Background
Accounting for 45% of all brain malignancies and 54% of all human gliomas, glioblastoma (GBM) is the most aggressive and lethal type of brain tumor [
1],[
2]. Despite multimodality therapies including maximal resection and adjuvant chemotherapy and radiotherapy, the overall outcome of patients with newly diagnosed GBM remains dismal. According to the most recent report of The Central Brain Tumor Registry of the United States (CBTRUS), less than 5% of GBM patients survive five years post diagnosis [
1]. Clearly more effective therapies are urgently needed and identification of valuable prognostic biomarkers and potential molecular targets is one key strategy to achieve this goal.
There are several different genetic alterations of important genes that may contribute to the pathogenesis of GBM, and these aberrations may differ from patient to patient. Therefore, treatment regimens for patients with GBM may be more effective if they are tailored toward the particular pathogenesis of patients' neoplasm. In recent years, substantial efforts have been made to explore molecular profiles to better understand the pathogenesis of GBM and biomarkers associated with patients' survival. There also have been several public resources that have provided insight into the pathogenesis of GBM through allowing researchers to correlate levels of gene expression with clinical features, including The Cancer Genome Atlas (TCGA) network [
3] and Repository of Molecular Brain Neoplasia Data (REMBRANDT) database [
4]. Gene expression studies of TCGA GBM tissues have identified several distinct GBM molecular subtypes, namely classical, mesenchymal, proneural and neural [
5]. Thus, uncovering new prognostic factors and molecular targets altered in GBM, and revealing the association of their expression profile with genetic alterations and molecular subtypes of GBM, may provide opportunities to improve the clinical outcome of GBM patients.
Tectonic family member 1 (TCTN1), was first identified in 2006 as a potential regulator of the Hedgehog pathway in patterning of the neural tube of mice, downstream of smoothened and Rab23, and named tectonic after the Greek word for builder due to its apparent involvement in a diverse range of developmental processes [
6]. In addition, a recent study showed that TCTN1 was part of a ciliopathy-associated protein complex and interacted with several other proteins associated with ciliopathies [
7]. Over the past several years, the primary cilium was found to be a complex signalling center where Hedgehog signalling was regulated [
8]-[
10], and its disregulation was associated closely to tumorigenesis [
11]-[
13]. Furthermore, Hedgehog pathway was involved in the regulation of embryonic development, cancer formation and maintenance, cancer stem cells [
14]-[
16], and particularly development and progression of human gliomas [
17],[
18]. However, the function and prognostic value of TCTN1 in human glioma have never been characterized.
In this study, we sought to investigate levels of TCTN1 expression in human GBMs using a tissue microarray (TMA) of a Chinese GBM cohort and estimate its prognostic value. We then validated the differential expression and prognostic significance of TCTN1 in another two independent datasets, namely the TCGA cohort and the REMBRANDT cohort. For the TCGA cohort, we also analysed the expression profile of TCTN1 according to subtypes and genetic alterations of GBM. Finally, we performed cell proliferation assays to explore the functions of TCTN1 in GBM cells.
Discussion
Glioblastoma (GBM) is the most malignant brain tumor with dismal prognosis despite multimodal therapies, and its pathogenesis is still far from elucidation. Molecular targeted therapy represents promising avenue for the future of effective treatment strategies for GBMs. Hence, more valuable prognostic biomarkers and potential molecular targets for gliomas are urgently needed to combat this devastating disease. The present study identified TCTN1 as a novel prognostic factor for GBM, which was overexpressed in GBM tissues and could also regulate the proliferation of GBM cells.
TCTN1 was a newly identified gene reported to be involved in developmental processes, Hedgehog pathway transduction and functions of primary cilium [
6],[
7]. Given that potent regulators of developmental processes are frequently disrupted in tumorigenesis [
44], and the primary cilium and Hedgehog pathway also play important roles in tumorigenesis [
11],[
16], it is to be expected that TCTN1 also contributes to tumor development yet there have been no reports on it. Hence, our study aimed to unveil the indispensable role of TCTN1 in GBM progression. Our TMA analysis and real-time PCR validation of a Chinese GBM cohort revealed that TCTN1 was up-regulated in GBMs compared to normal controls, and high TCTN1 expression could predict shorter overall and progression-free survival for GBM patients, as an independent prognostic factor. Due to differences of genetic background between populations [
45], we further validated these findings in another two independent international cohorts, namely the TCGA cohort and the REMBRANDT cohort.
It was noteworthy that our immunohistochemical staining experiments in GBM tissues found a nuclear localisation of TCTN1, which was beyond our expectation more or less, given the two important reports that linked TCTN1 to Hedgehog pathway and primary cilia by Dr. Jeremy F. Reiter' group [
6],[
7]. Actually, there were several limitations of these studies. The former revealed the involvement of Tctn1 (the mouse homolog of human TCTN1) in Hedgehog signaling mediated patterning of the neural tube of mice. Epistasis analyses further indicated that Tctn1 modulated Hedgehog signal transduction downstream of Ptch, Smo and Rab23. However, the findings were merely restricted in a mouse embryonic development context and lacked direct evidences using biochemical methods. The latter report found that Tctn1 was essential for ciliogenesis in some embryonic tissues such as the node and neural tube, and was required to localize some proteins to the cilium in several other tissues containing perineural and limb bud mesenchyme. They further discovered Tctn1 as part of a transition zone complex that controlled the organization of the transition zone and ciliary membrane composition using some mouse cell lines. However, the mechanisms underlying the tissue specificity of Tctn1 complex function remain unclear and the findings were also context dependent. Thus, whether TCTN1 regulate Hedgehog pathway still remains unclear, particularly in the context of human cancer biology.
Hedgehog signaling pathway was linked to tumorigenesis in recent years. The most typical examples were basal cell carcinoma (BCC) [
46] and medulloblastoma (MB) [
47], in which mutations were identified in the regulatory components of Hedgehog pathway. Although there were a few reports regarding the regulation of Hedgehog signaling on cancer stem cells in human gliomas [
48],[
49], the role of Hedgehog pathway in glioma remains in question.
To further investigate the relationship of
TCTN1 and Hedgehog pathway, we examined the transcriptional level of
GLI1, which is widely used to reflect Hedgehog pathway activity [
50], in TCGA database, and found it comparable between GBMs and normal controls (data not shown). In addition, we analyzed the relationship of
TCTN1 and two common target genes of Hedgehog pathway,
GLI1 and
PTCH1, and found no significant correlation (Additional file
1: Figure S3), indicating a rather weak link (if any) between TCTN1 and Hedgehog pathway in GBMs.
The signal transduction of Hedgehog pathway was regulated in the primary cilium, where TCTN1 was found to be a component of a protein complex. In the mammalian body, primary cilia were found on most epithelial and stromal cells, and interestingly, transformed cells commonly lack cilia [
51]. The role of primary cilia in cancer progression were still controversial, maybe according to the genetic background, as found in BCC [
13] and MB [
12]. In addition, the prevalence and role of cilia on glioma cells were poorly studied. It was reported that primary cilia were deficient in several established GBM cell lines compared to normal astrocytes [
52]. Consistently, in recently derived primary GBM cell lines and tumor biopsies, the majority of cells were unable to grow cilia [
53]. Furthermore, it seems that the observed cilia of a small portion of U251 GBM cells had no effect on cell proliferation, since depletion of Kif3a, a key component of ciliogenesis, did not significantly affect cell growth [
54].
A remarkable feature of ciliogenesis is its cell cycle-dependence [
51],[
55]-[
57]. In a system to study ciliary dynamics in the hTERT-RPE1 cell lines, most of the cells were ciliated following serum starvation [
56], which was widely used to induce ciliogenesis in cultured cells [
54],[
58],[
59]. However, it is a remarkable fact that ciliogenesis was enhanced by serum starvation in neither established nor recently derived primary GBM cell lines [
52],[
53], although that was the case in normal primary astrocytes [
52]. Recently, it was reported that a cell-cycle-related kinase (CCRK) may modulate ciliogenesis, and its regulation of cell cycle was dependent on cilia in NIH3T3 cells [
54]. In addition, they found that depletion of CCRK could restore cilia for a small fraction of U251 glioma cells, and inhibit cell growth in part dependent on cilia. However, it is interesting to note that depletion of CCRK did not block cells in G0/G1 phase, suggesting other underlying mechanisms.
Our immunohistochemical staining experiments showed primary expression of TCTN1 in cell nucleus through a scan of more than one hundred GBM patients, suggesting a weak link (if any) of TCTN1 and cilia in human gliomas. Functions and molecular mechanisms of TCTN1 in glioma warrant more investigations.
Characterized by dramatic molecular and histologic heterogeneity, GBM has recently been classified into distinct subtypes with clinical relevance, opening the way for treatments to be directed at subtype-specific mechanisms [
5],[
60]. In addition, for each molecular subtype, genetic alterations in several key genes were significantly different. The TCGA dataset offers an opportunity to investigate the relationship between gene expression, molecular subtypes and genetic alterations [
61]-[
64]. Therefore, we studied the expression preference of
TCTN1 in different subtypes and its association with genetic aberrations in the TCGA cohort. We found that
TCTN1 was dramatically decreased in the proneural subtype compared to other three subtypes, which is in concordance with the previous finding that the proneural subtype has a trend toward longer survival compared with other subtypes [
5]. For common genetic alterations of GBM,
TCTN1 was expressed in correlation with 10 of them, i.e. mutations of
TP53,
IDH1 and
ATRX, amplifications of
EGFR,
PDGFRA and
MYCN, deletions of
CDKN2A,
CDKN2B,
PTEN and
PARK2. Interestingly, for several of them (
TP53 mutation,
EGFR amplification,
PTEN deletion and
PARK2 deletion), the association was restricted in non-proneural or proneural subtype. For instance, within non-proneural subgroups, the status of
PTEN deletion did not influence the levels of
TCTN1 expression. However, within the proneural subtype, patients with no CNA of
PTEN had dramatically lower
TCTN1 expression compared to
PTEN deleted patients. These findings provided a clue for further research of the regulation of
TCTN1 expression in GBMs.
We also investigated the relationship of
TCTN1 expression and patients' clinical outcome stratified by different molecular subtypes and status of key genetic alterations. As a consequence, when we looked at
TCTN1 impact on survival based on molecular subtype, only the proneural and mesenchymal subtype retained significance. This analysis showed that the influence of
TCTN1 expression on survival outcome shows high subtype specificity with very strong effect in the proneural and mesenchymal subtypes and almost no effect in the other subtypes, thus the full sample analysis effectively showed a dilution of the effect in these two subtypes. In particular, patients within the proneural subtype are expected to have a slightly better prognosis compared to other subtypes [
5]. However, we noted that within the proneural subgroup patients with high
TCTN1 expression suffer from especially poor prognosis than those with low
TCTN1 expression. Moreover, we also investigated status of genetic alterations in TCGA dataset and stratified the patients with GBM into two subgroups by these molecular features. Our results showed that the effect of
TCTN1 expression on patients' survival rely on genetic background. It should be noted that,
TCTN1 could divide patients with no
PTEN copy number change into two subsets with totally distinct outcome, although there was no difference for survival of
PTEN deleted patients with different
TCTN1 expression, suggesting distinct effect of
TCTN1 on clinical outcome dependent on status of
PTEN deletion. Similar results could also be observed for other several alterations, in detail, high expression of
TCTN1 could predict poor prognosis for patients with no
EGFR change, no
PDGFRA change, no
MYCN change,
PARK2 deletion,
CDKN2A deletion or
CDKN2B deletion. However, further perspective studies are still warranted to unveil the underlying mechanisms.
Our analyses in these independent cohorts suggested a key role of TCTN1 gene in tumorigenesis and progression of GBM, yet there has been no direct report on its function in cancer biology. Thus we performed in vitro experiments in two GBM cell lines through enforced expression or depletion of TCTN1. Consequently, we observed that TCTN1 overexpression evidently promoted cell proliferation, whereas TCTN1 depletion dramatically hampered cell growth. These results were consistent with the augmented expression and prognostic value of TCTN1 in GBM clinical tissues, suggesting that its survival detriment role may be in part due to the ability of the TCTN1 protein to regulate proliferation of GBM cells. Functional study in cell lines further highlighted potential therapeutic value of TCTN1 in treatment of patients with GBM, albeit the molecular mechanisms were still far from elucidation.
Competing interests
The authors declare that they have no competing interests.
Authors' contributions
DM designed the study, performed data analysis, carried out experiments and drafted the manuscript. DM and YC revised the manuscript. YZ, SY and HC assisted with statistical analyses. YC, JW and DY participated in experimental studies. JC contributed to patient collection and clinical data interpretation. DL conceived of the study, participated in its design and coordination, and revised the manuscript critically. All authors read and approved the final manuscript.