. The sensitivity, specificity and diagnostic accuracy from MK were almost similar with the

Table 5 Data comparison in subgroups of the 131I treatment cohort Subgroup⁎ (case number)

Age (years old)

MK⁎⁎ (pg/ml)

Tg⁎⁎ (ng/ml)

TgAb⁎⁎ (IU/mL)

FT3⁎⁎ (pmol/L)

FT4⁎⁎ (pmol/L)

TSH⁎⁎ (μIU/mL)

1 (167) 2 (47) Total (214)

48.14 ± 12.77 52.38 ± 13.56 49.07 ± 13.04

277.05 ± 148.78 1501.98 ± 1063.49 546.08 ± 721.03

3.78 ± 6.00 145.36 ± 119.76 34.88 ± 81.10

27.81 ± 75.79 33.12 ± 136.04 28.98 ± 92.08

1.92 ± 0.60 2.29 ± 0.90 2.00 ± 0.69

5.14 ± 1.69 5.94 ± 2.44 5.32 ± 1.91

90.03 ± 40.81 86.15 ± 44.11 89.18 ± 41.49

T value(1):(2)# P value(1):(2)#

−1.983 0.049

−14.473 b0.001

−15.301 b0.001

−0.348 0.728

−3.302 0.001

−2.579 0.011

0.565 0.573

⁎ Subgroup 1 = successful ablation without metastases, subgroup 2 = 131I-avid metastases exist. ⁎⁎ MK = midkine, Tg = thyroglobulin, TgAb = thyroglobulin antibody, FT3 = free triiodothyronine, FT4 = free thyroxine, TSH = thyroid stimulating hormone. # Analyzed by independent samples T test.

Z. Meng et al. / Life Sciences 130 (2015) 18–24

23

Fig. 3. Scatter graphs were plotted to demonstrate the inter-subgroup differences of pre-131I-ablative serum midkine (A) and thyroglobulin (B).

parameters from Tg (Table 4). These results proved the efficacy of MK to be used as a promising pre-surgical biomarker for thyroid cancer screening. The prognostic value of MK was also evaluated in a lot of studies by the method of the Kaplan–Meier curve. Ota et al. [27] identified MK as a prognostic biomarker in oral squamous cell carcinoma. They found that patients with serum MK concentrations of ≥ 650 pg/ml had a 5-year survival rate of 56.6 %. However, if patients had serum MK concentrations of b 650 pg/ml, they would have a 5-year survival rate of 82.9 %. Rawnaq et al. [30] used 400 pg/ml as the threshold value for the study of gastrointestinal stromal tumors, and found a 5-year recurrence-free survival rate to be 50% if serum MK was lower than the threshold, yet the figure was only 25% if MK was higher than the threshold. Ikematsu et al. [8] defined low and high levels of serum MK by a threshold of 900 pg/ml when investigating neuroblastoma, and found significant higher cumulative survival rates in the low level group. In the study by Shimada et al. [34], the 5-year survival rate of esophageal squamous cell cancer in the high serum MK group was significantly worse than that in the low serum MK group, 46.2 % versus 81.2 %, respectively. In our analyses, we decided to use a more practical concentration of 500 pg/ml for MK and 20 ng/ml for Tg in order to evaluate their prognostic values for thyroid cancer, in specific, to predict whether metastases existed. The Kaplan–Meier curves showed significant better 131 I-avid metastasis-free survival in the lower level groups (Fig. 2D–E). These assessments evidently indicated MK to be a prognostic biomarker before the first 131I therapy. To the best of our knowledge, this is the first paper showing that presurgical serum MK can be used to differentiate malignant and benign

Table 6 Prediction of whether 131I-avid metastases exist according to ROC curve-related data MK⁎ Area under the curve Optimal cut-off value Sensitivity(%) Specificity(%) Accuracy(%) PPV⁎ (%) NPV⁎ (%)

0.876 504.71 pg/ml 83.00% 91.00% 89.25% 95.00% 72.22%

Tg⁎ 0.995 19.50 ng/ml 97.90% 96.40% 96.73% 99.38% 88.46%

⁎ MK = midkine, Tg = thyroglobulin, PPV = positive predictive value, NPV = negative predictive value.

thyroid nodule, and pre-131I-ablative serum MK to predict whether or not metastases exist. In fact, regarding thyroid cancer, we identified only three papers which used immunohistochemistry to study PTC. Kato et al. [14] first reported that PTC strongly expressed MK protein and mRNA. However, normal follicular epithelial cells in tissues adjacent to the cancer tissues expressed MK very faintly or not at all. Shao et al. [33] evaluated associations between MK expression and clinicopathological features of PTC. This investigation found that extrathyroidal invasion, lymph node metastasis, and tumor stage III/IV were associated with strong MK positivity and high expression scores. Our recent study showed that MK immunohistochemistry could be used for differential diagnosis between PTC and multi-nodular goiter, and for prediction of synchronous metastases [37]. These findings actually prompted us to do the current study. In an attempt to identify a serum biomarker for DTC, serum MK was determined in peripheral blood of DTC patients as well as control subjects in our investigation. We found significantly elevated pre-surgical MK levels in DTC patients in comparison to patients with benign thyroid nodules and health controls. This feature is of great importance, since MK is applicable to differentiate between malignant and benign thyroid nodules, in which the purpose of Tg is of inferior quality [4] and is not recommended by the guidelines [2,29]. We also found that pre-ablative MK concentration was able to predict metastases. Thus, blood MK can be as good as or even better than the conventional biomarker (Tg) for DTC. There are two major limitations in this study, which warrant further research. Firstly, we recruited a relatively small number of DTC patients and followed up for less than three years, so we did not perform prognostic analyses on MK's values to predict new metastases or recurrence. Besides, since DTC has relatively good survival rate, it does need a much longer follow-up time to study MK's value to predict long-term survival rate and to predict dedifferentiation. Therefore, three aspects of clinical studies need to be done, which include: 1) whether serum MK can be a good surveillance marker for DTC, 2) whether serum MK can predict overall survival rate and 3) whether serum MK can predict 131I therapeutic outcome and dedifferentiation. Secondly, it has been proved that MAPK and PI3K are two of the main pathways that regulate 131I uptake in DTC cells, and targeting one or several of these major signaling pathways can restore thyroid gene expression and increase 131I accumulation in thyroid cancer cells [6,16,20]. Very recently, the team from the Memorial Sloan-Kettering Cancer Center demonstrated that the MAPK inhibitor selumetinib could reverse refractoriness to 131I in

24

Z. Meng et al. / Life Sciences 130 (2015) 18–24

patients with metastatic DTC [5]. Cancer-related activities of MK have been clarified by numerous studies to be handled by MAPK and PI3K cascades via MK receptors. It will be very interesting to study whether MK inhibition (using modalities like anti-MK siRNA or anti-MK antibodies) can result in 131I uptake enhancement and synergism in DTC in vitro and in vivo. Currently, we are undertaking the above two arms of investigation. 5. Conclusions We discovered that MK can potentially be used to screen patients with thyroid nodules for DTC before surgery, and to predict whether metastases exist before 131I ablative therapy. The present results also indicated that MK could possibly be a candidate molecular target for therapy of DTC, which warrants further investigation. Conflict of interest statement The authors declare that they have no conflict of interest.

Acknowledgments This investigation was supported by the National Key Clinical Specialty Project (awarded to the Departments of Nuclear Medicine and Radiology). This study was also supported by the Tianjin Medical University General Hospital New Century Excellent Talent Program; Young and Middle-aged Innovative Talent Training Program from Tianjin Education Committee; and Talent Fostering Program (the 131 Project) from the Tianjin Education Committee, Tianjin Human Resources and Social Security Bureau (awarded to Zhaowei Meng). Appendix A Supplementary data Supplementary data to this article can be found online at http://dx. doi.org/10.1016/j.lfs.2015.02.028. References [1] M.O. Bernier, O. Morel, P. Rodien, J.P. Muratet, P. Giraud, V. Rohmer, et al., Prognostic value of an increase in the serum thyroglobulin level at the time of the first ablative radioiodine treatment in patients with differentiated thyroid cancer, Eur. J. Nucl. Med. Mol. Imaging 32 (2005) 1418–1421. [2] D.S. Cooper, G.M. Doherty, B.R. Haugen, R.T. Kloos, S.L. Lee, S.J. Mandel, et al., Revised American Thyroid Association management guidelines for patients with thyroid nodules and differentiated thyroid cancer, Thyroid 19 (2009) 1167–1214. [3] L. Giovanella, L. Ceriani, A. Ghelfo, M. Maffioli, F. Keller, Preoperative undetectable serum thyroglobulin in differentiated thyroid carcinoma: incidence, causes and management strategy, Clin. Endocrinol. 67 (2007) 547–551. [4] E. Guarino, B. Tarantini, T. Pilli, S. Checchi, L. Brilli, C. Ciuoli, et al., Presurgical serum thyroglobulin has no prognostic value in papillary thyroid cancer, Thyroid 15 (2005) 1041–1045. [5] A.L. Ho, R.K. Grewal, R. Leboeuf, E.J. Sherman, D.G. Pfister, D. Deandreis, et al., Selumetinib-enhanced radioiodine uptake in advanced thyroid cancer, N. Engl. J. Med. 368 (2013) 623–632. [6] P. Hou, E. Bojdani, M. Xing, Induction of thyroid gene expression and radioiodine uptake in thyroid cancer cells by targeting major signaling pathways, J. Clin. Endocrinol. Metab. 95 (2010) 820–828. [7] M. Ibusuki, H. Fujimori, Y. Yamamoto, K. Ota, M. Ueda, S. Shinriki, et al., Midkine in plasma as a novel breast cancer marker, Cancer Sci. 100 (2009) 1735–1739. [8] S. Ikematsu, A. Nakagawara, Y. Nakamura, M. Ohira, M. Shinjo, S. Kishida, et al., Plasma midkine level is a prognostic factor for human neuroblastoma, Cancer Sci. 99 (2008) 2070–2074. [9] S. Ikematsu, A. Yano, K. Aridome, M. Kikuchi, H. Kumai, H. Nagano, et al., Serum midkine levels are increased in patients with various types of carcinomas, Br. J. Cancer 83 (2000) 701–706. [10] D.R. Jones, Measuring midkine: the utility of midkine as a biomarker in cancer and other diseases, Br. J. Pharmacol. (2014). [11] K. Kadomatsu, S. Kishida, S. Tsubota, The heparin-binding growth factor midkine: the biological activities and candidate receptors, J. Biochem. 153 (2013) 511–521.

[12] K. Kadomatsu, T. Muramatsu, Midkine and pleiotrophin in neural development and cancer, Cancer Lett. 204 (2004) 127–143. [13] K. Kadomatsu, M. Tomomura, T. Muramatsu, cDNA cloning and sequencing of a new gene intensely expressed in early differentiation stages of embryonal carcinoma cells and in mid-gestation period of mouse embryogenesis, Biochem. Biophys. Res. Commun. 151 (1988) 1312–1318. [14] M. Kato, H. Maeta, S. Kato, T. Shinozawa, T. Terada, Immunohistochemical and in situ hybridization analyses of midkine expression in thyroid papillary carcinoma, Mod. Pathol. 13 (2000) 1060–1065. [15] O. Kemik, A. Sumer, A.S. Kemik, I. Hasirci, S. Purisa, A.C. Dulger, et al., The relationship among acute-phase response proteins, cytokines and hormones in cachectic patients with colon cancer, World J. Surg. Oncol. 8 (2010) 85. [16] T. Kogai, S. Sajid-Crockett, L.S. Newmarch, Y.Y. Liu, G.A. Brent, Phosphoinositide-3kinase inhibition induces sodium/iodide symporter expression in rat thyroid cells and human papillary thyroid cancer cells, J. Endocrinol. 199 (2008) 243–252. [17] M. Krzystek-Korpacka, D. Diakowska, K. Neubauer, A. Gamian, Circulating midkine in malignant and non-malignant colorectal diseases, Cytokine 64 (2013) 158–164. [18] M. Krzystek-Korpacka, M. Matusiewicz, D. Diakowska, K. Grabowski, K. Blachut, I. Kustrzeba-Wojcicka, et al., Serum midkine depends on lymph node involvement and correlates with circulating VEGF-C in oesophageal squamous cell carcinoma, Biomarkers 12 (2007) 403–413. [19] J.I. Lee, Y.J. Chung, B.Y. Cho, S. Chong, J.W. Seok, S.J. Park, Postoperative-stimulated serum thyroglobulin measured at the time of 131I ablation is useful for the prediction of disease status in patients with differentiated thyroid carcinoma, Surgery 153 (2013) 828–835. [20] D. Liu, S. Hu, P. Hou, D. Jiang, S. Condouris, M. Xing, Suppression of BRAF/MEK/MAP kinase pathway restores expression of iodide-metabolizing genes in thyroid cells expressing the V600E BRAF mutant, Clin. Cancer Res. 13 (2007) 1341–1349. [21] S. Lucas, T. Reindl, G. Henze, A. Kurtz, S. Sakuma, P.H. Driever, Increased midkine serum levels in pediatric embryonal tumor patients, J. Pediatr. Hematol. Oncol. 31 (2009) 713–717. [22] S. Maeda, H. Shinchi, H. Kurahara, Y. Mataki, H. Noma, K. Maemura, et al., Clinical significance of midkine expression in pancreatic head carcinoma, Br. J. Cancer 97 (2007) 405–411. [23] H. Muramatsu, T. Muramatsu, Purification of recombinant midkine and examination of its biological activities: functional comparison of new heparin binding factors, Biochem. Biophys. Res. Commun. 177 (1991) 652–658. [24] T. Muramatsu, Structure and function of midkine as the basis of its pharmacological effects, Br. J. Pharmacol. 171 (2014) 814–826. [25] T. Muramatsu, K. Kadomatsu, Midkine: an emerging target of drug development for treatment of multiple diseases, Br. J. Pharmacol. 171 (2014) 811–813. [26] Y. Obata, S. Kikuchi, Y. Lin, K. Yagyu, T. Muramatsu, H. Kumai, Serum midkine concentrations and gastric cancer, Cancer Sci. 96 (2005) 54–56. [27] K. Ota, H. Fujimori, M. Ueda, S. Shiniriki, M. Kudo, H. Jono, et al., Midkine as a prognostic biomarker in oral squamous cell carcinoma, Br. J. Cancer 99 (2008) 655–662. [28] F. Pacini, M.G. Castagna, L. Brilli, G. Pentheroudakis, Thyroid cancer: ESMO Clinical Practice Guidelines for diagnosis, treatment and follow-up, Ann. Oncol. 21 (Suppl 5) (2010) v214–v219. [29] F. Pacini, M. Schlumberger, H. Dralle, R. Elisei, J.W. Smit, W. Wiersinga, European consensus for the management of patients with differentiated thyroid carcinoma of the follicular epithelium, Eur. J. Endocrinol. 154 (2006) 787–803. [30] T. Rawnaq, M. Kunkel, K. Bachmann, R. Simon, H. Zander, S. Brandl, et al., Serum midkine correlates with tumor progression and imatinib response in gastrointestinal stromal tumors, Ann. Surg. Oncol. 18 (2011) 559–565. [31] E. Robenshtok, R.K. Grewal, S. Fish, M. Sabra, R.M. Tuttle, A low postoperative nonstimulated serum thyroglobulin level does not exclude the presence of radioactive iodine avid metastatic foci in intermediate-risk differentiated thyroid cancer patients, Thyroid 23 (2013) 436–442. [32] K. Sakamoto, K. Kadomatsu, Midkine in the pathology of cancer, neural disease, and inflammation, Pathol. Int. 62 (2012) 445–455. [33] H. Shao, X. Yu, C. Wang, Q. Wang, H. Guan, Midkine expression is associated with clinicopathological features and BRAF mutation in papillary thyroid cancer, Endocrine (2013). [34] H. Shimada, Y. Nabeya, M. Tagawa, S. Okazumi, H. Matsubara, K. Kadomatsu, et al., Preoperative serum midkine concentration is a prognostic marker for esophageal squamous cell carcinoma, Cancer Sci. 94 (2003) 628–632. [35] R. Siegel, J. Ma, Z. Zou, A. Jemal, Cancer statistics, 2014, CA Cancer J. Clin. 64 (2014) 9–29. [36] M. Toubeau, C. Touzery, P. Arveux, G. Chaplain, G. Vaillant, A. Berriolo, et al., Predictive value for disease progression of serum thyroglobulin levels measured in the postoperative period and after (131)I ablation therapy in patients with differentiated thyroid cancer, J. Nucl. Med. 45 (2004) 988–994. [37] Y. Zhang, Z. Meng, M. Zhang, J. Tan, W. Tian, X. He, et al., Immunohistochemical evaluation of midkine and nuclear factor-kappa B as diagnostic biomarkers for papillary thyroid cancer and synchronous metastasis, Life Sci. (2014). [38] W.W. Zhu, J.J. Guo, L. Guo, H.L. Jia, M. Zhu, J.B. Zhang, et al., Evaluation of midkine as a diagnostic serum biomarker in hepatocellular carcinoma, Clin. Cancer Res. 19 (2013) 3944–3954.