1. Introduction
Glioblastoma, also designated as Isocitrate Dehydrogenase wild-type (IDH-wt) glioma in the World Health Organization (WHO) Classification of Central Nervous System (CNS) Tumors in 2021, is the most frequent malignant adult brain tumor and is classified as CNS WHO grade 4 [
1]. In addition to exhibiting rapid cellular growth, the most important characteristic of glioblastoma cells is the ability to easily infiltrate and invade the surrounding healthy brain tissue, resulting in high recurrence and death rates [
2].
Invasive cancer cells use their innate migratory capacity and form specialized protrusions termed invadopodia to invade adjacent tissues [
3]. Although the driving force for the formation of invadopodia is provided by the polymerization of actin filaments lying below the cell cortex, microtubules are needed for subsequent elongation of protrusions [
4]. The formation of these specialized protrusions, which equip tumor cells with the ability to invade, is comparable to the formation of axonal branching in neurons. Neuronal axonal branching is provided by microtubule-severing proteins that cut microtubules into small pieces at branch sites to promote arborization [
5]. Although the role of microtubule-severing proteins in neurons is well understood, their involvement in cancer cell migration and invasion has yet to be elucidated.
Microtubule-severing protein Spastin is a member of the AAA (ATPases Associated with diverse cellular Activities) protein family and cuts microtubules in an ATP-dependent manner [
6]. Spastin is classified into two main isoforms, each of which is coded from a different translational initiation site. Translation from the first ATG results in the full-length protein consisting of 616 amino acids (~68 kDa), whereas translation from the second ATG results in the short form consisting of 530 amino acids (~58 kDa) that lacks the first 86 residues containing hydrophobic region in the N-terminal of the full-length Spastin [
7]. Except for the hydrophobic region, both isoforms have identical sequence and domains: The N-terminal region of each Spastin isoform contains the microtubule-interacting and trafficking (MIT) domain, which is responsible for interaction and intracellular trafficking, while the C-terminal portion of the Spastin is the AAA domain, which catalyzes the hydrolysis of ATP required to cut microtubules. Moreover, the microtubule-binding domain (MBD) comprises amino acids from 270 to 328, which are necessary for the binding of Spastin to microtubules to cut [
8,
9,
10]. Aside from the fact that both Spastin isoforms are responsible for microtubule severing, their intracellular locations and hence cellular activities differ. While the hydrophobic region at the N-terminus of the full-length Spastin isoform (also known as M1) permits it to localize to the endoplasmic reticulum, the short isoform (known as M87) lacking this region is distributed in the cytoplasm [
11]. Since pathogenic mutations in the SPG4 gene encoding Spastin lead to the development of hereditary spastic paraplegia (HSP), earlier investigations on Spastin function have mostly focused on functions of M1 isoform which is thought to be responsible for HSP in neuronal cells [
11,
12,
13]. However, the function of M87-Spastin in dividing cells needs to be investigated further since, unlike M1-Spastin, which is prevalent in the spinal cord and expressed at a low level in other tissues and cell lines, M87-Spastin is expressed ubiquitously [
12]. Spastin has recently been linked to glioblastoma cell motility, as it was shown that the level of M87-Spastin increased in correlation with glioblastoma cell invasion capability, and Spastin was found to be co-localized with actins in migratory glioblastoma cells [
14]. Nevertheless, the molecular mechanism underpinning Spastin’s motion from microtubules to actin filaments, as well as its influence on cell migration, has yet to be uncovered.
During cell migration, serine/threonine kinase-dependent signaling pathways may transiently shift the subcellular localization of migration regulatory proteins [
15]. Pin1, a peptidyl-prolyl cis–trans isomerase enzyme, recognizes phosphorylated Ser/Thr-Pro (p-Ser/Thr-Pro) motifs in its target proteins and regulates their functions by causing changes in their structures [
16,
17,
18]. Previously, we determined that Pin1, whose expression also increases in correlation with the tumor grade of gliomas [
19], contributed to tumorigenic properties of gliomas by playing an active role in cell growth, migration, and angiogenic potential [
20]. Since Pin1 is known to alter the subcellular localization of its target proteins [
21], in this study, we investigated whether Pin1 plays a role in the transient localization change in Spastin in migrating T98G glioblastoma cells. Here, we identified Pin1 as a novel interaction partner of Spastin and showed that this interaction is required for transient co-localization of Spastin with actin filaments to enhance migration and invasion of T98G glioblastoma cells.
2. Materials and Methods
2.1. Vector Construction
The Pin1-Flag vector was a kind gift from Dr. Aslı Kumbasar. All Spastin vectors designed to be used in this study are shown schematically in
Figure 1. The Spastin
M1&M87 vector expressing both M1 and M87 isoforms due to the presence of CCACC Kozak sequence at the beginning of the sequence was constructed by cloning the full-length Spastin-coding sequence into pcDNA3.1 (+)/myc-His A vector (Addgene, Watertown, MA, USA) at HindIII/XhoI restriction sites. Spastin
M1&M87_mutA (T292A & T303A) and Spastin
M1&M87_mutD (T292D & T303D) mutants were obtained by site-directed mutagenesis in the Spastin
M1&M87 vector using a Q5 Site-Directed Mutagenesis Kit (New England Biolabs, Ipswich, MA, USA) and analyzed by Sanger sequencing. Spastin
MBD vectors (Spastin
MBD, Spastin
MBD_mutA, and Spastin
MBD_mutD) containing only the MBD domain (198-338 aa) of Spastin and M87-Spastin vectors (Spastin
M87, Spastin
M87_mutA, and Spastin
M87_mutD) expressing only the M87 isoform of Spastin (87-617 aa) were constructed by cloning the DNA fragments amplified from wild-type and mutant full-length Spastin vectors into pcDNA3.1 (+)/myc-His A vector (Addgene, Watertown, MA, USA) at HindIII/XhoI restriction sites. The primers used in polymerase chain reactions (PCR) performed for the amplification of DNA fragments or site-directed mutagenesis are listed in
Table S1.
2.2. Cell Culture, Transfection and Treatment
T98G glioblastoma cells and HEK293T embryonic kidney cells were cultured in a 37 °C CO2 incubator in Dulbecco’s modified Eagle’s medium (DMEM) (Gibco, Life Technologies, Carlsbad, CA, USA) supplemented with 10% fetal bovine serum (Gibco, Life Technologies, Carlsbad, CA, USA). For pull-down assays, HEK293T cells were transfected with 5 μg of each Spastin vector and/or 4 μg of Pin1-Flag vector using polyethylenimine (PEI) in a 1:3 ratio. Moreover, 50 ng/mL epidermal growth factor (EGF) (Gibco, Life Technologies, Carlsbad, CA, USA) was applied to T98G cells for 5 h. For immunocytochemistry (ICC) analysis, T98G cells were transfected with 1.5 μg of each Spastin vector using PEI in a 1:4 ratio. Separately, T98G cells were treated with 50 ng/mL EGF for 24 h or treated with 10 µM paclitaxel (Sigma-Aldrich, St. Louis, MO, USA) for 18 h. For real-time cell migration analysis, T98G cells were transfected with 50 nM non-targeting (NT) (GE Healthcare Dharmacon, Lafayette, CO, USA) or Spastin siRNAs (GE Healthcare Dharmacon, Lafayette, CO, USA) using DharmaFECT 2 transfection reagent (GE Healthcare Dharmacon, Lafayette, CO, USA) for 48 h. T98G cells were also transfected with 5 µg mock or each mutant M87-Spastin vector using PEI in a 1:4 ratio for real-time cell migration and invasion analysis.
2.3. Antibodies
Pin1 antibody (#sc-46660, Santa Cruz Biotechnology, Dallas, TX, USA) was used for immunoprecipitation (IP) of endogenous Pin1. The following antibodies were used for immunoblotting: Pin1 (#sc-46660, Santa Cruz Biotechnology, Dallas, TX, USA); Spastin (#PA5-44807, Invitrogen, Carlsbad, CA, USA); Myc-tag (#2276S, Cell Signaling Technology, Danvers, MA, USA); Myc-tag (#2278S, Cell Signaling Technology, Danvers, MA, USA); Flag-tag (#F7425-2MG, Sigma-Aldrich, St. Louis, MO, USA); β-actin (#8457S, Cell Signaling Technology, Danvers, MA, USA); β-actin (#3700S, Cell Signaling Technology, Danvers, MA, USA); Anti-mouse IgG VeriBlot for IP secondary antibody (ab131368, Abcam, Cambridge, UK); Anti-mouse IgG HRP-linked antibody (#7076S, Cell Signaling Technology, Danvers, MA, USA); IRDye 800CW Goat Anti-Mouse (#5257S, Cell Signaling Technology, Danvers, MA, USA); and IRDye 680RD Goat Anti-Rabbit (#5366S, Cell Signaling Technology, Danvers, MA, USA). Antibodies and dyes used in ICC experiments are as follows: Spastin (#H00006683-M02, Novus Biologicals, Centennial, CO, USA); ꞵ-tubulin (#ab6160, Abcam, Cambridge, UK); Myc-tag (#2276S, Cell Signaling Technology, Danvers, MA, USA); Alexa Fluor® 647 Conjugate Anti-rat IgG (#4418, Cell Signaling Technology, Danvers, MA, USA); Alexa Fluor® 594 Conjugate Anti-mouse IgG (#8890, Cell Signaling Technology, Danvers, MA, USA); and phallacidin (#B607, Invitrogen, Carlsbad, CA, USA).
2.4. Immunoprecipitation and Pull-Down Assays
For the endogenous Spastin–Pin1 interaction analysis, T98G cells were lysed in IP-lysis buffer (50 mM Tris-HCl, pH 7.6; 150 mM NaCl; 0.5 M EDTA, 1% NP-40, and protease inhibitors). T98G cell lysates (750 µg) were immunoprecipitated with 2 µg Pin1 antibody (#sc-46660, Santa Cruz Biotechnology, Dallas, TX, USA) and incubated overnight, at 4 °C. The samples were then rotated with DynabeadsTM Protein G (Invitrogen, Carlsbad, CA, USA), at 4 °C, for 4 h. After the incubation, the samples were placed on a magnetic platform, and the supernatant was removed. After magnetic beads were washed three times with IP-lysis buffer, the beads were boiled in sodium dodecyl sulfate (SDS)-loading buffer.
For the analysis of Spastin–Pin1 interaction based on the mutations in Pin1 recognition motifs in the MBD region of Spastin, 700 µg lysates from Spastin and/or Pin1-overexpressed HEK293T or T98G cells were incubated with 30 µL MagneHisTM Ni Particles (Promega, Madison, WI, USA), at 4 °C, for 2.5 h. After the incubation, magnetic beads were washed three times with washing/binding buffer. Then, nickel particles were resuspended in 30 µL MagneHis elution buffer. Finally, to investigate the co-precipitation of Pin1 with either endogenous or exogenous Spastin, Western blotting (WB) was used to separate precipitated protein samples and 3% input samples used as an expression control. WB was performed as described in the immunoblotting section.
2.5. Immunoblotting
Protein samples were resolved by SDS-polyacrylamide gel electrophoresis and transferred onto a nitrocellulose membrane (Santa Cruz Biotechnology, Dallas, TX, USA). Membranes were blocked with 5% non-fat dry milk in TBS-T at room temperature (RT) for 1 h. Then, the membranes were incubated with the specified primary antibodies, at 4°C, overnight, and with proper secondary antibodies, at RT, for 1 h. After the incubation with each primary and secondary antibody, the membranes were washed three times with TBS-T. Then, the visualization was performed using the ChemiDoc XRS Imaging System (BioRad, Hercules, CA, USA) or the Licor Odyssey CLx Near-Infrared Fluorescence Imaging System. Densitometric analyses were performed using ImageLab (Version 6.1), Image Studio (Version 5.2) and Adobe Photoshop CS6 software.
2.6. Immunocytochemistry
Immunocytochemistry (ICC) analysis was performed by using two alternative fixation protocols, including either paraformaldehyde (PFA) only or glutaraldehyde and PFA. For fixation with PFA protocol, T98G cells were fixed with 4% PFA at RT for 15 min. Then, cells were washed with 1X phosphate-buffered saline (PBS) for 5 min. Fixation using mixed-aldehyde protocol was carried out as follows: T98G cells were fixed with mixed-aldehyde fixation buffer [4% PFA, 0.2% glutaraldehyde, 1X PHEM (PIPES, HEPES, EGTA, MgCl2) and 0.1% Triton] at RT for 15 min. Then, cells were washed with 1X PHEM and PBS for 5 min and then permeabilized with 0.25% Triton-X-100 in 1X PBS at RT for 10 min. After washing twice with 1X PBS, cells were treated with 1% sodium borohydride in 1X PBS. After the washing steps, all samples were blocked with blocking buffer (3% (w/v) bovine serum albumin and 0.1% Triton X-100 in PBS) for 1 h at RT. Then, cells were incubated with specified primary antibodies, at 4 °C, overnight. The next day, samples were washed with 1X PBS and treated with secondary antibodies and phallacidin at RT for 1 h. Samples were then washed three times with 1X PBS. ProLong™ Diamond Antifade Mounting Medium (Invitrogen, Carlsbad, CA, USA) was added onto slides, and coverslips were placed onto samples. Cells were visualized by either TCS SP2 SE Confocal Microscope (Leica, Germany) using a 63× oil immersion objective or Leica TCS SP8 MP Confocal Microscope (Leica, Wetzlar, Hessen, Germany) using a 40× objective. All images were captured in a single focal plane using Leica Application Suite X (LAS X) Version 3.5.1 software and then processed with Adobe Photoshop CS6 Version 21.0.2 software.
2.7. Cell Migration and Invasion Analysis
XCELLigence Real-Time Cell Analysis (RTCA) DP system (Roche, Basel, Switzerland) was used to analyze cell migration and invasion. For the migration analysis, the lower chamber of a CIM-Plate16 (Agilent Technologies, Santa Clara, CA, USA ) was filled with 160 µL of DMEM (10% FBS, 1% Pen/Strep), and then the upper chamber of the dishes was placed on the lower chamber and incubated, at 37 °C, in 5% CO2 for 1 h. After the incubation period, each upper chamber was filled with 50 µL of DMEM containing 3% FBS, and it was taken to be the blank. Spastin siRNA-treated or mutant SpastinM87_mutA or SpastinM87_mutD-overexpressed T98G cells were trypsinized and resuspended in DMEM containing 3% FBS. Then, 4 × 104 cells were seeded in each well on the upper chamber, and the CIM-Plate16 was maintained in an incubator for another 30 min to allow cell attachment. The xCELLigence system automatically monitored the impedance value of each well every 15 min for 48 h, and a CI value was obtained. Before the real-time invasion analysis, the top chamber of a CIM-Plate16 was coated with 15% Matrigel Basement Membrane Matrix (Corning Incorporated, Corning, NY, USA) in a CO2 incubator for 4 h. Then, the procedures applied during the real-time migration analysis were repeated.
2.8. Statistical Analysis
GraphPad Prism 8.0.2 software (GraphPad Software Inc., CA, USA) was used for the statistical analyses. The two-tailed Student’s t-test was used to evaluate the difference between the means of two groups. One-way ANOVA was utilized for multiple group comparisons. Error bars in the graphs were generated using ± standard deviations (SD). Statistical significance was set at p < 0.05.
4. Discussion
Studies on the function of Spastin have been mostly focused on the neuronal axonal transport process, which is dysfunctional in spastic paraplegia type 4 caused by mutations in the
SPG4 gene that produces Spastin [
11,
12,
13]. Investigations in mitotic cells have not proceeded beyond examining the function of Spastin in the ER tubular network; only a study performed by Draberova and colleagues (2011) revealed that Spastin levels are increased in glioblastoma cells interrelatedly with their invasion capacity and that it is co-localized with actin in migratory T98G glioblastoma cells [
14], indicating that Spastin might also play a role in cell motility. Although this study implies that Spastin may participate actively in cell migration, there is no evidence in the literature to support this claim.
In this study, we demonstrated that Spastin plays an active role in cell migration by using real-time migration analysis of T98G glioblastoma cells in the presence and absence of Spastin (
Figure 2B), establishing for the first time a relationship between Spastin and cell migration. To reveal the molecular mechanism behind Spastin’s function in cell migration, we firstly investigated the intracellular localization of Spastin in T98G cells in which migration was stimulated or not. ICC analysis showed that the co-localization of endogenous Spastin with actin filaments at the periphery or extended regions of the T98G cells was significantly enhanced upon EGF treatment (
Figure 3A,B), which triggers the EGFR kinase activity that has been shown to accelerate angiogenesis and invasion of glioblastoma [
30,
31]. According to studies, EGFR, whose expression is known to increase in association with the severity of glioblastomas [
32], is thought to be involved in a signal network within the cell that includes 122 proteins and 211 interactions [
33]. Furthermore, Olsen and colleagues (2006) reported that EGF treatment activates EGFR, resulting in the global phosphorylation of 2244 proteins at 6600 sites, with threonine sites accounting for around 12% of these phosphorylation sites [
25]. After confirming that the expression level of Spastin was not altered by EGF treatment (
Figure 3C), we inquired to identify if it was possible that phosphorylation induces the orientation of Spastin to actin filaments. Therefore, we investigated putative phosphorylation sites in the Spastin sequence, and a bioinformatic analysis showed that Spastin has six different Pin1 recognition motifs, two of which are found in the Spastin’s MBD region (Thr292Pro and Thr303Pro) (
Figure 4A), which is primarily responsible for the microtubule binding [
26]. Therefore, we inquired to find if there would be an interaction between Spastin and Pin1, whose interaction depends on phosphorylation of the target residues. In addition, co-IP analyses have proved that Pin1 is an interaction partner of Spastin (
Figure 4B). Furthermore, we discovered that endogenous Spastin is directed to actins in T98G glioblastoma cells upon Pin1 overexpression, which is comparable to EGF treatment (
Figure 4C). Pin1 expression is transcriptionally controlled by E2 transcription factor 1 (E2F1) and has also been shown to be upregulated via E2F1 triggered by H-Ras-oncogenic signaling mediated by EGFR activation [
34]. Moreover, Pin1 expression in gliomas is known to increase in correlation with tumor grade, similar to EGFR expression [
19]. In accordance with this evidence, our results suggest that since both the phosphorylation of Spastin and the increase in Pin1 expression required for this interaction are induced by EGF treatment, the enhanced EGFR kinase activity in glioblastoma must play a significant role in the orientation of Spastin to actin filaments as a result of its interaction with Pin1.
Pull-down assays using phospho-mimetic or non-phosphorylatable mutants of the Spastin’s MBD demonstrated that phosphorylation of Pin1 recognition motifs in the MBD region mediates interaction between Spastin and Pin1 (
Figure 5B,C,E). Nevertheless, which kinases might have a role in the phosphorylation of Spastin at Pin1 recognition motifs located in its MBD region remains to be clarified. To further look into this, we searched for possible kinases that may phosphorylate Thr292 and/or Thr303 residues of Pin1 recognition motifs in the MBD of Spastin using the bioinformatic tools NetPhos 3.1 [
35], PPSP [
36], and GPS 5.0 [
37]. The data obtained by bioinformatic analysis are given in
Spreadsheet S1, and our findings led us to propose that the Pin1 recognition motifs located in the MBD region of Spastin might be phosphorylated by Ser/Thr protein kinases that are known to be triggered due to EGFR activation, such as CDKs, MAPKs, PKC, PKA, GSK3, or CKs [
25,
27]. However, further experimental investigations are needed to determine which specific kinase(s) may have roles in the phosphorylation of MBD and which Thr residues are phosphorylated.
Although we do not exclude possible interactions of Pin1 with Spastin through other putative Pin1 recognition motifs present in Spastin’s N-terminal and AAA-ATPase domains, pull-down (
Figure 5E,F) and ICC analyses (
Figure 6 and
Figure 7) using phospho-mimetic or non-phosphorylatable mutants of the full length or MBD Spastin constructs were able to show that Pin1 interaction through the MBD region of Spastin is required in determining Spastin’s re-localization from microtubules to actins. Although both Spastin
M1&M87_mutA and Spastin
M1&M87_mutD proteins were found to co-precipitate with actin (
Figure 5C), analyses of Spastin proteins containing only the MBD region proved that only Spastin
MBD_mutD co-precipitated with actin (
Figure 5E). The co-precipitation of the full length Spastin
M1&M87 or Spastin
M1&M87_mutA with the actin could be related to the interaction through another Spastin-interacting protein with actin. Pull-down experiments utilizing cell lysates do not necessarily reveal direct interaction with the bait and target proteins, but it will actually show that they are present together as members of a complex [
38]. Therefore, we decided that in situ localization analysis of the Spastin is the most accurate method to examine whether it is directed to actin filaments due to its interaction with Pin1 through its MBD region. Therewithal, ICC results revealed that only Spastin
M87_mutD is enriched at the cell periphery or in protrusions where actins are abundant, whereas Spastin
M87_mutA is distributed throughout the cell (
Figure 6C,D). Similarly, we determined that Spastin
M87, which could not be adequately phosphorylated due to its overexpression, was broadly disseminated in the cell but only could be directed to actin filaments when its phosphorylation was triggered by EGF (
Figure 6A). Altogether, this evidence proved unequivocally that Spastin’s orientation to actin filaments is dependent upon the interaction with Pin1 through Spastin’s MBD region.
Although we discovered that M87-Spastin was directed to actins owing to phosphorylation in its MBD region (
Figure 6B,C,D), we were unable to detect co-localization of actin filaments with Spastin
M1&M87_mutD, which predominantly expresses the M1-Spastin isoform as well as the M87-Spastin isoform (
Figure S3). This result indicated that while both Spastin isoforms have the potential to interact with Pin1 due to the presence of MBD in their structures, M87-Spastin could perform specific cellular functions in cell migration and invasion due to its cytoplasmic localization, as opposed to M1-Spastin, which is ER-resident.
On the other hand, ICC analyses investigating the localization of solely overexpressed M87 protein revealed that while some phospho-mimetic Spastin
M87_mutD protein was found to be oriented towards actins, the majority of the Spastin
M87_mutD protein was found to be scattered throughout the cytoplasm (
Figure 6C,D). In contrast to singly overexpressed M87, when Pin1 protein was overexpressed dually with M87 proteins, Spastin
M87_mutA protein remained scattered in the cytoplasm, while the majority of Spastin
M87_mutD protein was enriched at the cell periphery where actin filaments are abundant (
Figure 7). These results indicated that the majority of the Spastin
M87_mutD protein could not be directed to actin when there was an insufficient amount of endogenous Pin1 to interact with the exogenous phospho-mimetic M87-Spastin. Moreover, it is also understood from these results that phosphorylation of Spastin in its MBD is insufficient for its orientation to actin, and that this orientation was only achievable as a result of its interaction with Pin1 and hence its isomerization.
Additionally, we also detected that the microtubule severing activity of both non-phosphorylatable and phospho-mimetic M87-Spastin proteins persist, and their overexpressions result in loss of microtubule mass and disruption of filamentous microtubule structures. While enrichment of short non-filamentous microtubules was observed at the cell periphery where Spastin
M87_mutD was concentrated (
Figure 6C), microtubule mass was diminished without any specific enrichment regions in cells overexpressing Spastin
M87_mutA (
Figure 6C), as was detected in cells overexpressing wild-type Spastin
M87 (
Figure 6B; EGF (-)). Upon observing the presence of enriched short, non-filamentous microtubule structures at the cell periphery with Spastin
M87_mutD, we questioned if the altered localization of Spastin owing to its interaction with Pin1 was dependent on microtubule dynamics. Therefore, we assessed the impact of paclitaxel, a microtubule stabilizing agent, on the endogenous Spastin localization. ICC analyses demonstrated that when microtubules were stabilized with paclitaxel, endogenous Spastin could not co-localize with actins in the cell periphery, even if EGF was applied (
Figure 8). On the other hand, we also observed that in cells treated with paclitaxel, overexpressed Spastin
M87_mutD protein could still co-localize with actin filaments as paclitaxel treatment was not able to protect microtubule from excessive severing activity of the overexpressed Spastin (
Figure S4). These results suggest that Spastin’s co-localization with actin requires not only phosphorylation-dependent Pin1 interaction via its MBD region, but also requires dynamic microtubules and presence of severed short microtubules. Based on this evidence, we believe that Pin1 might act as a mediator of Spastin’s functions. It is tempting to suggest that phosphorylation-dependent Pin1 interaction and hence isomerization of Spastin could facilitate its interaction with motor proteins to carry the cargo microtubules and further process microtubule fragments at the cell periphery to support actin-driven protrusions required for cell migration. A recent study by Kumari and colleagues (2021) identified dynein as a novel interaction partner of Pin1. It is reported in this study that C-terminal domain of the light intermediate chain 1 subunit of dynein (LIC1-CTD) phosphorylation recruits Pin1 to the mitotic dynein complex, and LIC1-CTD phosphorylation regulates mitotic dynein function both directly and through selective Pin1 engagement with a subset of dynein complexes [
39]. Likewise, phosphorylation dependent Pin1-Spastin binding could cause conformational change in Spastin through prolyl isomerization, thus resulting in the engagement with adaptors for other functions. The possibility of Pin1-mediated regulation warrants further detailed investigation.
Real-time cell analyses indicating that T98G glioblastoma cell migration or invasion was only induced in cells where Spastin
M87_mutD protein was overexpressed (
Figure 9) imply that Spastin would only induce cell migration or invasion in its phosphorylated form in MBD and, hence, was orientated to actin filaments owing to its interaction with Pin1. Additionally, the decrease in T98G cell migration caused by the silencing of Spastin (
Figure 2) indicates that the MBD region of the endogenous Spastin is likely to be present in the phosphorylated form in T98G glioblastoma cells. Spastin, which has been reported to increase in expression in correlation with the invasion capacity of glioblastoma tumor cells, causes severe microtubule loss when overexpressed in dividing cells [
12]. Excessive microtubule-severing would normally lead to loss of microtubule cytoskeleton and therefore cell death. However, high proliferation ability [
40] and the absence of cellular death [
14] of glioblastoma cells in the presence of high levels of Spastin is possibly due to phosphorylation of Spastin because of the increased EGFR activity in glioblastoma cells. As a result of the phosphorylation of Spastin’s MBD region, its interaction with Pin1 prevents excessive microtubule severing and cell death by directing Spastin to the cell periphery. Indeed, Spastin’s orientation to the cell periphery would result in glioblastoma cells having a high migration and invasion ability, causing them to be highly aggressive tumors.
Glioblastoma treatment basically consists of surgical resection, radiation and chemotherapy. However, the most lethal characteristic feature of glioblastomas is their ability to easily invade surrounding healthy brain tissue. As a result, even if a considerable portion of the tumor mass is surgically removed during therapy, due to the high invasion capacity of glioblastoma cells, tumor cells within two centimeters of the resection margin often recur and result in mortality [
2]. Moreover, glioblastoma cells with a high migration–invasion capacity are known to be more resistant to therapeutic treatments than proliferative cells [
41]. Therefore, it is crucial to identify proteins implicated in the invasion process to improve current glioblastoma therapies. Since our real-time cell analyses clearly show that the interaction between Spastin and Pin1 promotes the migration and invasion capacities, this interaction might be a therapeutic target for glioblastoma treatment. Moreover, because this interaction is unlikely to occur in healthy glial cells due to the low expression of Spastin and Pin1 [
14], targeting this interaction may only influence cancer cells; hence, targeting this interaction may be a safe alternative for improving glioblastoma therapy. Given this, developing a bio-mimetic medication that may interfere and disrupt Spastin’s MBD interaction with Pin1 may be a critical step toward treating and preventing the recurrence of glioblastoma tumors.
In conclusion, our findings demonstrated that M87-Spastin has a role in the T98G glioblastoma cell migration and invasion processes, and this involvement is controlled by its shift from microtubules to actin filaments via its interaction with Pin1 through its MBD. Our findings imply that identifying kinases that phosphorylate Pin1 recognition motifs in MBD, will aid in the development of therapeutic drugs that may be beneficial in the treatment of cancer types with high invasion capacity such as glioblastoma.