Serum thrombomodulin as a metastatic and prognostic marker in soft tissue sarcomas

Abstract

BACKGROUND:

Thrombomodulin (TM) has multiple biological functions and modulates not only anti-coagulation, but also cell proliferation, adhesion, and anti-inflammation activities. The main function of TM is to activate the anticoagulant pathway of protein C. Soluble TM is related to metastasis by its inactivation of thrombin.

OBJECTIVES:

To clarify the correlation between serum TM levels and clinicopathological parameters.

METHODS:

The plasma TM levels (FU/ml) of 135 primary soft tissue tumors (benign, 67; soft tissue sarcoma (STS), 68) were measured before biopsy or treatment. TM levels were analyzed and compared to various clinicopathological parameters. Log-rank test and Cox proportional analysis were used to evaluate recurrence-free survival, metastasis-free survival, and overall survival.

RESULTS:

STS tumors had significantly higher TM values (15.9) than benign tumors (13.7) ($p=$ 0.0138). 5-year MFS was 81.1% in low TM and 40.0% in high TM ($p=$ 0.00671), and 5-year OS was 85.5% in low and 52.5% in high TM in grades 1–3 ($p=$ 0.0673). In multivariate COX proportional analysis, high-TM showed a significant difference (MFS: HR 4.37, $p=$ 0.0147; OS: HR 3.60, $p=$ 0.0557) in grades 1–3.

CONCLUSIONS:

We demonstrated that a high level of soluble TM has the potential to be a significant predictor of metastasis and poor prognosis in STS patients. TM is a candidate molecular marker for high metastatic potential and can be clinically useful for guiding therapeutic strategy.

1. Introduction

Thrombomodulin (CD141) (TM), a transmembrane glycoprotein with 557 amino acids, comprises an NH${}_2$-terminal lectin-like region followed by epidermal growth factor (EGF)-like structures, an O-glycosy- lation-site rich domain, a transmembrane domain, and a cytoplasmic tail. TM has multiple biological functions and modulates not only anti-coagulation but also cell proliferation, adhesion, and anti-inflammation [1, 2, 3].

The main function of TM is as a receptor for thrombin, and the complex with thrombin activates the anticoagulant pathway of protein C. Activated protein C proteolytically inactivates coagulation factors V and VII. The expression of TM leads to an anticoagulatory state in the microenvironment around cells. The anti-inflammatory effect is due to the N-terminal lectin-like domain of TM that binds to and neutralizes the inflammatory mediator HMGB1 (high-mobility group-B1 DNA-binding protein). Horowitz et al. reported that mice expressing a mutant form of TM with reduced thrombin affinity exhibit a remarkable metastatic phenotype, and mice expressing a mutant form of TM lacking the lectin-like domain have no metastatic potential. The critical site that promotes tumor metastasis is therefore the thrombin binding domain and not the lectin-like domain [4]. As to whether the circulating soluble form of TM has anti-metastatic functions, Hosaka et al. isolated soluble TM from human urine, which indicated complete activation of protein C. Intravenous injection of the extracted TM from human urine reduces lung metastasis in mice [5]. This means that soluble TM from human urine has the ability to influence metastatic potential in vivo.

Given this, it is thought that circulating soluble TM possibly possesses anti-metastatic effects and leads to improved metastasis-free survival and overall survival. We hypothesized that an increase in soluble TM is related to malignancy or improvement of recurrence, metastasis, or overall survival in STS patients. The purpose of the present study was to look for correlations between serum TM levels and clinicopathological parameters in 135 soft tissue tumor patients and to predict recurrence, metastasis, or prognosis in STS patients.

2. Materials and methods

2.1. Patients

A total of 135 patients (76 men and 59 women) with primary soft tissue tumors who visited Mie University Hospital from 2012–2014 were enrolled in this study. Patients who had local recurrence or who were referred for additional resection after inadequate resection in a previous hospital were excluded from this study. Histopathological diagnosis and the histological grade were verified based on the French Federation of Cancer Centers Sarcoma group system (FNCLCC) by independent pathologists. Blood samples from all patients were obtained before biopsy or treatment. The serum TM level was quantitatively measured by enzyme immunoassay. Written, informed consent was obtained from each patient. For patients below the age 19 years, informed consent was obtained from their parents or legal guardian. This study was approved by the Ethics Committee of the Mie University Graduate School of Medicine (approval number: 1435). All procedures performed in studies involving human participants were in accordance with the ethical standards of the Ethics Committee of Mie University and with the Helsinki declaration of 1975.

2.2. Statistical analysis

Clinicopathological analysis was performed to compare the serum TM levels to various clinical parameters using the Mann-Whitney $U$-test or Kruskal Wallis test for quantitative data. To evaluate the threshold for detecting a recurrence, metastasis, or mortality due to disease, receiver operating characteristic (ROC) curves were generated. ROC curves were created by plotting the sensitivity on the y-axis and the false positive rate (1-specificity) on the x-axis. To measure the effectiveness of the TM level as a marker for recurrence, metastasis, or mortality due to disease, the area under the curve (AUC) was assessed. Local recurrence-free survival (RFS) was defined as the time from the initial treatment to the date of clinically documented local recurrence. Metastasis-free survival (MFS) was defined as the time from the initial treatment to the date of clinically documented distant metastasis. Overall survival (OS) was defined as the time from the initial treatment to the date of mortality attributed to the neoplasm. Kaplan-Meier survival plots and log-rank tests were used to assess the differences in time to local recurrence, distant metastasis, or overall survival. To adjust the imbalance in prognostic factors among patients, Cox proportional hazard analysis was used. 0.05 was considered statistically significant.

3. Results

3.1. Patient and tumor characteristics

The median age of the patients was 56.0 years (range: 12–92 years), and the median tumor size was 10.3 cm (range: 1–31 cm). The primary tumor sites were in the extremity for 68 patients and in the trunk for 67 patients. The tumor depth was superficial in 51 patients and deep in 84 patients. Histopathological diagnosis was as follows: 26 lipomas, 13 schwannomas, 7 fibromatosis, 5 hematomas, 3 neurofibromas, 3 pigmented villonodular synovitis, and 10 others in 67 benign tumors; and 29 liposarcomas (17 well-differentiated liposarcomas, 6 dedifferentiated liposarcomas, and 6 myxoid liposarcomas), 13 undifferentiated pleomorphic sarcomas, 9 myxofibrosarcomas, 4 leiomyosarcomas, 3 synovial sarcomas, 3 malignant peripheral nerve sheath tumors, and 7 others in STS. All patients with benign tumors received tumor resection and all patients with STS received curative treatment (wide resection: 41 patients, marginal resection: 17 patients, amputation: 3 patients, ion beam radiotherapy: 3 patients, no curative treatment: 4 patients). The median follow-up in STS patients was 35.8 months (range: 0.6–62.8 months).

In benign soft tissue tumors and STS, median TM levels between patients based on sex, tumor size, location, and depth were not significantly different. Patients over 60 years of age showed significantly higher median TM values than those below 60. STS patients had significantly higher TM values than those with benign tumors (Table 1). From the ROC analysis, the diagnostic accuracy for identifying STS was evaluated by the area under the curve. Serum TM levels were not useful for distinguishing STS from benign tumors according to area under the curve analysis (AUC: 0.611, 95% confidential interval (CI): 0.516–0.706) (data not shown). In STS, nineteen tumors were of histological grade 1, 22 were of grade 2, and 23 were of grade 3. According to the 7${}^{\text{th}}$ edition of the American Joint Committee on Cancer (AJCC) classification of soft tissue sarcomas, 19 patients were classified as stage I, 22 patients were classified as stage II, 21 were stage III, and 5 were stage IV. Median TM levels of between patients based on sex, tumor size, location, depth, FNCLCC grade, and AJCC stage were not significantly different (Table 2).

Table 1.

Patient characteristics in benign soft tissue tumors and STS

Characteristics	$n$	TM (median)	$p$-value
Sex
Male	76	15.2	0.	234
Female	59	14.3
Age
60	75	12.8	0.	000001
60 $⩽$	60	17.2
Tumor size
10	78	14.4	0.	453
10 $⩽$	57	15.4
Location
Extremity	68	14.5	0.	381
Trunk	67	15.1
Tumor depth
Superficial	51	15.6	0.	132
Deep	84	14.3
Malignancy
Benign	67	13.7	0.	0138
STS	68	15.9

Sex, age, tumor size, location, tumor depth, FNCLCC grade, and malignancy are listed and noted by number. TM values were compared for each parameter by the Mann-Whitney test. Age and malignancy showed significant differences.

Table 2.

Patient characteristics in STS

Characteristics	$n$	TM (meidan)	$p$-value
Sex
Male	36	15.2	${}^{*}$0.	672
Female	32	14.3
Age
60	28	13.1	${}^{*}$ 0.	0001
60 $⩽$	40	17.8
Tumor size
10	30	14.4	${}^{*}$0.	433
10 $⩽$	38	15.4
Location
Extremity	37	14.5	${}^{*}$0.	506
Trunk	31	15.1
Tumor depth
Superficial	21	15.6	${}^{*}$0.	465
Deep	47	14.3
FNCLCC
Grade 1	19	16.6	${}^{\mathrm{\#}}$0.	429
Grade 2	23	14.8
Grade 3	26	16.3
AJCC stage
I	19	16.6	${}^{\mathrm{\#}}$0.	441
II	23	14.8
III	21	15.7
IV	5	18.5
Treatment
Wide resection	41	15.6	${}^{\mathrm{\#}}$0.	0529
Marginal resection	17	16.9
Amptation	3	12.5
Ion beam radiotherapy	3	10.6
No curative treatment	4	20.3
Chemotherapy
$-$	52	17.1	0.	000152
$+$	16	12.0
Radiotherapy
$-$	54	15.7	0.	802
$+$	14	16.6

TM values by age showed significant difference by ${}^{*}$Mann-Whitney test in STS.

3.2. Characteristics of soft tissue sarcoma

During this study, 7 patients suffered recurrence, 22 patients suffered metastasis, and 20 patients died of disease (DOD). The TM levels of patients with metastasis and DOD were significantly higher (Table 3). From the ROC analysis, the diagnostic accuracy for identifying metastasis and DOD was evaluated by the area under the curve in grades 1 to 3 (metastasis: AUC 0.661, 95% CI 0.511–0.811; DOD: AUC 0.769, 95% CI 0.627–0.912) (Fig. 1A and B). Especially for high grade sarcoma (grades 2 and 3), the ROC analysis showed higher AUC for identifying metastasis (AUC: 0.738, 95% CI: 0.595–0.882) and DOD (AUC: 0.828, 95% CI: 0.702–0.828) than in grades 1 to 3 (data not shown). The sensitivity and specificity by threshold of serum TM of 18.5 FU/ml for identifying metastasis and DOD were delimited with 54.4%, 82.2% (metastasis in G1-3, Fig. 1A); 70.6%, 84.3% (DOD in G1-3, Fig. 1B); 54.5%, 88.0% (metastasis in G2 and 3, data not shown); and 70.6%, 90.3% (DOD G2 and 3, data not shown).

Table 3.

Over the study period, patients with recurrence were 7, with metastasis 22, and DOD 20. The TM levels of patients with metastasis and DOD indicated significantly higher TM values by Mann-Whitney test

Characteristics	$n$	TM (median)	$p$-value
Recurrence
$-$	60	15.9	0.647
$+$	7	15.6
Metastasis
$-$	45	14.6	0.0337
$+$	22	18.4
Dead of disease
$-$	47	13.8	0.000965
$+$	20	21.5

Figure 1.

ROC analysis. Results of AUC analysis for metastasis (AUC 0.661, 95% CI: 0.511–0.811) (A), and DOD (AUC 0.769, 95% CI: 0.627–0.912) (B). The sensitivity and specificity by a threshold of serum TM of 18.5 FU/ml were delimited as 52.4% and 83.7% (metastasis, A) and 58.8% and 83.0% (DOD, B).

From the ROC analysis, a threshold of 18.5 FU/ml was adopted to divide low ($⩽$ 18.5) and high (18.5 ) TM groups. To elucidate the association of multiple factors for identifying recurrence, metastasis, or DOD, multiple logistic regression analyses were performed. No factors had significant differences for diagnosing recurrence. TM higher than 18.5 FU/ml was associated with a significantly increased risk of metastasis and DOD (metastasis: OR 10.8, 95% CI: 2.22–52.6, $p=$ 0.0032, DOD: OR 10.5, 95% CI: 2.12–52.1, $p=$ 0.0214). Age was not related to the risk of metastasis or DOD (Table 4).

Table 4.

Multiple logistic analysis to identify recurrence, metastasis, or DOD is shown. Size and high TM value in metastasis and high TM in DOD showed significant difference

Characteristics	Recurrence			Metastasis			Dead of disease
	OR	95% CI	$p$-value	OR	95% CI	$p$-value	OR	95% CI	$p$-value
Male	2.13 $\times$ 10${}^8$	/	0.995	0.81	0.23–2.82	0.739	1.30	0.27–5.12	0.706
Age	0.98	0.91–1.05	0.511	0.99	0.96–1.04	0.823	1.05	0.98–1.10	0.0566
Size	0.96	0.83–1.11	0.552	0.89	0.81–0.99	0.0318	0.96	0.88–1.08	0.365
Superficial	0.50	0.04–5.70	0.576	0.60	0.15–2.35	0.463	0.98	0.23–4.97	0.982
Trunk	0.84	0.15–4.87	0.848	1.17	0.35–3.89	0.803	2.35	0.57–10.3	0.202
TM $>$ 18.5	1.98	0.18–22.5	0.580	10.8	2.22–52.6	0.00319	5.78	1.30–25.7	0.0214

3.3. Prognostic analysis

For prognostic analysis, 4 patients who received no curative treatment and 5 patients with distant metastasis at the first visit were excluded. RFS, MFS, and OS between the low- and high-TM groups were compared by Kaplan-Meier analysis and analyzed by log-rank tests. Five-year RFS showed no significant difference in grades 1 to 3 (low TM 90.0%, high TM 85.1%, $p=$ 0.538) (Fig. 2A). In high grade tumors (grades 2 and 3), RFS showed no difference as well (5 years: low TM 90.0%, high TM 85.1%, $p=$ 0.538, data not shown). The high-TM group had a significantly lower MFS rate in grades 1 to 3 (5 years: low TM 81.1%, high TM 40.0%, $p=$ 0.00671) (Fig. 2B). In high grade tumors, MFS showed a significant difference (5 years: low TM 72.2%, high TM 15.0%, $p=$ 0.00107, data not shown). Additionally, the high-TM group had a significantly poorer OS rate in grades 1 to 3 (5 years: low TM 85.5%, high TM 52.5%, $p=$ 0.00673) (Fig. 2C). In high grade tumors, OS showed a significant difference as well as MFS (5 years: low TM 78.6%, high TM 30.0%, $p=$ 0.00227, data not shown).

Figure 2.

Kaplan-Meier curves in STS. The high-TM group had a significantly lower MFS rate (5 years: low TM $=$ 81.1%, high TM $=$ 40.0%, $p=$ 0.00671) (B). The high-TM group had a significantly poorer OS rate (5 years: low TM $=$ 85.0%, high TM $=$ 52.5%, $p=$ 0.00673) (C).

Furthermore, as older patients had a high median value of TM, to adjust the imbalance in prognostic factors among patients, Cox proportional hazard analysis was used. In RFS, there were no significant factors in univariate and multivariate analysis in grades 1 to 3. In MFS, only the high-TM group showed a significant difference both univariate and multivariate analyses (univariate: HR: 3.62, 95% CI: 1.34–9.54, $p=$ 0.0111; multivariate: HR: 4.73, 95% CI: 1.34–14.3, $p=$ 0.0147). In OS, high TM (HR: 4.03, 95% CI: 1.35–12.0, $p=$ 0.0123) and age (HR: 1.06, 95% CI: 1.01–1.11, $P=$ 0.0172) showed significant differences in univariate analysis. However, in multivariate analysis, high TM (HR 3.60, 95% CI: 0.97–13.4, $p=$ 0.0557) showed only a tendency for poor prognosis (Table 5). These results indicate that high serum TM levels have high metastatic potential and lead to poor prognosis of overall survival.

Table 5.

Univariate and multivariate COX proportional hazard analysis

	Univariate analysis			Multivariate analysis
	HR	95% CI	$p$-value	HR	95% CI	$p$-value
RFS
Male	5.97 $\times$ 10${}^8$	/	0.998	6.93 $\times$ 10${}^8$	/	0.998
Age	1.007	0.95–1.06	0.809	0.99	0.92–1.06	0.713
Size	0.98	0.87–1.09	0.673	0.92	0.79–1.08	0.317
Deep	2.29	0.27–19.6	0.449	2.56	0.18–1.26	0.487
Trunk	2.85	0.52–15.6	0.228	1.48	0.24–9.12	0.670
TM $>$ 18.5	1.69	0.31–9.26	0.543	5.95	0.55–64.3	0.142
MFS
Male	0.58	0.21–1.59	0.289	0.64	0.22–1.87	0.416
Age	1.03	0.99–1.07	0.189	1.01	0.97–1.05	0.615
Size	0.97	0.90–1.04	0.399	0.94	0.88–1.01	0.0912
Deep	0.98	0.33–2.85	0.976	1.62	0.52–5.09	0.409
Trunk	1.10	0.41–2.97	0.848	1.57	0.52–4.76	0.422
TM $>$ 18.5	3.62	1.34–9.54	0.0111	4.37	1.34–14.3	0.0147
OS
Male	0.85	0.29–2.54	0.777	0.76	0.22–2.59	0.659
Age	1.06	1.01–1.11	0.0172	1.05	0.99–1.10	0.106
Size	0.997	0.93–1.07	0.943	0.97	0.90–1.05	0.448
Deep	0.997	0.31–3.24	0.996	1.85	0.47–7.31	0.383
Trunk	1.64	0.55–4.88	0.375	2.01	0.58–6.95	0.269
TM $>$ 18.5	4.03	1.35–12.0	0.0123	3.60	0.97–13.4	0.0557

In univariate analysis, high TM in MFS and age and high TM in OS showed significant difference. In multivariate analysis, only high TM showed significant difference in MFS.

4. Discussion

Our hypothesis was that the increase of circulating soluble TM prevents metastasis and leads to improved MFS and OS, because soluble TM has been shown to function in reducing metastasis in previous in vitro and in vivo studies [4, 5]. However, elevated soluble TM resulted in the opposite outcomes. TM values were significantly higher in STS than in benign tumors. In STS, patients with metastasis or DOD patients had significantly higher soluble TM values than patients without metastasis or surviving patients. This indicated that soluble TM could not have a sufficient anti-metastatic effect and prolong the lives of STS patients. It has been reported that soluble TM has only 30–50% of the activity of cellular TM [6]. The decreased activity may be insufficient to decrease of metastasis. We speculate that elevated TM probably results from one of the outcomes of tumor aggravation.

Soluble TM is not only a functional protein; the elevation of circulating soluble TM is widely accepted as a marker of endothelial cell damage [7]. The circulating soluble form is generated by proteolytic cleavage of trans-membrane TM, which is cleaved and released by matrix metalloproteinases (MMPs) or neutrophil-derived enzymes upon stimulation by TNF-alpha, IL-1beta, and IFN-gamma [8, 9]. Various disorders characterized by endothelial damage, like atrial fibrillation, organ failure, sepsis, disseminated intravascular coagulation (DIC), vasculitis, and venous thrombosis are associated with an increase of circulating soluble TM [10, 11, 12, 13]. Additionally, increases in soluble TM levels have been reported in prostate, pancreatic, colorectal, and other cancers with the progression of cancer stage [14, 15]. Additionally, inflammation, which is the cause of TM cleavage, is related to tumor aggravation. Elevated CRP, an inflammatory marker, is associated with poor prognosis in many malignant tumors [16, 17, 18, 19, 20, 21, 22, 23, 24]. In STS, vascular invasion is a significant prognostic factor for metastasis by histological analysis [25]. Increased inflammation, as caused by elevated IL-6 or CRP, is associated with poor prognosis in STS patients [26, 27]. As vascularity and inflammation are important factors for the progression of STS, the microenvironment at an inflammatory site possibly induces endothelial cell damage and release of TM. However, in this study, TM was not significantly correlated with CRP in high grade sarcomas by spearman’s rank test (rho: 0.0915, $p\mathrm{=}$ 0.555, data not shown). CRP could not predict TM values. Other inflammation system may be involved in increase of soluble TM. This needs further study.

Furthermore, from ROC analysis, a TM level higher than 18.5 FU/ml was a useful threshold to identify patients with metastasis or DOD. This threshold was delimited with 54.4% sensitivity and 82.2% specificity (AUC 0.661), or 70.6% sensitivity and 84.3% specificity (AUC 0.769), respectively. The threshold of 18.5 FU/ml was also useful for predicting 5-year MFS and OS. Five-year MFS was 81.1% in low TM and 40.0% in high TM, and 5-year OS was 85.5% in low TM and 52.5% in high TM with significant difference. This data included 19 FNCLCC grade I tumors, like well-differentiated liposarcomas, which basically had no potential for metastasis. Excluding grade 1 STS, a significant difference was observed for 5-year MFS (low TM 72.2%, high TM 15.0%) and 5-year OS (low TM 78.6%, high TM 30.0%). Given this, the measurement of soluble TM was helpful for diagnosis and predicting prognosis of metastasis or DOD.

This retrospective study includes some limitations. The number of patients is small and we could not statistically analyze them by each subtype, because soft tissue tumors including sarcomas have many subtypes and the incidence rate of each is low. Many studies have had to analyze STS as a whole rather by each histological classification. Soluble TM is elevated by atrial fibrillation, organ failure, sepsis, disseminated intravascular coagulation (DIC), vasculitis, venous thrombosis, etc. The background in this study was not included in the statistical analysis. These background conditions generally increase with age, and TM is thought to be increased by age as well. Our multivariate analysis excluded a relationship between poor MFS and OS in the high TM group and in older patients. However, we believe that measuring soluble TM may be useful for identifying metastatic and poor prognostic potential in STS.

5. Conclusion

High levels of serum soluble TM ($⩾$ 18.5) have the potential to be a significant predictor of metastasis and poor prognosis in STS patients. TM is a candidate molecular marker for high metastatic potential and can be clinically useful for devising a therapeutic strategy.

Conflict of interest

The authors declare that they have no conflict of interest.