Background
Neuroblastoma is a common extracranial pediatric solid tumor, arising from the sympathetic nervous system. At the time of diagnosis, more than half of the patients already exhibit metastases in distant organs, including lymph nodes, bone marrow, bone and other organs. Even if central nervous system (CNS) involvement is rarely described in neuroblastoma and restricted to tumor relapse, its frequency is increasing and its prognosis is dismal [
1]. Efforts have been made to develop treatments for CNS metastases, but they are generally ineffective in preventing progression and subsequent death [
2]. The pattern of spread is mostly believed to occur via the hematogenous route, implying that circulating cancer cells reach the CNS after crossing the interfaces existing between the blood and the brain. Among these, the blood–brain barrier (BBB) lies in the restrictive microvascular walls, whereas the blood-cerebrospinal fluid barrier (BCSFB) is formed by the choroid plexus epithelial cells [
3]. Many studies have focused on the migration of cancer cells through the blood–brain barrier, whereas the BCSFB is underestimated as a potential route to reach the CNS. In the frame of neuroblastoma, CNS metastatic lesions are indeed not only parenchymal, but intraventricular lesions and meningeosis have also been described [
4,
5]. Our aim was to elucidate the migratory potential of neuroblastoma cells to cross the intact BCSFB. For this purpose, we studied the transmigration of two neuroblastoma cell lines—IMR32 and SH-SY5Y cells—through a well-established human in vitro BCSFB model. Our results show that these neuroblastoma cell lines could actively cross the choroid plexus epithelium within 24 h to a similar extent, and that the interactions with cancer cells did not cause a breakdown of the cellular barrier.
Discussion
The improvement of treatment strategies has allowed for remissions in a large subset of high-risk neuroblastoma patients, but in parallel, an increasing number of children will later develop focal and metastatic relapse [
7,
8]. Among the involved organs, the brain is one of the most common sites of isolated relapse in neuroblastoma patients [
1,
7], and up to 16 % of patients develop CNS metastases [
1,
9,
10]. The prognosis of these patients is still dismal despite extensive therapeutic efforts [
1,
8,
11]. Therefore, understanding how neuroblastoma cells manage to reach the CNS despite its protective barriers is of utmost importance. The blood–brain barrier (BBB), located in brain microvascular endothelial cells, is thought to be the main entry gate for cancer cells into the CNS via the hematogenous route [
12]. In neuroblastoma, patients not only develop parenchymal metastases, but also leptomeningeal and intraventricular lesions have been described [
4]. In these patients, it might well be that the tumor cells have transmigrated through the BCSFB, after having crossed the fenestrated capillaries of the choroid plexus. However, rare studies have considered the BCSFB as a potential route for cancer cells into the brain [
13], and therefore the significance of this phenomenon has not been explored until now.
Since in vitro models prove useful in deciphering cellular mechanisms involved in cancer cell-brain barrier interactions [
14,
15], our aim was to use a human in vitro BCSFB model for studying the migration potential of neuroblastoma cells through the BCSFB. As already described [
6], this human BCSFB model exhibits proper barrier features, as shown by its TEER and low permeability to LY. The present results are the first to demonstrate that two neuroblastoma cell lines (IMR-32 and SH-SY5Y) are able to cross a choroid plexus epithelial cell layer, representing the anatomical basis of the BCSFB. Transmigrating neuroblastoma cells could be detected at different stages of the process by fluorescence microscopy imaging and 3-dimensional reconstruction. No impairment in the barrier integrity could be observed after 6–24 h of contact and no focal destruction of the epithelial layer could be observed around the transmigrating neuroblastoma cells; therefore we assume an active transmigration process of neuroblastoma cells through the choroid plexus epithelial cell layer. From fluorescence microscopy imaging we cannot definitely distinguish a transcellular from a paracellular route. Although the significantly increased transmigration rate in case of BCSFB disruption by cytochalasin D argues for a paracellular way, electron microscopy imaging is warranted to gain further insights into this process. Previous studies have highlighted some alterations of endothelial biomechanical properties by cancer cells, involving actin cytoskeletal remodeling via Rho kinase signaling [
16,
17]. The described rearrangement of endothelial actin filaments can promote the trans-endothelial migration of cancer cells by impacting tight junction assembly [
17]. Similar processes might occur during the trans-epithelial migration of neuroblastoma cells herein described. Different molecules secreted by cancer cells are able to influence neighboring cells’ phenotype, in particular at the level of the cellular barriers with which they interact during the metastatic process [
18]. Among these, transforming growth factor beta-1, which is secreted by neuroblastoma cells [
19], can promote the extravasation process by inducing actin cytoskeleton reorganization [
18,
20,
21]. Recently, microRNA-181c-containing extracellular vesicles secreted by breast cancer cell lines have been shown to induce BBB disruption via actin microfilament disorganization [
22]. It cannot be excluded that neuroblastoma cells could cause cytoskeleton changes in choroid plexus epithelium by secreting such vesicles.
Brain metastases of extracranial malignancies are not restricted to parenchymal lesions. Metastases within the choroid plexus or the ventricles as well as leptomeningeal tumor spread have been described both in neuroblastoma [
4] and various other malignancies in children and adults [
23‐
25]. In these patients, the choroid plexus might have been an entry site of tumor cells to the CNS [
23], with circulating cancer cells migrating from the lumen of the fenestrated choroidal capillaries to that of ventricles. Concerning adhesion to and transmigration across endothelial cells, cancer cells share many mechanisms with leukocytes including binding to selectins, integrins, and immunoglobulin-like cell adhesion molecules to invade various organs [
18]. Since the epithelial cells of the BCSFB share the expression of many of these adhesion molecules with BBB endothelial cells [
26,
27], and these cellular interactions are—at least in part—responsible for tumor cell adhesion to and transmigration through the BBB [
28], we hypothesize that the BCSFB within the choroid plexus might represent a relevant site of tumor invasion to the CNS. However, most previous studies focused on the molecular and cellular interaction of tumor cells with the endothelial cells of the BBB [
14,
15,
29], and therefore the role of the choroid plexus epithelial cells in the development of brain metastases still needs to be addressed. In this regard, human in vitro models of the BCSFB hold out the prospect of contributing significantly to our knowledge of molecular and cellular interactions between tumor cells and choroid plexus epithelial cells.
Methods
BCSFB in vitro model
The BCSFB model is based on the cultivation of human choroid plexus papilloma cells (known as HIBCPP) on the lower surface of cell culture inserts, as already described [
6] (Fig.
1a). In brief, HIBCPP were cultured in DMEM/HAM’s F12 1:1 (Ref. 31330, Gibco) supplemented with 15 % (v/v) heat-inactivated fetal calf serum (FCS), 4 mM
l-Glutamine, 5 mg mL
−1 insulin (Ref. I9278, Sigma), 100 U mL
−1 penicillin and 100 mg mL
−1 streptomycin (HIBCPP-medium with 15 % FCS). Cells were seeded on the lower surface of cell culture inserts (ThinCert™ Cell Culture Inserts for 24-well plates, pore diameter 3.0 µm; Ref. 662631, Greiner Bio-One, Frickenhausen, Germany) that were flipped over and placed in medium-flooded 12-well plates. Cells were fed the following day, and the filters were flipped over again and placed in a 24-well plate 2 days after seeding. Cell culture was continued in HIBCPP-medium containing 1 % FCS to increase TEER. When the TEER reached at least 250 Ω cm
2, experiments could be launched.
Cultivation of neuroblastoma cell lines
Human neuroblastoma cell lines SH-SY5Y and IMR32 were purchased from the European Collection of Cell Cultures (Cat n° 94030304 and 86041809, respectively). They were cultured in vitro under standard culture conditions (37 °C, 5 % CO2 under humidified atmosphere) in RPMI 1640 medium (Ref. F1215, Biochrom AG) supplemented with 10 % FCS, 4 mM l-Glutamine (Ref. 25030, Gibco), 1 % (v/v) non-essential amino acids (Ref. 11140, Gibco), 100 U mL−1 penicillin and 100 µg.mL−1 streptomycin. Medium was renewed every 2–3 days. Before reaching confluence, cells were routinely harvested for passaging using 0.25 % trypsin—0.02 % EDTA (ethylenediaminetetraacetic acid) solution (Ref. 25200, Gibco).
Trans-epithelial electrical resistance (TEER) measurements
The TEER of HIBCPP cultures grown on inserts was measured using an epithelial tissue Volt-Ohm Meter using the STX-2 electrode system (Millicell ERS-2, Millipore, Schwalbach, Germany), before and after each transmigration experiment. The values measured for empty filters (or for filters only containing neuroblastoma cells in the case of transmigration experiment) were subtracted from the overall TEER values, in order to consider the resistance of the HIBCPP layer itself.
Transmigration assays
Transmigration experiments were performed using the same medium in the upper and lower compartments (serum-free medium containing 0.5 % (w/v) Bovine Serum Albumin, BSA).
When experiments were carried out to determine the amount of transmigrated cells after 24 h, neuroblastoma cells were harvested, dissociated and labelled with the cell-permeant fluorescent dye BCECF-AM (2′,7′-Bis-(2-Carboxyethyl)-5-(and-6)-Carboxyfluorescein, Acetoxymethyl Ester), before being added to the upper compartment of the inserts, which represents the blood compartment in this inverted BCSFB model (Fig.
1a; 200,000 neuroblastoma cells added per insert). In parallel, the same number of neuroblastoma cells was added to the upper compartments of inserts alone, to assess the migration rate in the absence of any cellular barrier. After 24 h, the fluid in the lower compartments was centrifuged (5 min at 300×
g) to pellet neuroblastoma cells located at the bottom of the wells. These cells were lysed using 1 % (w/v) Triton X-100 in PBS and their amount determined by fluorescence measurement with a Tecan 200 M Infinite Multiwell reader.
For the experiments that aimed at observing the transmigration process using immunocytochemistry (see next paragraph), neuroblastoma cells were instead loaded with Cell Tracker Green CMFDA (fixable dye) according to manufacturer’s instruction (Ref. C7025, Invitrogen). In this case, the experiment was stopped after 6–8 h.
Immunocytochemistry
At the end of the transmigration experiment, cells were fixed with 4 % formaldehyde, diluted in PBS-CMF (phosphate-buffered saline, calcium- and magnesium-free), for 10 min and rinsed twice with PBS-CMF before cutting the membranes. They were subsequently permeabilized with 0.5 % (w/v) Triton X-100 diluted in PBS-CMF containing 1 % (w/v) BSA (PBS-CMF-1 %BSA) for 1 h. After rinsing three times with PBS-CMF-1 %BSA, samples were blocked with PBS-CMF-1 % BSA for 30 min and were then incubated with the primary antibody overnight at 4 °C (rabbit anti-occludin, Ref. 71-1500, Invitrogen; dilution 1/200 in PBS-CMF-1 %BSA).
The following day, samples were rinsed three times with PBS-CMF-1 %BSA, and incubated with the secondary antibody for 1 h at RT (chicken anti-rabbit IgG 594, Ref. A21442, Invitrogen, diluted 1/500 in PBS-CMF-1 %BSA).
After two rinsing steps with PBS-CMF-1 %BSA, nuclei were stained with 4′,6-Diamidino-2-phenylindole dihydrochloride (DAPI, dilution 1.5:50000 in PBS-CMF-1 %BSA) solution and the actin cytoskeleton with phalloidin (Phalloidin Alexa fluor® 660, Invitrogen; dilution 1:250 in PBS-CMF-1 %BSA,) for 5 min. Finally, the preparations were rinsed three times with PBS-CMF and mounted under coverslips with Prolong Antifade reagent (Invitrogen).
Images were acquired and processed with Zeiss Apotome and Axiovision software (Carl Zeiss, Jena, Germany) using a 63x/1.4 objective lens. This system provides an optical slice view reconstructed from fluorescent samples (structured illumination microscopy).
Permeability of the barrier to a low-molecular-weight integrity marker
To assess the integrity of the BCSFB in vitro, its permeability to Lucifer Yellow (LY, 457 Da, Ref. L0259, Sigma-Aldrich) was measured in control filters and after each transmigration experiment. The quantity of LY having crossed the BCSFB in vitro was measured in the lower compartment after 1 h, using a microplate reader (Tecan Infinite M200 Multiwell reader, Tecan, Switzerland). For permeability coefficient (Pe) calculations, the method from Cecchelli et al. [
30] was used.
Statistical analyses
The results were expressed as mean ± SEM from three independent experiments, each performed in triplicates. Statistical significance was assessed by unpaired t-tests. A p value <0.05 was considered as significant. All statistical analyses were performed using GraphPad Prism 5.0 for Windows (GraphPad Software, San Diego, California, USA).
Authors’ contributions
EV designed and performed experiments, analyzed and interpreted the data, wrote the manuscript and acted as corresponding author. CSG carried out experiments, gave technical support and helped with fluorescence microscopy acquisition. MLL performed experiments and corrected the English. TT, HR, CS, MK, MM and HS critically revised the manuscript. HI provided HIBCPP cells. MM performed experiments. MK developed the basic concept of the article, supervised the development of the work, helped in data interpretation, evaluated and wrote parts of the article. MD developed the basic concept of the project and critically revised the manuscript. HS supervised the work, obtained funding, gave precious advices concerning data interpretation and finally approved the article. All authors read and approved the final manuscript.