Background
Nodal is a member of the TGF-β superfamily of secreted proteins that signals through the serine/threonine kinase receptors family triggering the phosphorylation of Smads 2 and 3 [
1]. Among a range of biological functions attributed to Nodal protein is its classical role during embryonic development [
1‐
3], stem cell development and differentiation [
4]. However, recent studies have shown that Nodal also regulates the maintenance of pluripotency in embryonic stem cells [
5], carcinogenesis [
6], and tumor cell progression and development [
7‐
15].
Glioblastoma (GBM; grade IV astrocytoma) is the most common primary brain tumor characterized by aggressive invasiveness, high proliferative rate, insensitivity to radio- and chemotherapy and a short survival period [
16‐
18]. It has been reported that the tumoral mass is generated by a rare fraction of cells displaying self-renew capacity named tumor-initiating cells [
19‐
22] that are involved in tumor growth and resistance to chemotherapy [
23]. Sub-population of GBM cells with stem-like properties may be the source of tumors since, apparently, these stem cells are highly resistant to current cancer treatments and survive to regenerate new tumors [
16,
24,
25].
Regardless the physiological function of Nodal has been extensively described elsewhere during tumor progression and tumorigenesis, the characterization of Nodal availability in this context has not been addressed so far. In this study we have investigated by an original approach the dynamics of Nodal intracellular distribution and extracellular availability in both stem and more differentiated GBM cells. Strikingly, we found that in GBM stem cells Nodal co-localizes in early endosomal vesicles and is abundantly available both intra and extracellularly. On the other hand, Nodal was found to co-localize to late endosomal compartments, including lysosomal vesicles, and was less available in the extracellular medium in more differentiated GBM cells. Altogether, our results propose for the first time that the Nodal availability is controlled by an endocytic pathway during GBM tumorigenesis shedding light on the molecular pathways that might emerge as putative targets for GBM therapy.
Methods
Cell culture
The cells were cultured as described in [
26]. The human glioblastoma cell line GBM011 (mdGBM) [
27] and OB1 stem cells (GBMsc) [
28] were obtained in previous studies. The human GBM U87MG cell line (mdGBM) was purchase from ATCC (Manassas, VA, USA). The experiments with human cells were regulated by the license MS, CONEP 2340. DU145 prostate cell line was purchased from the Cell Bank of Rio de Janeiro—UFRJ.
Immunocytochemistry
Permeabilization was done with 3 % Triton X-100 in PBS followed by 10 % BSA in PBS incubation. Primary anti-Nodal (1:25, rabbit, H-110, Santa Cruz Antibodies), anti-Nestin (1:200, Promega), anti-EEA1 (1:100, Santa Cruz sc5939), anti-Rab5 (1:100, Santa Cruz sc46692), Rab7 (1:100, Santa Cruz Antibodies sc6563) and Rab11 (1:100, Santa Cruz sc9020) were used followed by secondary antibody incubation (1:300, Molecular Probes) plus DAPI nuclear stain (1 μg/1 μl). The slides for Nodal immunostaining were incubated with tyramide (1:100, TSA kit #13, Life Technologies, T-20923) and were mounted with Vectashield (Vector Laboratories). Leica TCS SP5 AOBS confocal microscope was used. The images were handled in Image J. Intensity and co-localization analysis was perfomed by Leica Application Suite (LAS), Leica Microsystems and Pearson’s correlation coefficient (PCC) was used for statistic quantifying colocalization. We analyzed three different fields for each marker, where we performed the co-localization analysis in a mean of three spheroids or 20–30 cells per field. The prevalence of Nodal colocalization with different endocytic markers in each cell line was analyzed by Tukey’s test. For comparison and analysis of Nodal/endocytic markers between two cell lines, we used Sidak’s multiple comparisons test.
Western blot
RIPA buffer solution was used in the presence of protease inhibitors. The protein samples were separated by electrophoresis and blotted to PVDF membrane, followed by blocking with 5 % non-fate dry milk powder in 0.1 % PBS-Tween and incubated overnight with primary antibody for Nodal. The secondary antibody used was the same for Nodal immunofluorescence assays. The reaction was developed using SuperSignal® West Pico Chemoluminescent Substrate (Thermo Scientific) and gray scale analysis of protein bands was performed using image software. Loading control and normalization was performed through α-tubulin and actin immunobloting, and quantifications were performed in Image J.
Conditioned medium acquirement
OB1 stem cell and OB1 cells subjected to differentiation were cultivated with DMEM in the absence of Fetal Serum Bovine (FSB) for 3 days to acquire Conditioned Medium (CM). Proteins present in the CM were precipitated with 80 % acetone at −20 °C overnight and the protein extracts were processed by Western Blot assays as described above. Here, loading control and normalization was performed through densitometry of Coomassie Blue R staining of the gel. Quantifications handled on Image J.
Statistical analysis
GraphPad Prism (v6.0, La Jolla, CA) was used for ordinary one-way or two-way ANOVA analysis where appropriate. If the ANOVA produced a significant result, post hoc pair-wise comparisons were tested for significance in which the P value was adjusted (P adj < 0.05) by the Tukey’s method for multiple comparisons inside each group and by the Sidak’s method for multiple comparisons among the individual groups. Results are presented as mean ± SD and statistical relevance was defined as P < 0.05.
Discussion
We provide novel data regarding Nodal protein dynamics during GBM tumorigenesis, a process that remains poorly characterized. Besides Nodal has already been shown as a substantial engine disposed by different types of cancer, its availability and dynamics has not been addressed so far. Since we consider the understanding of basic regulation of cancer machinery a fundamental factor for its approach, our study quest a better understanding of Nodal protein in a subcellular level on GBM. Using an original approach, we analyzed the dynamics of Nodal distribution and availability in GBM cells with distinct differentiation status. Through detailed subcellular immunofluorescence analysis, we showed for the first time that Nodal is dynamically regulated during GBM cell differentiation, with clear differences found between the stem and the more differentiated cells. We were able to observe not only that the Nodal cytoplasmic distribution varied between GBMsc and mdGBM, but also to confirm that it is strongly dependent upon the differentiation status to which each cell type is induced. Even more interestingly, we showed that the availability of Nodal is regulated by different endocytic pathways during GBM tumorigenesis, being associated with distinct endosomal vesicles depending on the differentiation status of the cells.
Vesicular intracellular trafficking has important roles in cell signaling coupling processing and endocytosis in signal-receiving cells [
31,
32]. Endosomal markers are restrained at specific subcellular compartments, which dictates the stage of endosomal maturation of the vesicle [
32]. Rab5 proteins are localized in early endosomes, which play essential roles in endocytosis, signaling regulation, motility and invasion [
33‐
35]. Moreover, Rab5/EEA1 vesicles provide faster and direct recycling than through recycling endosomes (1–2 vs 15–20 min) [
36]. We found that Nodal is packed in EEA1/Rab5-positive vesicles in GBMsc. A possible consequence of the connection between Nodal and the endocytic pathway through Rab5/EEA1 vesicles is a higher rate of turnover of endocytated Nodal back to the membrane, maintaining Nodal signaling in the GBMsc itself, keeping in this way, high levels of Nodal in the extracellular media. This fact would contribute to keep cells in a more undifferentiated state. In fact, previous works have show that Nodal inhibition forces pluripotent embryonic stem cells into a differentiation pathway [
37].
On the other hand, Rab7-positive vesicles are a key factor for enzymatic regulation and internalization of membrane surface proteins [
38]. Our data show that mdGBM cells mostly presented Nodal co-localization with Rab7 and Rab11. This result indicates that these cells actively internalize Nodal and direct it for degradation or for recycling endosomes. The subsequent lower rate of Nodal addressed to the extracellular media could contribute to gradually downregulate its signaling in mdGBM cells. Altogether, our current results suggest that the endocytic pathway plays a key role for Nodal availability regulation during GBM differentiation.
We also have observed a change in Nodal intracellular distribution that was dependent upon the degree of cellular differentiation. GBMsc displayed a vesicle-like pattern of Nodal staining symmetrically distributed in the cytoplasm (Fig.
1a, b, arrow). On the other hand, mdGBM cells stained for Nodal in a more asymmetric and perinuclear fashion (Fig.
1c). We speculate that these changes directly result from the different vesicles that carry Nodal in the two cell types. EEA1/Rab5-positive vesicles would maintain Nodal closer to the membrane in GBMsc, whereas Rab7-positive vesicles would bring Nodal closer to the perinuclear region. As a consequence, GBMsc may present a higher rate of Nodal maintaining its signaling levels in the GBMsc itself.
Our immunostaining for OB1 stem cells showed that Nodal protein staining decreased as OB1 stem cells undergone differentiation (Fig.
2b, d). Our results clearly show that Nestin-positive cells are Nodal negative only when morphologically spread out, an additional indicator of differentiation. Conversely, U87MG cells forced to acquire a more undifferentiated morphology, shifted back to a symmetric pattern of Nodal cytoplasmic distribution (Additional file
4: Figure S4). These results indicate that Nodal expression heavily depends on cellular differentiation status, and we may speculate that Nodal is progressively downregulated during tumor development. In accordance, previous studies show that the TGF-β/Activin/Nodal branch is necessary to keep the pluripotent and undifferentiated state of human embryonic stem cells [
5,
39]. Altogether, our results support the notion that the cancer stem cell presents a similar behavior to the embryonic stem cell in terms of Nodal function. Despite the tight correlation between Nodal signaling and the stem cell phenotype, we still noticed the presence of Nodal in mdGBM cells. We speculate that tumor cells, although they might be in a differentiation pathway, never fully acquire a final, normal differentiated phenotype—due to the innumerous transformations they have incurred. Furthermore, the similar pattern, presented by the primary culture GBM011, to the behavior of well established cell lines suggests that Nodal protein can display an innovative marker on primary culture and biopsies analysis. However, tumor heterogeneity, especially in GBM, consists in a possible limitation for the accuracy of this approach. It would be necessary to not exclude the possibility of the subpopulation taken in the sample be precisely the more differentiated subpopulation. Thus, multiple markers would consist on a foremost solution. Still, since Nodal has been correlated to cancer aggressiveness and resistance [
6,
8,
10,
11], its verification in such samples would indicate the differentiation state and, consequently, improve prognosis.
Taken together, our results indicate for the first time that Nodal is consistently involved in GBM differentiation, since it is highly expressed in GBMsc and downregulated in mdGBM cells. Dedifferentiated GBM cells upregulate Nodal further supporting this hypothesis. This dynamics displayed by Nodal is tightly connected to the endocytic pathway in which Nodal is inserted. Our results also shed light on a new approach to evaluate GBM differentiation by analyzing Nodal protein subcellular distribution, shedding light on the molecular pathways that might emerge as putative targets for GBM therapy.
Authors’ contributions
Conceived and designed the experiments: MCON SAK LC FL VMN KC. Performed the experiments: MCON SAK ALOB. Analyzed the data: MCON SAK GV WQ JMB KC. Contributed reagents/materials/analysis tools: LGD TCLSS GV LC WQ JMB FL VMN KC. Wrote the paper: MCON SAK WQ FL VMN KC. All authors read and approved the final manuscript.