Thyroid cancer, similarly to breast cancer, has a good prognosis, with a 10-year survival rate of 80-85% (1,2). However, 10-15% of cases present with repeated recurrence, metastasis, and further genetic mutations, resulting in difficult-to-manage cases thereby possibly leading to death. If prognosis can be inferred for each case, it can influence the subsequent treatment plan, e.g., addition of adjuvant therapy, such as radiotherapy after surgery. Therefore, pathological factors, clinical stage, and blood thyroglobulin levels have been studied as prognostic factors for thyroid cancer (3). In this study, we created an ELISA kit using a monoclonal antibody (MoAb), JT-95, for thyroid cancer made from the cell membrane protein of papillary thyroid cancer, as an antigen. Using this kit, the blood expression levels of sialyl fibronectin (S-FN), an antigen of the MoAb JT-95, were measured across various tumors, since previous studies have shown that S-FN is present in several cancer types (4). In particular, we examined the relationship between blood S-FN expression levels and pathological factors, blood thyroglobulin levels, and prognosis in thyroid cancer. Additionally, we considered the possibility that S-FN can serve as a prognostic factor of thyroid cancer.

Examined cases and blood sample collection. In total 182 cases were investigated: 63 cases of thyroid cancer, 33 cases of thyroid benign tumors, 7 cases of parathyroid benign tumors, and 79 cases of breast cancer. Pathological diagnosis showed 61 cases of papillary thyroid cancer, 1 case of follicular carcinoma, and 1 case of undifferentiated cancer in 63 thyroid cancers. Of the 33 cases of benign thyroid tumors, 28 cases of adenomatous goiter, 5 cases of follicular adenoma, and 7 cases of parathyroid tumors, were all adenomas. In 79 cases of breast cancer, pathological diagnosis showed 52 cases of invasive ductal cancer, 9 cases of invasive lobular cancer, and 18 cases of ductal cancer in situ (DCIS).

Peripheral venous blood samples (7.5 ml) from 182 cases were collected on EDTA tubes (Thermo Scientific, Tokyo, Japan) at the Hospital of the Jikei University from 2011 to 2019. Blood samples were separated by centrifugation 1900×g for 10 min at 4˚C. The supernatant was divided to new tubes and stored at 4˚C until ELISA assay.

Sandwich ELISA using the JT-95 MoAb. Briefly, each 96-well microplate (Thermo Scientific, Waltham, MA, USA) was coated with the JT-95 MoAb (4) using 100 μl of 14.4 μg/ml in Dulbecco’s phosphate buffered saline (DPBS) and stored overnight at 4˚C. Next, 100 μl of patient plasma, diluted 10 times with DPBS, was added to each well and incubated at 18˚C for 2 h. After washing, 100 μl of MoAb JT-95 conjugated with biotin at a concentration of 20 μg/ml was added and incubated at 18˚C for 1 h. Streptavidin-HRT (Elite ABC standard kit, PK6100, Vector Laboratories, Newark, CA, USA) and the OptEIA TMB Substrate Reagent Set (Fisher Scientific, Waltham, MA, USA) were used for the reactions. The absorbance was measured at 450 nm using a plate reader (Model 680 XR; Bio Rad, Hercules, CA, USA), and the quantity of S-FN expression, which is a type of glycosylated protein and an antigenic substance of JT-95 in the blood, was measured according to the manufacturer’s protocol (Model 680 XR reader) and compared to standardized S-FN values. Patient plasma levels were measured three times in each case, and those with an average value of <0.0001 were evaluated as ELISA-negative, whereas those with a value >0.0001 were evaluated as ELISA-positive.

Tissue preparation for immunohistochemical staining. Stage IVa,b twenty-four formalin-fixed, paraffin-embedded thyroid cancer tissue sections (3-μm slices) were prepared. These 24 stage IVa,b tumors represent advanced cancers that had multiple lymph node metastases or extrathyroidal infiltration of tumor at the time of surgery.

The anti-MoAb JT-95 antibody was used as the primary antibody for these tissue sections at a concentration of 0.005 μg/ml, and the i-view DAB Universal Kit (Ventana Medical Systems, Tucson, AZ, USA) was used as the secondary antibody. The staining process was completed using a BenchMark-Ultra autostainer (Ventana Medical Systems).

Tissue preparation for electron microscopy. For electron microscopic observation, the immunostained paraffin sections were fixed with 1% osmium tetroxide in 0.1 M phosphate buffer (pH 7.3) at 4˚C for 30 min. Dehydration was achieved using a graded series of ethanol, and sections were embedded in Epok 812 reagent (Oken, Tokyo, Japan). Ultrathin sections were prepared with a diamond knife and observed using a JEM-1400Plus (JEOL, Tokyo, Japan) electron microscope at 80 kV.

Statistical analysis. All statistical analyses were performed using the R software version 4.0.5 (The R Foundation for Statistical Computing, Vienna, Austria). The correlation between blood S-FN levels and various clinicopathological parameters and prognostic factors was calculated using the Fisher’s exact test for categorical variables and the two-tailed Student’s or Welch’s t-test for continuous or non-continuous variables, respectively. Statistical significance was set at p<0.05. Spearman’s rank correlation was used for statistical correlation analysis between the blood S-FN and thyroglobulin levels.

Ethics statement. According to the Declaration of Helsinki, written informed consent was obtained from each patient, and the study was approved by the Ethics Committee of Jikei Medical University, Tokyo, Japan, in 2011 [No. 27-112 (7997)].

Expression levels of blood S-FN in various tumors. In this study, an ELISA kit was prepared using the monoclonal antibody JT-95 for thyroid cancer, and the expression levels of blood S-FN were examined in various tumors. Of the 182 cases, there were 63 cases of thyroid cancer, 33 of benign thyroid tumors, 7 of benign parathyroid tumors, and 79 of breast cancer. Among the 63 cases of thyroid cancer, 24 (38%) were found to be S-FN-positive and 39 (61.9%) were S-FN-negative. Of the 40 benign neck tumors, 16 (40%) were S-FN-positive and 24 (60.0%) were S-FN-negative. Of the 79 breast cancer cases, 20 (25.3%) were S-FN-positive and 59 (74.6%) were S-FN-negative.

Comparison of S-FN expression in various tumors. S-FN expression was examined in S-FN-positive cases in each tumor group, and a significant difference was observed between thyroid cancer and breast cancer cases (p=0.007, Welch’s t-test). However, there was no significant difference between the thyroid cancer and benign neck tumor cases (Table I).

Association between blood S-FN levels, pathological factors, and clinical stages in thyroid cancer. In the 24 blood S-FN-positive patients and 39 blood S-FN-negative thyroid cancer patients, we analyzed the association between a) T-factor (tumor size), b) N-factor (lymph node metastasis), c) Ex-factor (extra-membrane infiltration), d) Ly-factor (lymphatic vessel infiltration), and e) v-factor (vascular infiltration) (pathological factors), and f) clinical stage (clinical factor).

Regarding the association between blood S-FN expression levels, pathological factors, and clinical stage, only the presence of vascular infiltration (v-factor) was related to blood S-FN levels in thyroid cancer (p=0.03) (Table II).

Relationship between blood S-FN levels and prognosis (locoregional recurrence and remote metastasis) in thyroid cancer.

Cases of locoregional recurrence and remote metastasis. The observation period for the 63 thyroid cancer cases was 6-25 years (median=9.8 years) after surgery. Fifteen patients (15/63, 23.8%) experienced locoregional recurrence, in which cancer cells had metastasized to lymph nodes, or distant organs, such as the lung and bone. Among the 24 blood S-FN-positive cases, recurrent metastasis was present in four: two cases of lung metastasis and two of cervical lymph node recurrences (4/24, 16.7%). These four patients developed recurrent metastasis from clinical stage IVa,b status. They had central and lateral cervical lymph node metastases or extrathyroidal infiltration of tumor status at the time of surgery and nine patients were in the same clinical stage IVa,b, and blood S-FN positive cases (4/9, 44%). Of these four cases, one patient died during the observation period due to cerebral infarction and heart failure. Among the 39 blood S-FN-negative cases, recurrence or metastasis was observed in 11 cases, including 3 cases of lung metastasis, 1 of bone metastasis, 2 of longitudinal lymph node recurrence, and 5 of cervical lymph node recurrence (11/39, 28.2%). All 11 recurrent and metastatic patients also evolved from 16 cases of stages IVa and IVb, and blood S-FN-negative cases (11/16, 69%). Of the 11 cases, three patients died during the observation period due to progression of thyroid cancer metastasis.

Relationship between blood S-FN levels and prognosis. In all 63 thyroid cancer cases, there was a significant difference between the 15 patients with recurrence and metastasis in terms of progression of T, N, Ex, and clinical stage (Table III).

Relationship between 24 blood S-FN-positive thyroid cancer cases and prognosis. There was no significant difference in the progression of T, N, Ex, and clinical stage in the 24 blood S-FN-positive cases and 15 patients with recurrence and metastasis (Table III).

Relationship between 39 blood S-FN-negative thyroid cancer cases and prognosis. In the 39 blood S-FN-negative cases, there was a significant difference between 15 cases of recurrence and metastasis in the progression of T, N, Ex, and clinical stage (Table III).

Therefore, blood S-FN negativity in all 63 thyroid cancer cases had a relevance between recurrence and metastasis. However, there was no association between recurrence and metastasis and blood S-FN positivity.

Association between blood S-FN levels in thyroid tissues and infiltrating cells in stage IVa and IVb and recurrent metastatic thyroid cancer. Since blood S-FN is secreted from the thyroid tissue, the relationship between localization and expression of S-FN in benign and recurrent metastatic thyroid tissues was examined by MoAb JT-95 immunohistochemical staining. V-factor, which showed a significant difference between blood S-FN-positive cases and blood S-FN-negative cases as a pathological factor, was also evaluated.

MoAb JT-95 immunohistochemical staining. All 15 cases of recurrent and metastatic thyroid cancer occurred from 25 stage IV cases with the status of having multiple lymph node metastases or extrathyroidal infiltration of tumors at the time of surgery (stage IVa,b: 24 cases), or the status of remote metastasis (stage IVc: 1 case). Of these, 24 stage IVa,b tissue sections were prepared, excluding one blood S-FN-negative case that already had massive tracheal invasion and multiple pulmonary metastases at the time of the consultation and could not undergo surgery (stage IVc: 1 case). Pathological evaluation of the 24 stage IVa,b cancer cases revealed 22 cases of papillary thyroid cancer, one of follicular cancer, and one case of undifferentiated cancer combined with papillary cancer. There were nine blood S-FN-positive cases and 15 blood S-FN-negative cases among these 24 stage IVa,b cancers. The 14 cases of metastasis and recurrence occurred from 24 stage IVa,b cases during the 6-25 years observation period after surgery, included 4 blood S-FN-positive cases (4/9, 44.4%) and 10 blood S-FN-negative cases (10/15, 66.7%).

Expression range of S-FN-positive cells in stage IVa,b thyroid cancer tissue. The staining results confirmed S-FN expression in cancer cells in 23 out of 24 cancer tissue preparations. No S-FN expression was observed in normal thyroid or connective tissues.

Blood S-FN-positive cases of stage IVa,b thyroid cancer (9 cases). S-FN was observed in cancer cells in nine cases. The area covered by S-FN-positive cells was less than 50% of the whole cancer tissue in four of these cases and 51% or more in five cases (Figure 1A and Table IV).

Blood S-FN-negative cases of stage IVa,b thyroid cancer (15 cases). Out of fifteen cases, in one case of follicular thyroid cancer, cancer cells were not stained at all, and in one case of undifferentiated cancer combined with papillary cancer, only parts of papillary cancer cells were stained, while undifferentiated cancer cells parts were not stained (Figure 1G). The area in which S-FN-positive cells were present, including these two cases, was less than 50% of the whole cancer tissue in eight cases, and 51% or more in seven cases (Figure 1B and Table IV).

S-FN abundance levels in thyroid cancer cells. S-FN abundance levels in thyroid cancer cells varied. It was observed in either the cell membranes inside the lumen structure or the cell membrane and cytoplasm, with heterogeneity in both blood S-FN-positive and -negative cases. However, in blood S-FN-positive cases, cells with a high staining intensity, wherein the whole cell was stained, were more prevalent than in S-FN-negative cases (Figure 1C and D).

V-factor examination. Of the 24 stained stage IVa,b cases, 20 showed vascular infiltration of cancer cells (20/24, 83.3%).

Blood S-FN-positive cases of stage IVa,b thyroid cancer (9 cases). Eight of nine cases (8/9, 88.8%) had infiltrating cells, and the proportion of S-FN-positive cells of infiltrating cells was 0-50% in one case and 51-100% in seven cases. Seven of eight patients had an S-FN-positive cell count ratio of more than 51% (7/8, 87.5%) (Figure 1E and Table IV).

Blood S-FN-negative cases of stage IVa,b thyroid cancer (15 cases). Twelve of 15 patients (12/15, 80%) were recognized as having intravascular infiltrating cells. The proportion of S-FN-positive cells of infiltrating cells was 0% in two cases, 1-50% in eight cases, and 51-100% in two cases. The proportion of S-FN-positive cells was more than 51% in two cases (2/12, 16.7%) and that of S-FN-negative cells was 50% or more in 10 of 12 blood S-FN-negative cases (10/12, 83.3%) (Figure 1F and Table IV). Hence, vascular invasion occurred at high rates in both stage IVa,b blood S-FN-positive and -negative cases. More than 51% of S-FN-positive cells of infiltrating cells were present in the blood S-FN-positive cases (7/8, 87.5%). However, more than 51% of S-FN-negative cells infiltrated the vessels in the upper 80% of the blood S-FN-negative cases (10/12, 83.3%).

Detailed intracellular and intravascular localization of S-FN using electron microscopy observation. The S-FN particles stained with MoAb JT-95 were recognized as black particles. When the red square part of Figure 1C is enlarged, S-FN particles could be found on the outermost cell membrane, along with immunostaining of S-FN-positive thyroid cancer cells (Figure 2A). Several S-FN particles were observed in the cytoplasm (Figure 2B). When the red square part of Figure 1E is enlarged, S-FN particles that were released into the bloodstream from the S-FN-positive cancer cells that infiltrated the blood vessel could be seen (Figure 2C).

Recurrent metastatic cases.

Examined cases. Of the 25 stage IV cases, 15 had recurrent metastases, in which cancer cells had metastasized to lymph nodes other than those to which it belongs or had metastasized to distant organs, such as lung and bone during the observation period. Among these, a total of 14 cases, including 4 blood S-FN-positive cases and 10 blood S-FN-negative cases, were stained with the MoAb JT-95.

Recurrent metastatic cases with positive blood S-FN. Of the eight positive blood S-FN stage IVa,b cases, in which vascular invasion by cancer cells was observed, four relapsed (4/8, 50%). In all four cases, the ratio of S-FN-positive infiltrating cells was 51-100% (4/4, 100%) (Table V).

Recurrent metastatic cases with negative blood S-FN. Of the 12 negative blood S-FN stage IVa,b cases, in which vascular invasion by cancer cells was observed, 10 cases had metastasis and recurrence (10/12, 83%). Among the 10 cases, eight (8/10, 80%) had an S-FN-positive infiltration cell ratio of 0%-50% and two cases had an S-FN-positive cell ratio of 51%-100% (Table V).

Vascular infiltrating cells were observed in all 14 cases of recurrent metastasis, regardless of whether the blood S-FN status was positive or negative. However, recurrent metastases rates were higher in the blood S-FN-negative cases with an S-FN-positive cell ratio of vascular invasive cells of 0-50%, compared to blood S-FN-positive cases with an S-FN-positive cell ratio of vascular invasive cells of 50% or more.

Association between blood S-FN and blood thyroglobulin levels in thyroid cancer. Estimation of blood thyroglobulin levels is a known tumor marker for thyroid cancer. Blood thyroglobulin levels in 63 patients with thyroid cancer were above normal (>33.7 ng/ml) in 26 cases (26/63, 41%). In contrast, blood S-FN was positive in 24 (24/63, 39%) cases. There was no significant correlation between blood S-FN expression and thyroglobulin levels in the 63 patients with thyroid cancer (Spearman’s rank correlation: r=0.02) (Figure 3).

As a result of previous amino acid analysis, the antigen substance of MoAb JT-95 has been found to be S-FN, which is secreted in the culture supernatant of SW1736, one of the thyroid cancer cell lines, and in the sera of thyroid papillary cancer patients (4,5). In this study, blood S-FN was expressed in approximately 38% of thyroid cancer cases, and the incidence was significantly higher than that of breast cancer. This expression rate was lower than the positivity rate of CA19-9 (Sialyl Lewis^a), a similar glycan antigen, in pancreatic cancer (70-80%), but showed the same positivity rate as tumor markers CA15-3 and BCA225 in breast cancer (6,7). In addition, since there was no correlation with blood thyroglobulin values, currently considered a tumor marker for thyroid cancer, the blood S-FN value cannot be used as an independent marker for thyroid cancer.

In medical care, the clinical stage is defined in consideration of the pathological progression of tumor size (T-factor), tumor infiltration outside the thyroid gland (Ex-factor), and the number and size of lymph node metastases (N-factor). The number of cases of recurrence and metastasis often increases with progression of clinical stage. In this study, 39 cases of negative blood S-FN and all 63 cases of thyroid cancer showed an association between progression of clinical stage based on pathological factors and locoregional recurrence and metastasis. However, no association was observed in 24 cancers with positive blood S-FN by statistical analysis.

In addition, all recurrences and metastases occurred in stage IV cases. Under histological staining by MoAb JT-95 in the 24 stage IV cases, intravascular infiltration by cancer cells was recognized in more than 80% of both stage IV blood S-FN-positive and -negative cases. In all 14 patients with recurrent metastasis, intravascular infiltration by cancer cells was observed in both groups. In the four recurrent metastasis cases with positive blood S-FN, S-FN-positive cells infiltrated more than 50% of the blood vessels, and the recurrence rate was 6.3% (4/63). In contrast, in 10 cases of negative blood S-FN with relapse and metastasis, S-FN-negative cells infiltrated more than 50% of blood vessels, and the recurrence metastasis rate was approximately 17.5% (11/63), which was approximately twice the recurrence rate among blood S-FN-positive cases. Hence, in the recurrent metastasis or stage IV cases with advanced medical conditions, blood S-FN-positive and -negative cases had the same probability of vascular invasion by cancer cells. However, in positive blood S-FN cases, with intravascular infiltration by S-FN-positive cells, it is thought that the probability of relapse and metastasis is lower than that in negative blood S-FN cases, intravascular infiltration by S-FN-negative cells.

Therefore, it can be concluded that a relationship observed between blood S-FN negative cases and progression of the condition in addition to recurrent metastasis was statistically significant, but the same was not observed in S-FN-positive cases. In addition, in the blood S-FN-positive cases, cells with high staining intensity infiltrated the vessels more than those in the blood S-FN-negative cases; thus, a significant difference was observed in the V-factor of the two groups.

It is presumed that S-FN-expressing cells have difficulty reaching metastatic target organs, such as the lungs and bones, even when infiltrating blood vessels, or that their ability to cause lesion metastasis is poor, even if they reach such organs compared to S-FN-negative cells. Our previous research suggests that lymphocytes and S-FN-expressing cancer cells have a high binding capacity (4,8). VLA-5 (α5β1), a cell adhesion factor belonging to the integrin family, exists as a receptor of FN on the surface of T-lymphocytes, and the binding FN and VLA-5 promotes T-lymphocyte activation and IL-2 secretion towards inhibition of cancer cell metastasis (9). These reports suggest that S-FN-expressing cells in thyroid cancer suppress their migration and the metastatic nest formation processes by activating immune system cells, such as T-lymphocytes by binding affinity of FN and VLA-5, even if they infiltrate into the blood vessels of cancerous tissue. This is possibly a reason why S-FN-positive cases have less recurrent metastasis compared with S-FN-negative cases.

S-FN is a glycan containing sialic acid. Sialyl Lewis^a (sLe^a), recognized by MoAb CA 19-9, is also an N-type glycan in which sialic acid is bound. It binds to selectin and acts as a ligand; it is a sialyl glycoprotein that belongs to the selectin family. With disease progression, sLe^a expression levels in blood increase in pancreatic, gastric, and colorectal cancers. Glycosylation of sLe^a and selectins is regulated by ST3GAL4 glycosyltransferase, which is present in the Golgi apparatus of cells. Up-regulation of ST3GAL4 also activates the c-Met signaling pathway and accelerates tumor invasion and cancer growth. These findings support the idea that malignancy accelerates according to the sLe^a increase in pancreatic, gastric, and colorectal cancers (10-12).

On the other hand, bisecting GlcNAc glycan, an N-type sugar chain, is present in combination with VLA-5, a member of the integrin family. The addition of bisecting GlcNAc to glycoproteins is regulated by β1,4-N acetyl glucosaminyl transferase III (GnT-III), which is present in the Golgi apparatus of cells. The overexpression of bisecting GlcNAc results in the suppression of the invasive ability and lung metastasis of the melanoma cell line B16. The affinity of the binding of integrin to fibronectin in vessel endothelial cells is significantly reduced by the overexpression of bisecting GlcNAc in integrin VLA-5 (α5β1) and inhibits spreading, migration, and focal adhesion of carcinoma cells to metastatic sites. These findings suggest that bisecting GlcNAc of the N-glycan of integrin suppresses tumor recurrence and metastasis (13-15).

S-FN, an antigen of the MoAb JT-95, is a glycoprotein with sialic acid in its N region. S-FN contains FN as a protein, and its expression site is on the cell membrane, as confirmed by immunostaining and electron microscopy. Moreover, it is almost identical to the positive site when thyroid cancer is stained with an anti-integrin antibody. Therefore, it is presumed to exist as a type of glycan protein on integrin α₅β₁ in thyroid cancer cells, and a part of it is released into the blood. In addition, since expression is not recognized by immunostaining of normal thyroid cells, S-FN appears to be overexpressed in cancer cells. In this study, the appearance of recurrent metastasis was lower in blood S-FN-positive cases than in blood S-FN-negative cases. One reason for this result is presumed to be that modification of the N-glycan of integrin by S-FN modulates the affinity of integrin, as well as bisects GlcNAc glycan, and induces the inhibition of spreading, migration, and adhesion to metastatic sites. Moreover, S-FN contains sialic acid at the end of the glycan structure. The presence of sialic acid residues on glycans of the β1 chain decreases binding to fibronectin in vessel endothelial cells (16). This may also be one of the reasons why S-FN-expressing cells cause inhibition of migration.

In summary, the reasons why blood S-FN-positive cases in thyroid cancer present less recurrent metastases compared to blood S-FN-negative cases, even if intravascular infiltration occurs in a similar ratio in both groups, are believed to be as follows: 1) activation of T-lymphocytes by FN, 2) modification of integrin by the binding of S-FN inducing down-regulation of cell adhesion and migration, and 3) the sialic acid residues of S-FN also decrease the binding and migration of cancer cells. S-FN-expressing cells are found in more than 90% of papillary cancers belonging to differentiated thyroid cancer by histological staining with MoAb JT-95, but not in poorly differentiated or undifferentiated cancers (4). In this study, S-FN was expressed in combined papillary cancer tissues but not in poorly differentiated and undifferentiated cancer tissues. It is thought that S-FN is mainly expressed in well-differentiated cancer and disappears in poorly or undifferentiated cancers with poor prognosis, with further gene mutations occurring and increasing malignancy. These histological findings also showed that blood S-FN-positive cases with high S-FN secretion in cells can be used as a good prognostic indicator.

Recently, in a case of thyroid cancer within 10 mm in diameter without lymph node metastasis, active surveillance has been recommended in clinical guidelines owing to its good prognosis (17). However, there are certain cases in which lymph node metastasis and undifferentiated conversion were observed during follow-up. Blood S-FN levels measurement may be a method for aiding active surveillance.

This study was performed at one facility only, and the number of cases examined is limited. Therefore, joint trials at multiple facilities are currently being planned to further examine the reliability of blood S-FN as a biomarker of thyroid cancer.

The Authors declare that there are no conflicts of interest.

H. Takeyama: Conceptualization, formal analysis, investigation, methodology, visualization, writing-original draft. Y. Manome: formal analysis, investigation, methodology, visualization, writing, reviewing, and editing.

Financial support: This study received financial support from the 52nd Japan Endocrine Surgery Congress 2019.