Abstract
Background/Aim: Perineural invasion (PNI) is a significant pathological feature in head and neck cancer. The molecular mechanisms of PNI are poorly understood. Contrary to the previous belief that cancer cells invade nerves, recent studies have shown that Schwann cells (SC) can dedifferentiate, intercalate between cancer cells, and promote cancer dispersion. Communication between cells through brain-derived neurotrophic factor (BDNF) activation of its receptor tropomyosin receptor kinase B (TRKB) may contribute to these cellular events. We aimed to determine the effect of TRKB inhibitor ANA-12 on the direction of cell migration and degree of SC-induced oral cancer cell dispersion. Materials and Methods: Cell migration and dispersion assays were performed in vitro using murine SC and oral carcinoma cell lines. Assays were performed with and without ANA-12. Results: Although SCs preferentially migrated towards cancer cells in control medium, there was minimal SC-associated cancer cell dispersion. In contrast, treatment with ANA-12 reduced migration of SCs and cancer cells towards each other and initiated more SC-associated cancer cell dispersion. Conclusion: This pilot study shows that BDNF–TRKB signaling may have a role in regulating interactions between SC and oral cancer cells that affect cell migration, intercalation, and cancer cell dispersion. Further research into these interactions may provide important clues about the molecular and cellular mechanisms of PNI.
- Perineural invasion
- Schwann cells
- cancer cell dispersion
- migration
- head and neck cancer
Head and neck cancer (HNC) is the sixth most common type of cancer in the world, with approximately 630,000 new cases diagnosed each year (1). Poor prognostic factors associated with HNC include increased age and advanced tumor staging (2, 3). Studies have shown that perineural invasion (PNI) is also a poor prognostic factor in HNC, particularly in oral cavity cancer (4-10).
PNI is the tracking of cancer cells along a neuron or within the neuronal sheath (7). Reports describing neoplastic invasion of peripheral nerves date as early as 1842 (11-13). In squamous cell HNC, PNI is associated with increased risk of local recurrence and reduced survival (14, 15), and therefore, presence of PNI in HNC is an indication for adjuvant radiotherapy (15). The exact mechanisms of PNI in HNC are not well understood and therefore effective therapies targeting neuronal involvement in cancer do not exist.
Although PNI was traditionally thought to be a cancer-driven process, the nerve microenvironment is now thought to be an important regulator. Schwann cells are glial cells that wrap around axons of peripheral nerves and aid in neuronal survival and repair after injury (16, 17). Schwann cells are able to extend cellular processes and dedifferentiate into migratory subtypes that act to repair neurons and guide neuronal growth (17-20). Using similar actions seen in nerve repair, Schwann cells can potentially promote or regulate PNI in cancer (21-27). Deborde and colleagues demonstrated how Schwann cells can dedifferentiate, become motile, and travel toward pancreatic cancer cells (22). They also showed how Schwann cells can intercalate and disperse cancer cells and aid in cancer cell migration towards nerves.
The mechanisms behind PNI may involve reciprocal signaling interactions between tumor cells and nerve components, in which nerves release trophic factors that stimulate cancer cells, and cancer cells release neurotrophic factors that stimulate nerve infiltration (9). In support of this theory, studies involving prostate and gastric cancer have shown that denervation can suppress tumorigenesis, tumor progression, and metastasis (28, 29). As it pertains to squamous cell HNC, several neurotrophic factors have been proposed to contribute to neurite outgrowth and PNI (9), including nerve growth factor, brain-derived growth factor (BDNF), and neurotrophin-3 (Figure 1A), which interact with their high affinity receptors, tropomyosin receptor kinase A, B, and C (TRKA, TRKB, and TRKC), respectively. In particular, BDNF and its high affinity receptor TRKB are highly expressed in squamous cell HNC and have been shown to be involved in cell proliferation in other cancer types (9, 30-34).
In this pilot study, we created an in vitro migration and dispersion assay by co-culturing murine Schwann cells and oral cavity squamous cell carcinoma. We used the compound ANA-12 (Selleckchem, Houston, TX, USA), a non-competitive inhibitor of TRKB (35), to study the effect of BDNF–TRKB signaling on cell-to-cell interactions between cell types.
Materials and Methods
Cell lines and culture. The B4B8 oral squamous cell carcinoma cell line was characterized by Thomas et al. (36) and obtained for our study as a generous gift. In brief, mucosal keratinocytes of BALB/c mice were extracted and transformed in vitro with 4-nitroquinolone-1-oxide before inoculation into severe combined immunodeficient (SCID) mice. B4B8 cells were re-isolated from tumors in the SCID mice and confirmed to express markers of epithelial origin. B4B8 cells were cultured in keratinocyte-SFM medium with epidermal growth factor-1, bovine pituitary extract, and glutamine factors (Thermo Fisher, Waltham, MA, USA). Adult rat Schwann cells were isolated from normal peripheral nerve and cultured according to methods reported by Andersen and Monje (37) and were obtained from the Miami Project to Cure Paralysis (Monje Laboratory, Miami, FL, USA) as a generous gift. In short, the methods of Schwann cell isolation involved removal of epineurium, extensive teasing of nerve fibers, removal of myelin debris, dissociation of cells, expansion of initial cell population in medium with chemical mitogens, and preparation of Schwann cells as cryogenic stocks for experiments. The Schwann cells were transfected with lentiviral fluorescent protein vectors and expanded at 37°C and 5% CO2 on 0.01% poly-L-ornithine (Sigma–Aldrich, St. Louis, MO, USA) and laminin (5 μg/ml; Thermo Fisher) coated plates. Mitogenic media consisted of Dulbecco's modified Eagle's medium (DMEM), 10% heat-inactivated fetal bovine serum (FBS; Seradigm, VWR, Radnor, PA, USA), and heregulin (10 nM; Sigma–Aldrich).
Cell viability assay. Cancer and Schwann cell viability were tested with different levels of ANA-12 (Selleckchem), a selective, non-competitive antagonist of TRKB. Cells were cultured on a 96-well plate that was precoated with laminin (5 mg/ml) and 0.01% poly-L-lysine. Approximately 20,000 cells were seeded in each well and incubated for 24 h at 37°C and 5% CO2. After 24 h, cells were cultured with DMEM with 10% FBS or different concentrations of ANA-12 up to 5 μM for 4 days. Subsequently, crystal violet colorimetric assay was performed and absorbances were measured for each well to determine cancer and Schwann cell viability in the presence of ANA-12. There was no significant difference in viability between control condition and within cancer and Schwann groups treated with different concentrations of ANA-12 (p=0.06 for cancer cells; p=0.58 for Schwann cells, data not shown). Therefore, it was determined that ANA-12 was not cytotoxic to the cancer or Schwann cells at the concentrations tested and ANA-12 at 5 μM was used for further experimentation.
Western blot. Western blots were performed to confirm expression of TRKB and BDNF in both cancer and Schwann cell lines. Protein was isolated using RIPA buffer (Thermo Fisher). Protein concentration was measured using the BCA Protein Assay Kit (Thermo Fisher), per manufacturer's instruction. For western blotting, 50 μg of proteins were loaded on a precast gel (Bio-Rad, Hercules, CA, USA), separated by gel electrophoresis, and transferred to a nitrocellulose membrane. Membranes were blocked in 2.5% bovine serum albumin and incubated overnight in primary antibody solution at 4°C and Alexa Fluor 488 and 647-conjugated secondary antibodies (1:1000; Life Technologies, Carlsbad, CA, USA) for 2 h at room temperature. The following primary antibodies from Abcam (Cambridge, MA, USA) were used: BDNF (1:500, Ab108319), TRKB (1:250, Ab18987), beta-actin (1:1000, Ab8227), and beta-tubulin (1:1000, Ab78078). Fluorescent images were obtained using the ImageQuant LAS 4000 Gel Biomolecular Imager (GE Healthcare Life Sciences, Marlborough, MA, USA). Images of the western blot confirmed expression of TRKB and BDNF in both Schwann and cancer cells (Figure 2).
Migration and growth assay. Spheroids consisting of cancer cells and Schwann cells were cultivated in three-dimensional (3-D) hanging drop plates (Perfecta3D®, Sigma–Aldrich) (38). Spheroids were then plated in two-dimensional (2-D) culture on 6-well plates precoated with laminin (5 mg/ml) and 0.01% poly-L-lysine (Figure 3A). Four internal controls included cancer cells alone, Schwann cells alone, cancer cells exposed to ANA-12 (5 μM), and Schwann cells exposed to ANA-12 (5 μM). Once in 2-D culture, differential interference contrast (DIC) light and fluorescent microscopy images were taken daily for 4 days. Each image was divided into 12 equivalent pie-shaped sections with the center originating at the spheroid (Figure 3B). The cells were counted in each section daily for 4 days. The percentages relative to the total number of cells were calculated and graphed on radar cell plots.
Cell dispersion assay. For the co-culture cell dispersion assay, cancer and Schwann cell spheroids were plated simultaneously on 6-well plates and cultured in control media or with ANA-12 (5 μM). Once in 2-D culture, differential interference contrast light and fluorescent images were taken daily for 4 days. Once the cancer and Schwann cells came into contact, cancer cells became dispersed by Schwann cells in various conditions. Dispersed cells were defined as any cancer cell separated from the cancer cluster and completely surrounded by Schwann cells (Figure 4). The number of dispersed cancer cells were counted during the first 4 days.
Statistical analysis. Statistical analysis was performed with analysis of variance test with Tukey post-hoc testing or Kruskal–Wallis test with multiple pairwise comparison tests. Significance was set at p<0.05.
Results
Spheroid migration and growth assay. For the internal control groups in control medium, cancer cells and Schwann cells grew and migrated outward from the center of the clusters over 4 days. Representative images are shown in Figure 4. The pattern of growth of both cancer cells and Schwann cells appeared to be random in all directions. When graphed on radar plots, the cancer cells in control medium showed a degree of nonrandom migration on days 1 and 2, however, the percentage of cells became consistent in each section on days 3 and 4. The Schwann cell plots remained circular, which demonstrated uniform migration in all directions in days 1-4 (Figure 5A, lower panel). We also noted that Schwann cells demonstrated a characteristic whorling pattern in culture (Figure 5A, upper panel).
Similarly to cancer cells and Schwann cells in control medium, both cancer cells and Schwann cells in the presence of ANA-12 demonstrated a random pattern of migration and growth (Figure 4B, lower panel). In addition, there was no effect of ANA-12 on the characteristic whorling pattern of Schwann cells. When exposed to ANA-12, the nonrandom migration by cancer cells seen on day 1 and 2 was less pronounced when compared to the control group (Figure 5B, upper panel). There was no difference in the random Schwann cell migration and growth in the ANA-12-treated group compared to control.
Co-culture migration and cell dispersion assay. Subsequently, Schwann cells and cancer cells were co-cultured in control medium. Time-lapse images initially showed intercalation of Schwann cells into cancer cell clusters; however, by day 4, the cancer appeared to be growing as a large mass, pushing Schwann cells to the periphery (Figure 6A). Co-cultures in control medium showed cancer and Schwann cells maintaining a well-defined border (Figure 6B and C) with minimal cancer cell dispersion.
In contrast, when Schwann and cancer cells were co-cultured in ANA-12, a different pattern of growth and migration was appreciated (Figure 7A). The well-defined border between cancer cells and Schwann cells was no longer apparent. Schwann cells appeared more active, intercalating between cancer cell clusters, directing cancer cell migration between their whorls, leading to cancer cell dispersion (Figure 7B-D). In these co-cultures exposed to ANA-12, there was a significant increase in what appeared to be Schwann cell-associated cancer cell dispersion in the ANA-12-treated group at all time points when compared to the control (Figure 8A).
With the cancer and Schwann cell co-cultures, we also measured the direction of migration and growth using radar plots of the co-cultures on the first 2 days. In control medium, cancer cells demonstrated migration and growth towards Schwann cells, and Schwann cells demonstrated migration and growth toward cancer cells. These results demonstrate there is preferential migration and growth of cancer cells and Schwann cells towards each other. However, this preferential migration was not seen in the presence of ANA-12 (Figure 8B).
Discussion
Schwann cells have recently come into focus as having an important role in regulating the tumor microenvironment and PNI. In a 3-D in vitro experiment, Demir et al. demonstrated that Schwann cells exhibited early and targeted migration towards human pancreatic cancer cells prior to cancer cells migrating toward peripheral neurons (39). Deborde et al. demonstrated that dedifferentiated Schwann cells directed pancreatic cancer cell migration towards nerves in a neural cell adhesion molecule 1 (NCAM-1)-dependent mechanism (22, 40). These studies provide some clues that cancer cells have the potential to initiate Schwann cell migration and activity that can promote PNI in HNC.
BDNF and TRKB are highly expressed in HNC and their expression have been associated with tumor growth, invasion, and metastasis (32). In squamous cell HNC, BDNF–TRKB signaling was linked to aggressive tumor behavior, poor prognosis, and resistance to chemotherapy (30, 32, 41-43). Furthermore, overexpression of BDNF and TRKB in salivary adenoid cystic carcinoma was correlated with invasion, metastasis, and poor prognosis (44). It has been suggested that BDNF and TRKB signaling may also be involved in the pathogenesis of PNI in HNC (9).
In our study, we chose to investigate how BDNF–TRKB signaling affects cell migration and cancer dispersion in co-cultures of murine Schwann cells and oral squamous cell carcinoma cells. To do this, we first performed a western blot to confirm that both Schwann cells and cancer cells express BDNF and TRKB (Figure 2). Then, we utilized a 3-D spheroid to 2-D plate assay (Figure 3) for our investigation and performed cell migration and growth assays individually for Schwann and cancer cells in the presence and absence of a TRKB inhibitor (ANA-12). Our study found that Schwann cells and cancer cells exhibited random growth and migration overall and that ANA-12 did not affect this finding (Figure 5). Lastly, we performed co-cultures of Schwann and cancer cell spheroids and performed migration and cell dispersion assays with and without ANA-12. Our findings demonstrated that Schwann cells and cancer cells preferentially migrate towards each other and that TRKB inhibition with ANA-12 blocked the preferential migration that was seen (Figure 8). However, although TRKB inhibitor ANA-12 reduced the preferential migration of Schwann and cancer cells towards each other, noticeable differences in the cell-to-cell interactions were observed under both treatment conditions when the two cell types came into direct contact with each other.
Upon cell contact, it appeared that the cancer cells were growing as a single mass invading into the Schwann cell cluster and herding Schwann cells to the periphery. Under the control conditions, it appears that the cancer cells disregard the boundaries set up by the Schwann cells (Figure 6). This pattern of cancer cell invasion may mimic what potentially occurs in direct PNI and perhaps the reciprocal signaling interactions between cell types is responsible for what was observed. However, in the presence of ANA-12, Schwann cells became more active against intrusion by the cancer cell cluster. Schwann cells intercalated between cancer cells, ultimately dispersing small collections of cancer cells from the much larger cancer cell cluster (Figures 4, 7 and 8). The Schwann cell activity seen is similar to that seen with dedifferentiated Schwann cells in nerve injury (24-27), which suggests that TRKB inhibition may activate intrinsic properties of Schwann cells that might regulate the cancer-and-nerve interactions associated with PNI. In addition, the cancer did not penetrate toward the center of the Schwann cell cluster as it did under the control conditions, suggesting that TRKB inhibition with ANA-12 may have some utility against perineural spread and PNI.
Our findings with oral squamous cell carcinoma are consistent with investigations in other cancer types that show Schwann cells can migrate toward cancer cells and promote cancer progression (39, 40). However, instead of Schwann cell migration as an initial event in migration and invasion assays, we show both Schwann and cancer cells migrate and grow towards each other in a similar manner. The addition of ANA-12 blocked this preferential migration of Schwann and cancer cells towards each other, possibly by preventing BDNF activation of the TRKB receptor. This is consistent with a study by Yamuachi et al., who showed how BDNF inhibits Schwann cell migration, while removal of BDNF using the BDNF scavenger TRKB-Fc enhanced Schwann cell migration and activity (45).
Upon contact though, we found cancer cells invaded Schwann cell clusters and pushed Schwann cells to the periphery, which may potentially represent a mechanism of perineural spread and PNI in oral squamous cell carcinoma. This pattern of cell-to-cell interaction seen with murine oral squamous cell carcinoma is similar to an investigation using pancreatic adenocarcinoma that showed Schwann cell contact-mediated cancer cell migration towards the nerve (22). Although we did not have neurons in our co-cultures, we did demonstrate how cancer cells grew and migrated toward the center of the Schwann cell cluster.
Unlike Deborde et al. (22), we did not show that Schwann cells induced cancer cell dispersion under the control conditions, but rather demonstrated an intriguing response from the Schwann cells when co-cultures were treated with ANA-12. With ANA-12, the passive Schwann cells appeared very active, intercalating between cancer cells and dispersing groups of cancer cells from the cancer cluster. This contradicts what is seen when pancreatic adenocarcinoma is co-cultured with dorsal root ganglia (22), which shows Schwann cell-induced cancer dispersion occurs through a NCAM1-dependent manner. Differences in these findings related to cancer cell dispersion in vitro may be related to differences in cell culture, assays, and cancer types. Further investigation into the role of Schwann cells in cancer cell dispersion and its relationship to PNI is warranted.
Although we demonstrate that BDNF and TRKB signaling regulates Schwann and cancer cell interactions that may contribute to PNI in HNC, there are several limitations to this study. Migration and dispersion assays were in vitro and performed on 2-D plates. In addition, human cell lines were not used in this pilot study. Furthermore, neurons were not included in co-cultures to avoid confounding elements in this preliminary investigation involving Schwann and cancer cells. In future studies, we hope to further uncover the mechanisms of PNI, using primary human squamous cell carcinoma and adenoid cystic cell lines, as well as 3-D in vitro and in vivo models of PNI.
Conclusion
BDNF–TRKB signaling can regulate Schwann cell activation and migration, and may contribute to cellular events involved with PNI in HNC. Our findings warrant further investigation to clearly delineate these relationships in vitro and in vivo. As we further understand the interactions between cancer cells and Schwann cells that promote PNI, we can begin to identify target-directed therapies for patients with HNC.
Footnotes
* These Authors contributed equally to this work and should be considered co-first authors.
Authors' Contributions
Ein, L: Conceived the experimental design of the study, performed experiments/data collection, data analysis and interpretation, provided revisions to scientific content, provided revisions to the article. Mei, C: Performed experiments/data collection, data analysis and interpretation, primary author of the article, provided revisions to the article. Bracho, O: Performed experiments/data collection, data analysis and interpretation, provided revisions to the article. Bas, E: Performed experiments/data collection, data analysis and interpretation, provided revisions to the article. Monje, P: Provided access to crucial research components, data interpretation, provided revisions to the article. Weed, D: Contributions to experimental design of the study, data interpretation, provided revisions to scientific content, provided stylistic/grammatical revisions to the article. Thomas, G: Contributions to experimental design of the study, data interpretation, provided revisions to scientific content, provided stylistic/grammatical revisions to the article, provided access to crucial research components. Sargi, Z: Contributions to experimental design of the study, data interpretation, provided revisions to scientific content, provided stylistic/grammatical revisions to the article. Dinh, C: Principal investigator, conceived the experimental design of the study, performed experiments/data collection, data analysis and interpretation, provided revisions to scientific content, provided stylistic/grammatical revisions to the article, provided access to crucial research components.
Conflicts of Interest
No conflicts of interest exist for any Author.
- Received August 19, 2019.
- Revision received September 23, 2019.
- Accepted September 30, 2019.
- Copyright© 2019, International Institute of Anticancer Research (Dr. George J. Delinasios), All rights reserved