Breast Cancers Secreting Sialyl-fibronectin Are Less Likely to Cause Epithelial-mesenchymal Transition and Have Good Prognoses
1Department of Breast, Thyroid, Endocrine Surgery, Jikei University School of Medicine, Tokyo, Japan
2Department of Molecular Cell Biology, Jikei University School of Medicine, Tokyo, Japan
Fibronectin (FN) is a relatively large glycoprotein with a molecular weight of 210-250 kDa and is roughly categorized into humoral FN, which is present in the bodily fluids, such as the blood and synovial fluid, and cellular FN that is present in the connective tissue of the stroma.
Cellular FN is mainly secreted by fibroblasts present in the connective tissue and is associated with connective tissue formation and maintenance. Additionally, cellular FN is produced not only by fibroblasts but also by epithelial cells, neutrophils, and macrophages (1). FN produced in epithelial cells is secreted extracellularly and stored in the extracellular matrix (ECM) around the cell membrane.
Notably, FN in the ECM acts as a ligand of the integrin family, such as α5β1, which is a transmembrane receptor on the cell surface and is involved in the adhesion between cells and stroma and maintenance of cell-to-cell bonds (2).
Furthermore, FN is evidently involved in carcinogenesis since the amount of cellular FN secretion changes with its progression (1,2). However, the relationship between its expression and the malignancy and prognosis of cancer remains controversial.
Therefore, this study aimed to develop an enzyme-linked immunosorbent assay (ELISA) system to measure the blood levels of sialyl-fibronectin (S-FN), a type of cellular FN secreted by breast cancer cells.
To investigate whether blood S-FN levels are associated with cancer malignant potential and recurrent metastases, we examined the relationship between blood S-FN levels and 12 breast cancer cases that recurred or metastasized during 11 years of observational studies. Furthermore, we also examined the association between blood S-FN levels and anatomical, pathological, and molecular biological prognostic factors for breast cancer.
Materials and Methods
First, peripheral venous blood samples (7.5 ml) from 89 patients were collected using ethylenediaminetetraacetic acid tubes (Thermo Scientific, Tokyo, Japan) at the Jikei University Hospital from 2011 to 2019. Next, blood samples were centrifuged at 1,900×
In this study, 89 patients with breast tumors were analyzed using the experimental sandwich ELISA assay and tissue immunohistochemistry, similar to that used in our previous study to measure the secretion of S-FN in the blood and expression of S-FN in tissues (3).
Formalin-fixed, paraffin-embedded breast cancer tissue sections (3 μm slices) were prepared from 12 cases. MoAb JT-95 antibody and the i-view DAB Universal Kit (Ventana Medical Systems, Tucson, AZ, USA) were used as the primary antibody for these tissue sections at a concentration of 0.005 μg/ml and the secondary antibody, respectively.
However, a significant difference between the blood S-FN-positive and S-FN-negative patients was found regarding three factors as follows: l) locoregional recurrence status (
Therefore, we examined the association between locoregional recurrence, remote metastases, and prognostic and predictive factors, including HMS, in blood S-FN-positive, blood S-FN-negative, and total breast cancer groups.
Of these 21 blood S-FN-positive patients, recurrent metastases were observed in four cases (4/21, 19.1%), including three locoregional recurrences and one remote metastasis. In contrast, recurrent metastases were seen in 61 blood S-FN-negative patients (8/61, 13.1%), including two locoregional recurrences and six remote metastases.
The prognostic factors (T-factor, N-factor, NG, MIB-1 index, HR status, HER-2 status, and HC) were almost identical among S-FN-positive and S-FN-negative patients.
Regarding HMS, a statistical association between the S-FN positive group and the S-FN negative group was found. Cancer cell residues were observed at the histological margins in all three S-FN-positive patients, which had local recurrences 6-13 years postoperatively. However, the two blood S-FN-negative patients were HMS-negative at the initial surgery but developed local recurrences 4-10 years postoperatively.
Treatments were partial breast resection (Bp)+ sentinel lymph nodes biopsy (SNB) or axillary lymph nodes dissection (Ax), followed by radiation therapy for the remaining breast, and then hormonal therapy (HT) until recurrence in all 5 patients. Additionally, in all HR-positive patients, a selective estrogen receptor modulator (SERM) or an aromatase inhibitor (AI) or a selective estrogen receptor down regulator (SERD) was administered as post- operative and recurrent metastases hormonal therapy.
Patient 3 of the S-FN-positive patients showed metastasis to the axillary lymph nodes at the time of surgery, thus HT therapy was preceded by adjuvant chemotherapy, including four cycles of anthracycline and cyclophosphamide and four cycles of docetaxel (four cycles of AC+Dx).
After recurrence, four patients (Patients 1, 2, 4, and 5) with breast recurrence underwent breast total mastectomy (BT) and continued with HT therapy. Patient 3 had a lymph node recurrence and received cyclin-dependent kinase 4/6 (CDK-4/6) inhibitor in addition to HT therapy. The prognosis after local recurrence was good, and no new local recurrence or distant metastasis were observed in all five patients.
Regarding HC, S-FN-positive patients were IDC (one patient), and the six S-FN negative patients were IDC (two patients) and ILC (four patients).
Regarding HER-2 type, neoadjuvant chemotherapy (NAC) using six cycles of fluorouracil+epirubicin+cyclophosphamide (FEC) and six cycles of trastuzumab+pertuzumab+docetaxel (TPDTX) was administered, followed by breast total mastectomy + axillary lymph nodes dissection (BT+Ax) surgery. Postoperative histopathological diagnosis (POPD) was pathological complete response (pCR) for the mammary glands and lymph nodes, and after NAC, the metastatic lung lesions also disappeared on the CT scan. Postoperatively, trastuzumab+pertuzumab was continued for 1 year, followed by hormonal therapy.
Patient 7 of the S-FN-negative patients had T3N1M0, stage IIIA cancer, and was also an HER-2 type like Patient 6; NAC therapy using six cycles of FEC and TPDTX was administered, followed by BT+Ax surgery. The POPD showed pCR in the mammary glands and axillary lymph nodes, similar to patient 6. The postoperative course involved the administration of trastuzumab+pertuzumab for 1 year. But during the 1.5 years of postoperative follow-up, bone and brain metastases developed.
Patient 11 had a ≥5 cm tumor (T3) and lymph node metastasis (N1). It was Stage IIIA. For this reason, NAC therapy was administered in four cycles of anthracycline+cyclophosphamide and four cycles of docetaxel (four cycles of AC+Dx). After NAC, a BT+Ax operation was performed.
HMS was negative, but the POPD was T1N1 (2/7). Capecitabine was administered for 6 months for residual tumors. However, bone metastasis developed 1 year post-operatively. In the S-FN-negative group, Patient 8 did not receive NAC, and Patients 9, 10, and 12 received surgery, radiation therapy, and hormone therapy, respectively.
As treatment for distant metastases, adjuvant chemotherapy and hormone therapy + CDK 4/6 inhibitor therapy were administered. In patient 12, partial lung resection was performed to remove the site of the lung metastasis in addition to hormone therapy (
Regarding prognosis, of the six S-FN-negative patients, two patients (Patients 10 and 12) were successfully treated, and no progression of the disease was observed, and one S-FN-positive patient (Patient 6) also had no disease progression. However, four patients of S-FN-negative group showed progression due to treatment resistance, and two patients died after multiple distant metastases.
The blood S-FN was positive in 21 (21/82, 25.6%) patients. However, no statistical correlation was found between S-FN blood levels and CEA and CA15-3 levels in the 82 patients with breast cancer (Spearman’s rank correlations: CEA: R=0.0039, CA15-3: R=-0.098) (
Breast cancer is the most prevalent disease among women in Japan, and its incidence is still increasing (4). Conversely, since the 1990s, the overall incidence of breast cancer has decreased in Europe and the United States. However, the number of cases of recurrent metastases with a poor prognosis has recently increased, particularly among younger patients; therefore, attention is required (5).
Notably, the 5-year survival rate is relatively good (approximately 80%) because of the advances in breast cancer treatment. Currently, the recommended breast cancer treatment is the combination of surgery, drug therapy, and radiotherapy while considering prognostic predictors and drug selection factors (6).
Conventionally, as a prognostic predictive factor for breast cancer, anatomical factors, such as T-factor, presence of axillary N-factor, presence of M-factor, and the clinical stage combined with these factors have been considered useful. Similarly, pathological factors, such as Ly-factor, v-factor, HC, histological grade (HG) or NG, and HMS, are also useful. Moreover, HR, HER-2, and MIB-1 status as biological factors are important prognostic factors and drug treatment selectors (7).
In clinical practice, locally advanced breast cancers with increased tumor diameter and many axillary lymph node metastases tend to have more recurrent metastasis in the future. Also, ILC, which is a special type of histological classification, accounts for approximately 5-10% of all breast cancers. It lacks E-cadherin, which is a cell adhesion factor, and the cells are easily released from the tissue, increasing the likelihood of recurrent metastasis compared with IDC, which accounts for 90% of all breast cancers (8). Additionally, it is known that in breast-conserving therapy cancer residue present in the margin is a risk factor for local recurrence (9).
In this study, of the seven benign breast lesions, blood S-FN expression was positive in only two cases of papilloma, which is a ductal epithelial benign tumor. Patients with phyllodes tumors and fibroadenomas, which are stromal proliferative lesions, do not express S-FN in the blood. However, among the 82 patients with breast cancer caused by malignant gene mutations in the normal duct epithelium, S-FN was detected in the blood of 21 of the 61 patients. These events suggest that blood S-FN is secreted from the tumorized ductal epithelium rather than the stroma or normal mammary epithelium.
No statistical association between the 21 blood S-FN-positive and 61 blood S-FN-negative cases was found with each of the following prognostic factor: T-factor, N-factor, Ly-factor, v-factor, NG-factor, HC-factor, MIB-1 index, HR status, and HER-2 status. However, a significant difference in HMS, which is a risk factor for local recurrence, was found between blood S-FN-positive and S-FN-negative patients (Table II). Additionally, the analysis of the association between blood S-FN expression and prognostic factors in the five locoregional recurrence cases revealed statistically significant association between blood S-FN-positive and HMS-positive (Table VI). This may be because, in the 21 S-FN-positive patients, all three patients with positive HMS developed locoregional recurrence in the breast (2 cases) and of cervical LN (metastasis; 1 case). In contrast, among the 61 blood S-FN-negative cases there were 2 breast recurrences; however, both patients were HMS negative (Table III).
These results suggested that local recurrence in S-FN-positive patients was largely due to the persistence of marginal carcinoma in S-FN-positive patients. However, it is presumed that local recurrence in blood S-FN-negative patients is because cell malignancy is higher in blood S-FN-negative patients than in blood S-FN-positive patients.
Many studies have reported that the amount of blood FN is associated with prognosis; however, the results are controversial. The mechanisms by which elevated blood FN levels become a poor prognostic factor for cancer are as follows:
1. As cancer cell proliferation progresses, FN, growth factors, and cytokines secreted by fibroblasts and macrophages in the tumor microenvironment (TME), which is the stroma around the tumor, induce epithelial-mesenchymal transition (EMT), which is a type of tumor transformation. Because of this phenomenon, cells that change from epithelial cells to stromal cell-like mesenchymal cells appear. Conventional epithelial and mesenchymal cells are mixed in the tumor, and a state of heterogeneity, which is recognized by pathological observation, is formed (10,11).
2. Cancer cells not inducing EMT, secrete FN from within the cells, and FN accumulates in the ECM; however, cancer cells that have changed to the mesenchymal cells do not produce FN; consequently, the FN in the ECM is reduced and depleted. In this state, mesenchymal cancer cells transform into fibroblast-like cells and secrete a large amount of FN (stromal fibrillar FN) into the TME (12,13).
3. This increased secretion of stromal fibrillar FN triggers the release of EMT-promoting substances, such as snail, N-cadherin, and vimentin, and the activation of proteolytic enzymes such as matrix metalloprotease 2 (MMP 2).
The action of these factors promotes a further increase in the number of mesenchymal cancer cells. It also promotes E-cadherin degradation. Furthermore, cell surface integrin α5β1, which binds to FN in the ECM and keeps cells to an aggregated state, and α2β1, which is fixed to pericellular collagen and forms an aggregated state with α5β1, are degraded by MMP 2 (14-16).
4. These phenomena mainly cause the release, invasion, and intravascular migration of mainly mesenchymal cancer cells from cancerous tissues. These cells cause recurrent metastases in target organs (12,13,17).
Specifically, fibroblast-like mesenchymal cancer cells induced by EMT secrete large amounts of stromal fibrillar FN and form recurrent metastases. Therefore, it is presumed that blood FN levels would become higher by stromal fibrillar FN. Clinical studies measuring actual blood FN levels have reported an association between high expression of blood FN and recurrent metastases (10,11,18).
In contrast, epithelial cancer cells not changed by EMT produce FN and store it in the ECM. This autocrine FN binds to the actin fiber and integrin α5β1 in the endoskeleton, maintaining the cytoskeleton, cell adhesion, and aggregation. These events prevent cell migration that causes recurrence of metastasis. Interestingly, these phenomena have been reported in
Additionally, epithelial cancer cells, which is still producing autocrine FN, and mesenchymal cancer cells, not producing autocrine FN, were intravascularly administered separately to nude mice, and lung metastatic ability was compared
Another mechanism involves the induction of natural killer (NK) cells in the TME during cancer development. NK cells produce interferon-γ (IFN-γ)
The total FN amount in the blood was estimated as the sum of the amount of fibrillar FN secreted from mesenchymal cancer cells in the stroma and that of autocrine FN secreted by the epithelial cancer cells.
The role of elevated blood FN levels in recurrent cancer metastasis is controversial because elevated stromal fibrillar FN levels suggest an increase in the mesenchymal cancer cells and promotion of recurrent metastases. Conversely, an increase in the amount of autocrine FN is believed to indicate that cell migration is blocked, and metastasis is suppressed.
Current blood FN measurement systems only measure the entire amount of FN in the blood. However, if autocrine and stromal fibrillar FN can be distinguished and identified, screening clinically favorable cases with epithelial cancer cells may be possible.
S-FN, an antigen of MoAb JT-95, is produced from the membrane component of papillary thyroid cancer and is detected in the culture supernatant of SW-1736, which is a thyroid cancer cell line, and in the blood of patients with thyroid cancer (3,30).
In this study, secretion of S-FN into the blood was observed in approximately 25.6% of the patients with breast cancer. Moreover, observing breast cancer tissue immunostained with MoAb JT-95 antibody using light and electron microscopes revealed secretion only from breast cancer cells rather than the stroma (Figure 1).
S-FN is a glycoprotein in which sialic acid is conjugated to FN. This makes it possible to distinguish autocrine FN from stromal fibrillar FN in the blood. Additionally, MoAb JT-95 loses its reactivity toward S-FN when sialic acid is following treatment with sialidase; it is believed to recognize the glycan structure near the junction of FN and sialic acid. Therefore, MoAb JT-95 does not react with the FN body, and therefore, does not detect stromal fibrillar FN (31).
From the above, it is considered that this blood S-FN measurement system using MoAb JT-95 can distinguish the autocrine FN secreted by epithelial breast cancer cells from the stromal fibrillar FN secreted from stromal mesenchymal cells altered by EMT. Therefore, S-FN-positive cases with a high level of S-FN in the blood indicate that the secretion of autocrine FN is maintained. This suggests that S-FN-positive cases exist in many epithelial cancer cells that do not undergo EMT and whose malignancy and metastasis are suppressed. Hence, the prognosis of blood S-FN-positive patients was considered better than that of S-FN-negative patients. Conversely, in S-FN-negative patients, where mesenchymal-status cells are more common in tumors than in S-FN-positive patients, locoregional recurrences may have occurred despite being HMS-negative. Additionally, a significant difference was found in remote metastasis between the blood S-FN-positive and S-FN-negative patients (Table II). It is presumed that six of seven remote metastases cases were found in blood S-FN-negative patients, and only one was found in S-FN-positive patients (Table IV). Notably, this suggests that blood S-FN-negative patients have worse prognoses than S-FN-positive patients, even with the M-factor. Regarding each prognostic factor in patients with remote metastasis, an association was found between the T-factor, N-factor, clinical stage, and HC in 82 total breast cancer cases (Table VI). This indicates a clinically commonly experienced tendency that hematogenous metastasis increases with an increase in T-factor and N-factor and increases due to histopathological features such as ILC (6-8). Therefore, the population of 82 patients examined in this study is considered to be common. In the blood S-FN negative patients, a relationship was found between the M-factor and HC, and N-factor (Table VI).
The HC was ILC in 9 and IDC+DCIS in 73 patients out of the 82 breast cancer cases in this study, and the special subtype ILC was approximately 11% of all cases, with a typical histopathological composition (8). A statistical association was found between remote metastasis and HC in S-FN-negative patients. This may have been statistically relevant because four out of the six S-FN-negative patients with distant metastases were ILC (4/6, 66.7%). Moreover, out of the nine ILC patients, two S-FN-positive patients did not develop metastasis.
ILC lacks E-cadherin, which is a cell adhesion molecule, in both blood S-FN-positive and S-FN-negative patients. However, in blood S-FN-negative patients, it is presumed that many mesenchymal cancer cells that secrete the autocrine FN (S-FN) were missing; this is important as these autocrine factors are involved in cells aggregation and sticking with surrounding tissues. It has also been suggested that blood S-FN–negative patients are prone to cancer cell migration, invasion, and M-factor due to this disappearance. Therefore, there were many distant metastasis patients among S-FN–negative patients compared with S-FN-positive patients, in which many epithelial cancer cells maintained S-FN secretion.
For the N-factor, which represents the number of lymph node metastases an association was found between the N-factor and M-factor in blood S-FN-negative patients (Table VI).
Clinically, this is probably because in the S-FN-negative patients, N-factor was observed in four out of six M-factors, and there was also one patient with ≥4 metastases (N2), while in the blood S-FN-positive patients, only one patient with inflammatory breast cancer had lymph node metastases (Table IV).
N-factor was more common in blood S-FN-negative patients and was associated with the M-factor, presumably because there are multiple mesenchymal cancer cells in the tumor in blood S-FN-negative patients as in HC. Regarding the T-factor, an association was found between blood S-FN-positive cases and remote metastases (Table VI).
In the blood S-FN-negative patients, five out of six M-factors had a diameter of ≥2 cm, and M-factors developed with an increase in T-factor. In contrast, among blood S-FN-positive patients, M-factor was observed in only one patient with T4d cancer (Table IV).
However, no statistically significant association between the M-factor and T-factor in blood S-FN-negative patients was found, whereas an association with the T-factor was observed in blood S-FN-positive patients. It is presumed that one M-factor patients with blood S-FN-positivity had inflammatory breast cancer, which is a special type of T4d cancer, where, regardless of T-factor, cancer cells widely invade the subcutaneous lymphatic vessels of the mammary gland, causing redness and swelling of the skin. Therefore, this special type of T4d cancer may have affected the association between an increase in T-factor and M-factor observed in blood S-FN-negative patients.
CA15-3 and CEA, which are markers for breast cancer, are glycoproteins whose levels tend to increase in the blood as breast cancer progresses. Specifically, they will likely increase when recurrent metastases develop (32). S-FN is also a glycoprotein; however, in the 82 patients examined in this study, no correlation was found between blood S-FN and CEA and CA15-3 levels (Figure 2). This may also indicate that S-FN expression is not involved in the exacerbation or progression of breast cancer. In this clinical study, S-FN-positive (Patient 6) and S-FN-negative (Patient 7) patients with HER-2-positive breast cancer were treated with the same FEC+TPD as NAC. Both patients had pCR in the axillary lymph nodes and mammary glands on postoperative histopathological examination, and M1 (lung metastases) of the inflammatory carcinoma in the S-FN-positive patient (Patient 6) was not detectable on the CT images.
However, patients who were S-FN positive were still in complete remission 5.8 years postoperatively, whereas those who were S-FN negative developed bone and brain metastases 1.5 years postoperatively (Table IV, Table V). In addition to the two HER-2 type patients, one blood S-FN-positive (Patient 3) and five blood S-FN-negative patients (Patients 6, 7, 9, 10, and 11) were treated with chemotherapy or a HT(SERD)+CDK 4/6 inhibitor as adjuvant therapy after recurrent metastases. Of these, one blood S-FN-positive patient (Patient 3) and one blood S-FN-negative patient (Patient 10) did not show any new recurrent metastases after additional treatment; however, in four blood S-FN-negative patients, new metastases appeared even after additional treatment, and the disease progressed (Table V).
Eribulin mesylate (eribulin) is a non-taxane microtubule dynamics inhibitor that has shown trends towards greater overall survival than progression-free survival in patients with late-stage metastatic breast cancer. This mechanism is believed to be due to mesenchymal-epithelial transition (MET) effects of eribulin, which reverses the mesenchymal cell status to epithelial cell status, enhances anticancer efficacy, and extends overall survival (33).
In this clinical study, even with the same anticancer drug, the effect of the treatment administered on patients with recurrent metastases, including the HER-2 and luminal types, was good in blood S-FN-positive patients; however, further progression was observed in blood S-FN-negative patients. This may be due to the presence of more mesenchymal cancer cells resistant to chemotherapy in blood S-FN-negative patients, than in blood S-FN-positive patients. Anticancer drug treatment is effective for breast cancer; however, it also has many side effects. Therefore, genetic analyses such as Oncotype DX are performed in HR-positive cases to prevent unnecessary anticancer drug treatment. In HER-2-positive cases, it is recommended to avoid anthracycline, which is cardiotoxic, in the early stages, such as T1N0 (6).
Therefore, in addition to the prognostic predictors, blood S-FN measurements may help differentiate cases with a favorable prognosis. This study has some limitations. First, it was conducted at a single institution. Second, the number of cases was insufficient. Therefore, we are currently planning a joint study with other institutions to better ascertain the relationship between the expression and prognostic significance of blood S-FN levels in breast cancer.
In this study, patients with positive blood S-FN, which is a type of autocrine FN secreted by epithelial cancer cells, had a better prognosis and lower treatment resistance than those with negative blood S-FN, with many mesenchymal cancer cells altered by EMT.
This study received financial support from The 52nd Japan Endocrine Surgery Congress 2019.
Conflicts of Interest
The Authors declare that there are no conflicts of interest in relation to this study.
All Authors contributed to the study conception and design. Material preparation and data collection and analysis were performed by H. Takeyama, and Y. Manome. The first draft of the manuscript was written by H. Takeyama, and all authors commented on previous versions of the manuscript. All Authors read and approved the final manuscript. Conceptualization: H. Takeyama. Methodology: Y. Manome. Formal analysis and investigation: H. Takeyama and Y. Manome. Writing - original draft preparation: H. Takeyama. Writing - review and editing: Y. Manome. Supervision: H. Takeyama.