Down Regulated Expression of Claudin-1 and Claudin-5 and Up Regulation of B-Catenin: Association with Human Glioma Progression

Abstract

Glioblastoma multiforme is the most common form of intracranial malignancy in humans, and is characterized by aggressive tumor growth, tissue invasion and neurodegenerative properties. The present study investigated the expression status of tight junction associated Claudin 1 (CLDN1), Claudin 5 (CLDN5) and Adheren junction associated β-catenin genes in the light of their critical role in the progression of both low- and high-grade human gliomas. Using quantitative PCR and Western blot methods the mRNA and protein status of CLDN1, CLDN5 and β-catenin genes were studied in a total of 25 human gliomas of World Health Organization (WHO) grades I-IV, non-cancerous control brain tissues and their corresponding model cell lines (C6, U373, U118, T98 and U87MG). Quantitative analysis of the transcript and protein expression data showed that CLDN1 and CLDN5 were significantly down regulated (p=<0.001) in tumors of all four grades and model cell lines. This decrease in expression pattern was in accordance with the increasing grade of the tumor. A 4-fold stronger reduction of CLDN1 when compared to CLDN5 was evident in high-grade tumors. Interestingly, β-catenin was up regulated in all tumor types we studied (p=<0.005). Our findings, suggest that down regulated CLDN1 and CLDN5 genes have potential relevance in relation to the progression of glioblastoma multiforme. Hence, their therapeutic targeting may provide both insight and leads to control the cellular proliferation and subsequent invasiveness among affected individuals.

1. Introduction

Glioblastoma multiforme (GBM) is the most common and aggressive malignant tumor of human brain. Besides its primary occurrence within the cerebral hemispheres, GBM is occasionally reported to occur in brain stem, cerebellum and spinal cord. It accounts for 52% of all astrocytic brain tumors and 20% of all intracranial tumors. The key histopathological features exhibited by GBM include necrotizing tissue, pronounced hypercellularity, hyperplastic blood vessels and excessive vascularisation [1–3]. Aggressive tumor growth, tissue invasion and neurodegeneration are the hallmark properties of these tumors. GBM constitutes a common cause of cancer related deaths in children of 3 to 12 years and adults of 50 to 70 years old, worldwide. Despite the conventional treatments that include surgical debulking, radiotherapy and adjuvant chemotherapies for these tumors, prognostic outcomes evaluated by median survival still remains <1 year for the majority of patients. The specific molecular mechanism(s) underlying the pathogenesis of gliomas are yet to be fully elucidated. However, tracking cellular and molecular alterations may enable us to better understand disease biology and may also aid in identifying potential molecular therapeutic targets.

Neurodegeneration is considered a key hallmark of malignant glioma, along with uncontrolled cell proliferation and tissue invasion. In the past, select studies have shown that glioma cells exhibit glutamate excitotoxicity in brain tumors that induces neuronal cell death of surrounding normal brain tissues [4, 5]. Another study reported that an up regulation of astrocyte elevated gene (AEG)-1 in glioma reduced Anti-Excitatory Amino Acid Transporter 2 (EEAT2) protein expression, which leads to neurodegeneration due to excessive glutamate [6].

The Claudin (CLDN) family proteins are essential in the formation of tight junctions (TJs) in epithelial and endothelial cells. It has been observed that approximately 24 proteins are involved in the formation of TJs, which have a critical role in regulating paracellular transport and the maintenance of cell polarity. It is believed that various CLDN family members can confer different properties to epithelial cell permeability and account for the selective variability of different cellular barriers. The strength of a TJ is determined, in large part, by the combination of CLDN proteins expressed in a particular tissue. Recent observations have demonstrated that mutations within CLDN genes are responsible for various diseases, which include neonatal sclerosis cholangitis (CLDN1) [7–9], nonsyndromic recessive deafness (CLDN14) [10–12] and familial hypomagnesemia (CLDN16) [13, 14].

Recent knowledge indicates that CLDNs are able to form TJs in the absence of occludins. Upon fibroblasts transfection with CLDN1 and CLDN2, CLDN1 forms a TJ with the protoplasmic face (P-face), whereas CLDN2 forms exocytoplasmic fracture face (E- face) [15]. Expression of CLDN1 and CLDN5 has been reported within the brain [16–18], whereas CLDN11 expression has only been described in oligodendrocytes [19]. Both CLDN1 and CLDN5 are the important elements of blood-brain barrier (BBB) endothelial TJs. Earlier reports suggested that the particle association with the P-/E face in endothelial cells is believed to be a combination of CLDN1 and CLDN5 [64]. Furthermore, several studies have reported that CLDN gene expression is frequently altered in various cancers of the lung, pancreas, brain, breast and prostate [20– 27]. For example, the down regulation of CLDN1 and CLDN7 have been observed in breast cancer, prostate cancer and esophagus cancer, respectively [28– 30]. A loss of CLDN1 and CLDN5 has been described in blood vessels of GBM [31], and a significant reduction of CLDN1 expression associated with BBB changes in West Nile virus infected encephalitis [32]. Finally, down regulation of CLDN5 has been associated with BBB leakage in both mouse brain tissue with leukemia [33] and rat brain glioma [34].

Disruption of TJs causes the loss of cohesion, invasiveness and lack of differentiation which ultimately leads to the tumorigenesis in epithelial cells. The phosphorylation of CLDN family proteins has been identified as to be responsible for disruption of TJs in cancer [35]. Interestingly up regulation of CLDN genes was also observed in some cancers. For example CLDN3 and CLDN4 are over expressed in ovarian cancer [35]. Additionally, CLDN3 and CLDN4 are reported up regulated in breast, prostate and pancreatic cancers [36– 39]. In nasopharyngeal cancers, increased CLDN1 expression was found to confer cell death [40]. Immunohistochemical analysis revealed a significant association between high CLDN1 expression and basal-like breast cancer [41]. TJs are an integral component of epithelial junction complexes, which play a vital role in maintaining epithelial integrity and cell polarity. Disruption of TJs is not only a hallmark of epithelial cancer development and malignant expression but also associated with a number of pathological conditions such as kidney disorders, inflammatory bowel disease, pulmonary edema, diarrhea and jaundice [42– 45].

Adheren junctions are known to be crucial for the development and maintenance of epithelial and endothelial TJs [46– 49]. Adheren junctions of endothelial cells are composed of cadherins and catenins, both of which in conjunction with TJ proteins participate in the maintenance of paracellular barriers. The catenins include β-catenin and plakoglobin, and both of these belong to the Armadillo protein family that is characterized by multiple repeats of a 42-amino acid sequence named the Armadillo or arm repeat. These catenins mediate the linkage of transmembraneous cadherins to the cytoskeleton to form a functional adheren junction that plays an important role in intracellular signal transduction [50, 51]. To achieve these functions, β-catenin present in cytoplasm is translocated into the nucleus and generates a complex with lymphoid enhancer factor-1/T-cell factor (LEF-1/TCF) transcription factors. The tumor suppressor genes adenomatous polyposis coli (APC) and glycogen synthase kinase 3β (GSK) regulate β catenin expression and its degradation [52, 53]. In addition, catenins also act as transcriptional factors in several developmental processes in humans [54–56]. Furthermore, studies have shown that TJ opening can occur consequent to removal of extracellular Ca²⁺ from adheren junctions [57]. The abundance of β-catenin in tumor microvessels can alter TJ proteins and the morphology of undifferentiated microvessels within brain tumors that, in turn, can impact the blood brain characteristics [31].

Owing to the importance of CLDNs and adheren junction proteins in the progression of various malignant tumors, and the lack of such data on gliomas, we investigated the transcript and protein expression of CLDN1, CLDN5 and β-catenin in human gliomas of grades I–IV to evaluate whether there are molecular differences between high- and low-grade gliomas. As a cross validation these findings in human gliomas were compared with widely used immortal glioma cell culture lines. Our studies demonstrated a positive correlation between the down regulation of CLDN1 and CLDN5 and the up regulation of β-catenin in both patient samples and model cell cultures.

2. Materials and Methods

2.1 Patient samples

A total of 25 gliomas were collected from patients while they underwent neurosurgical resection of their tumors at the Neurosurgery Department of the Krishna Institute of Medical Sciences (KIMS), Hyderabad, India. Glioma tumor typing was performed by the Pathology Department, KIMS, as per the WHO classification based on histopathological characteristics [Figure 1]. Of the gliomas biopsies obtained, 2 were of Pilocytic astrocytomas (grade I), 7 were of Diffuse astrocytomas (grade II), 4 were of Anaplastic astrocytomas (grade III) and 12 were of GBM (grade IV) type. Additionally, biopsies of non-cancerous brain tissues of 3 epilepsy patients who underwent surgical resection for the treatment of medically intractable epilepsy were used as control tissues [58]. All tissues were immediately snap frozen in liquid nitrogen and stored at −80°c until they were further processed for RNA and protein isolation. Protocols followed for collecting patient samples and consent was in accordance to the approved guidelines of the KIMS institutional ethics committee.

Figure 1.

Hematoxylin and Eosin staining of low- and high-grade glioma tissues. Low grade glioma A Pilocytic astrocytoma, WHO grade I (20x) and B Diffuse astrocytoma, WHO grade II (20X), and high grade gliomas C. Anaplastic astrocytoma WHO grade III (20x) and D. Glioblastoma multiforme WHO grade IV (20x)

2.2 Glioma Cell cultures

The malignant glioma cell lines comprising C6 (low-grade), U373, U118, T98MG, U87MG (high-grade) and the control non-CNS cell line ECV304 were used in the current study, and were obtained from the cell line repository of the National Centre for Cell Science (NCCS), India, and from American Type Culture Collection (ATCC). All cell lines were maintained in Dulbecco’s modified Eagle’s media [DMEM] with supplementations of 10% heat-inactivated fetal bovine serum, 2 mM L-glutamine, 50 U/mL penicillin, and 50 μg/mL streptomycin. All were grown up to 80% confluence in 75 cm² culture flasks at 37°C in a humidified incubator supplied with 5% CO₂. Cell line sub culturing was performed for 48 hours with a cell scraper, centrifuging and resuspending in fresh DMEM. Finally, the cultured cells were scraped from culture flasks and single cell suspensions were prepared in TRIZOL reagent (Invitrogen) by passing pieces of cells through series of sequentially smaller hypodermic needles (22–30 gauge). Cells were stored at −80°c until processed further for RNA and protein isolations.

2.3 Cell viability assays

The viability of cultured cells was determined by a standard trypan blue exclusion test. In brief, 15 μl of trypan blue solution was added to 15 μl of cell suspension containing approximately 2×10⁴ cells at logarithmic growth phase. Cell viability was evaluated using a hemocytometer to aid quantify viable cell number.

2.4 RNA extractions and Real-time polymerase chain reaction analysis

Total RNA from biopsies (approximately 50 mg) or cultured cells (approximately 10⁶ cells) was isolated using TRIZOL reagent (Invitrogen) according to the manufacturer’s recommendations. After DNaseI treatment, all RNAs were reverse-transcribed into cDNAs with Superscript II reverse transcriptase (Invitrogen, USA) and oligo (dT) primers in accordance to the manufacturer’s protocol instructions. The specific mRNA quantities of CLDN1, CLDN5 and β-catenin in all cDNA samples were determined by Real-time reverse transcriptase (RT)-polymerase chain reaction (PCR) method using ABI Prism 7000 Sequence Detection System. Each 25 μl reaction mixture contained 12.5 μl of 2X Power SYBR Green PCR Master Mix (Applied Biosystems), 5 μl of cDNA, and 10 pm primer pairs (listed in Table 1). Oligonucleotide primers for CLDN1, CLDN5 and β-catenin were selected from published reports [59]. To counterbalance variations in PCR efficiency, standard curve analysis with a serially diluted pool of cDNAs was undertaken for primer sets in each reaction set up. PCR reaction conditions included 2 mins at 95°C for initial denaturing, then 40 cycles of 95°C for 20 s, 63°C for 30 s, and 72°C for 30s annealing, followed by melting analyses from 55°C to 95°C. The real time PCR reactions of CLDN1, CLDN5 and β-catenin for each sample were performed in triplicate in 96-well plates. Melting curves were checked to verify the melting temperatures of PCR amplicons. Additionally, PCR amplicons from the real-time master plate were subjected to electrophoresis on 2% agarose gel to confirm the success of the PCR reaction. GAPDH was used as an endogenous control gene. The relative expression levels for CLDN1, CLDN5 and β-catenin were calculated according to the ΔΔCt approximation method, by normalizing estimates of CLDN1, CLDN5 and β-catenin to GAPDH levels [XN=2(−Δ Ct), where ΔCt=(Ct of CLDN1, CLDN5 or β-catenin-Ct of GAPDH)]. The normalized levels of the transcripts in gliomas were then expressed in the form of 2(−ΔΔCt) [60].

Table 1.

Primer sequences

Gene	Primer sequence	Product Size
CLDN1-FP	5′-GATGAGGTGCAGAAGATGAGG-3′	200
CLDN1-RP	5′-AGAAGGCAGAGAGAAGCAGC-3′
CLDN5-FP	5′-TTCGCCAACATTGTCGTCC-5′	232
CLDN5-RP	5′-TCTTCTTGTCGTAGTCGCCG-3′
β-catenin-FP	5′-GTGCTATCTGTCTGCTCTAGTA-3′	152
β-catenin-RP	5′-CTTCCTGTTTAGTTGCAGCATC-3′
GAPDH-FP	5′-TTCGTACCTGGCATTGACTGG-3′	225
GAPDH-RP	5′-GAAGGTGAAGGTCGGAGT-3′

2.5 Preparation of soluble glioma tumor tissue lysates

All frozen tumor tissues were washed twice with ice-cold phosphate buffered saline (PBS) and scraped into Radio-Immunoprecipitation Assay (RIPA) buffer. After sonication for 2–3 min, insoluble material was separated by centrifuging at 14000g for 15 min at 4°C. The resulting supernatant was collected as a whole cell lysate and frozen at −80°C before it was used for further protein analysis. Quantification of protein concentrations in cell lysates were determined by direct UV measurement at 280 nm using a NanoVue^™ Spectrophotometer.

2.6 Western immunoblotting analysis

Tumor tissue lysates were subjected to electrophoresis on SDS-polyacrylamide gels and transferred onto nitrocellulose papers at 100V and 4°C, for 1.5 h using Towbin buffer. After blocking the nitro cellulose paper in non-fat dry milk (5%) in Tris Buffered Saline (TBS) (10mM Tris (pH 7.5), 150 mM NaCl) for 1 h at room temperature, the membrane was incubated with primary antibody CLDN1(1:100), CLDN5 (1:100) and β-catenin (1:200) overnight. The blot was thereafter incubated with a secondary antibody, i.e. goat anti-mouse IgG (dilution rate of 1:2000) conjugated to alkaline phosphatase (ALP), for 1–2 h at room temperature. Before and after incubation of blots with secondary antibodies, each blot was washed twice with TBS, 0.05% Tween 20 and once more in TBS. Immunoreactivity was determined by incubating the blots with BCIP (5-bromo-4-chloro-3-indolyl-phosphate) - NBT (nitro blue tetrazolium) chromagen solution.

2.7 Immunofluorescence Analysis

The resected tissue samples were prepared for immunofluorescence analysis by cutting them into small halves and they were subjected to subsequent fixation for 3 h at 4°C in 2% paraformaldehyde plus 0.2% glutaraldehyde solution. All the specimens were washed with PBS three times before they were sectioned by microtome (Leika) and subsequently layered onto slides. Thin tissue sections were permeabilzed using acetone and methanol mixture, which was followed by overnight primary antibody incubation. Single immunolabeling was carried out with CLDN1 (diluted 1:20 in blocking buffer BB: 1xPBS, 1% bovine serum albumin) and CLDN5 (diluted 1:50 in BB). Further sections were incubated with secondary antibody rabbit anti-mouse conjugated to FITC (1:200). Before and after incubating the tissue sections with secondary antibodies, the tissues were thoroughly washed with PBS, 0.05% Tween 20. After DAPI staining, immunofluorescence images were taken with an Immunofluorescence microscope camera (Olympus).

2.8 Statistical analysis

All continuous variable data derived from the sets of gene expression experiments are presented as mean and standard deviation values. The relative differences between the expression ratios of transcripts were compared by use of a standard two-tailed t-test using InStat Software (GraphPad, San Diego, CA). A statistically significant difference between the expression of individual transcripts is reported when a p value <0.05 was obtained.

3. Results

3.1 Down regulation of CLDN1 in glioma patient samples and glioma cell lines

In order to evaluate the CLDN1 expression status across tumors, we analyzed the transcript expression levels of CLDN1 in our human glioma samples. Figure 2 reveals that CLDN1 expression is significantly decreased in gliomas of all four grades. Specifically, higher grade tumors (e.g., grade IV GBM) showed a 4-fold reduction and grade III anaplastic astrocytomas a greater than two-fold reduction, as compared to tumors of lower grade and control samples (Table 2a). A significant down regulation of CLDN1 expression was also evident in glioma cell lines of higher grade, of more than three-fold in U373, U87MG, U118 and T98, as compared to their lower grade counterpart and the ECV 304 control cell line [Figure 5 and Table 2b]. Western blot analysis of protein lysates from low- (I and II) and high-grade (III and IV) tumor tissues was performed to evaluate correlations between transcripts and protein expression levels. Notably, Western blot analysis results similarly demonstrated a decreased expression pattern of CLDN1 protein in grade IV (GBM) and grade III (anaplastic astrocytomas) tumors versus low-grade diffuse astrocytomas and epilepsy control tissue [Figures 8a&9, and Table 2c].

Figure 2.

Transcript expression of CLDN1 in human glioma tumors grades (I–IV)

Table 2a.

CLDN1 mRNA expression levels in glioma tumor grades.

WHO Grade	Mean (SD)
Control	1.85±0.11
I	1.42±0.05
II	1.19±0.082
III	0.68±0.03
IV	0.42±0.03

Figure 5.

Transcript expression of CLDN1 in human glioma cell lines

Table 2b.

CLDN1 protein expression levels in glioma tumor grades.

WHO Grade	Mean (SD)
Control	0.8465±0.03
II	0.55±0.02
III	0.41 ±0.005
IV	0.415±0.007

Figure 8.

Western blot results of CLDN1, CLDN5 and β-catenin in glioma tumor grades of I–IV

Figure 9.

Protein expression of CLDN1 in human glioma tumors grades (I–IV)

Table 2c.

CLDN1 mRNA expression levels in glioma cell lines.

WHO Grade	Mean (SD)
ECV 304	2.05±0.11
C6	1.47±0.13
U373	0.60 ± 0.03
U87MG	0.58±0.06
U118	0.475± 0.03
T98	0.489± 0.002

Evaluation of the microvascular expression of CLDN1 in tumor tissues and non neoplastic tissues was performed by Immunofluorescence microscopy. Whereas CLDN1 expression was evident in control epilepsy tissue, its expression was not detected in high-grade tumors (GBM). In synopsis, there was concordance between the results obtained by real time PCR and Western blot analysis of glioma tumors as well as cell lines supporting a down regulated expression of CLDN1, and this decreased expression significantly associated with the high-grade of the glioma. Moreover, this result was further supported by Immunofluorescence data in which decreased expression of CLDN1 was evident in high-grade gliomas.

3.2 CLDN5 is down regulated in glioma patient samples and glioma cell lines

As illustrated in Figure 3, CLDN5 expression was significantly reduced by in excess of two-fold in grade IV GBM, two-fold in grade III anaplastic astrocytoma, one-fold in grade II diffuse astrocytoma, but not in grade I pilocytic astrocytoma, as compared to control non-neoplastic samples (Table 2d). Inter-tumor differences in CLDN5 expression revealed that m-RNA levels were down regulated by approximately 50% and 10% in grade IV, as compared to grade II and grade III gliomas, respectively. In support to this finding, experiments on glioma cell lines also indicated a down regulation of CLDN5 in higher grade glioma cell lines (i.e. U373, U118, T98 and U87MG), as compared to lower grade C6 glioma cell line and the control cell line ECV 304 [Figure 6 and Table 2e]. Results from Western blot analysis demonstrated a decline in CLDN5 protein expression patterns in high grade IV GBMs and grade III anaplastic astrocytomas versus lower grade II diffuse astrocytoma and epilepsy control tissue [Figure 8b &10, and Table 2f]. In control samples, an intense staining pattern of CLDN5 was observed at endothelial cell borders whereas in GBM microvessels it was undetectable. These results indicate that a reduction in CLDN5 expression is associated with progressively high-grade gliomas.

Figure 3.

Transcript expression of CLDN5 in human glioma tumors grades (I–IV)

Table 2d.

CLDN5 mRNA expression levels in glioma tumor grades.

WHO Grade	Mean (SD)
Control	1.61 ±0.077
I	1.44±0.04
II	1.24±0.01
III	0.85 ±0.09
IV	0.67±0.01

Figure 6.

Transcript expression of CLDN5 in human glioma cell lines

Table 2e.

CLDN5 protein expression levels in glioma tumor grades.

WHO Grade	Mean (SD)
Control	0.355±0.002
II	0.32±0.009
III	0.297±0.002
IV	0.275±0.007

Figure 10.

Transcript expression of CLDN5 in human glioma tumors grades (I–IV)

Table 2f.

CLDN5 mRNA expression levels in glioma cell lines.

WHO Grade	Mean (SD)
ECV 304	2.7
C6	1.64±0.007
U373	1.47±0.01
U87MG	1.46±0.005
U118	1.22±0.004
T98	1.242±0.006

3.3 Increased β-catenin mRNA expression in glioma patient samples and glioma cell lines

The relative transcript expression ratio of the β-catenin gene was significantly increased in relation to increased glioma grade [Figure 4]. High-grade (IV) GBM and anaplastic astrocytoma (grade III) showed an elevated β-catenin expression of more than three-fold, as compared to lower grade tumors (diffuse astrocytoma (grade II) and grade I pilocytic astrocytoma) and control tissue samples (Table 2g). β-catenin transcript analysis in glioma cell lines demonstrated a 4-fold elevated expression in high-grade tumor cell lines (U373, U118, T98 and U87MG), as compared to lower grade C6 glioma cell line and control the non-CNS ECV 304 cell line [Figure 7 and Table 2h]. Western blot analysis yielded similar results with increased β-catenin expression of in excess of three-fold in high-grade (GBM (IV) and anaplastic astrocytomas (III)), compared to grade II diffuse astrocytoma and non-neoplastic controls [Figure 8c & 11, and Table 2i].

Figure 4.

Transcript expression of Beta-catenin (or β-catenin) in human glioma tumors grades (I–IV)

Table 2g.

β-catenin mRNA expression levels in glioma tumor grades.

WHO Grade	Mean (SD)
Control	0.83±0.01
I	0.94±0.01
II	1.32±0.02
III	2.96 ±0.06
IV	3.25±0.043

Figure 7.

Transcript expression of Beta-catenin in human glioma cell lines (I–IV)

Table 2h.

β-catenin protein expression levels in glioma tumor grades.

WHO Grade	Mean (SD)
Control	0.355±0.002
II	0.32±0.009
III	0.297±0.002
IV	0.275±0.007

Figure 11.

Protein expression of Beta-catenin (or β-catenin) in human glioma tumors grades (I–IV)

Table 2i.

β catenin mRNA expression levels in glioma cell lines.

WHO Grade	Mean (SD)
ECV 304	0.67±0.02
C6	1.48±0.01
U373	2.70±0.181
U87MG	2.81±0.07
U118	2.86±0.02
T98	2.69±0.02

In synopsis, results of Western blot, real-time PCR and Immunofluorescence analyses correlated well with one another, demonstrating alike changes in both amount and direction. CLDN1 was down regulated and demonstrated marked differences between high-grade tumors (III, IV) and high grade cell lines as compared to their counterpart low-grade tumors, cell lines and controls. CLDN5 was found down regulated, but with smaller differences evident between high-grade tumors and cell lines versus lower grade ones and controls. By contrast, the expression of β-catenin progressively increased from low- to high-grade gliomas, as compared to control samples.

4. Discussion

In the present study we examined the expression pattern of the TJ proteins CLDN1, CLDN5 and adheren junction protein β-catenin in different grades of human glioma samples and from low- and high-grade human glioma cell lines. This study was undertaken to evaluate the biological significance of altered TJ protein expression in glioma progression. Several studies have defined a role for CLDN in forming TJs [61], conferring ionic selectivity [62] and functioning as a barrier [63]. In this regard, Morita and colleagues [64] demonstrated that occludin along with CLDN1 and CLDN5 are the important components of blood-brain barrier well over a decade ago. Despite this, there are relatively few published reports characterizing the expression of CLDNs in gliomas, albeit they are differentiated and often imaged in brain consequent to their altered blood-brain barrier properties that additionally can impact their response to treatment regimens [65–67].

Our data demonstrated decreased CLDN1 expression in all 12 GBM (grade IV) samples and 4 anaplastic astrocytoma (grade III) tumor samples, while showing its increased expression in 2 pilocytic astrocytoma samples (grade I) and 7 diffuse astrocytoma (grade II) tumor samples. Interestingly, the CLDN5 expression pattern was similar to the status of CLDN1 expression, as evident in high- (IV and III) and low-grade (II and I) samples. Recent studies have shown CLDNs as candidate markers for detection, prognostic evaluation and therapy of various human cancers [68]. Our studies can be compared to those of Liebner and colleagues [31], who reported the frequent loss of CLDN1 in GBM, compared to normal brain, but no such down regulation was evident for CLDN5 expression [31]. Ishihara and colleagues [69] studied 24 cases of low- and high-grade gliomas, and a loss of CLDN1 expression was observed in high-grade ones. Our results are in accord with these studies [31, 69] in relation to high-grade (IV, III) glioma patient tumor samples, and extend them in glioma cell lines demonstrating reduced CLDN1 expression levels versus low-grade pilocytic astrocytoma (I) and diffuse astrocytomas (II) tumors and control tissue. A reduction in CLDN1 expression can be attributed to the altered cellular proliferation and differentiation evident within the high-grade gliomas, potentially correlating with disease progression in patients. It may additionally relate to local proinflammatory cytokine levels [70] as reductions in CLDN1 have also been reported in hepatitis C [71]. Interestingly, the ectopic expression of CLDN-1 in mice prevents blood-brain barrier loss of permeability in the EAE model of multiple sclerosis, compared to control littermates [72].

Several studies have shown that a reduced expression of certain CLDNs have a greater impact on tumor aggressiveness than others, even though differential expression of CLDN proteins are also observed in different tissues [58]. The loss of CLDN1 in higher grade breast cancer and the unchanged expression of CLDN3 and CLDN4 further support this observation [28]. Regardless of the cellular origin, it is now widely accepted that an alteration in CLDN expression can lead to tumorigenesis. Re-expression of CLDN1 in vitro leads to apoptosis in breast cancer spheroids by another mechanism [73]. Hence, although up- or down-regulation of certain CLDNs may have great impact in altering TJs and contributing to neoplastic processes, the specific molecular mechanism via which such processes occur are not yet clearly understood, but likely involve regulation at multiple levels including microRNAs [70, 74].

As previously discussed, in prostate, breast and thyroid cancers the prognostic value of a reduced expression of CLDN1 is already reported [29, 75, and 76]. A loss of CLDN1 expression has also reported in hepatocellular malignancy [77]. Surprisingly, the up regulation of CLDN1 leads to increased cell motility and invasiveness in squamous cell carcinomas and melanomas [78, 79]. The occurrence of epithelial-mesenchymal transition, an early step during cancer progression due to the alteration of TJ proteins, is already well supported [80, 81].

Liebner and colleagues [31] have reported a strong correlation between blood-brain barrier maturation and the differential expression of β-catenin and plakoglobin. In the case of tumor microvesssels, β-catenin expression is elevated whereas plakoglobin expression is reduced. In our study, real time PCR analysis yielded similar findings that β-catenin expression was markedly increased in high-grade tumors (III, IV) and had moderate expression in pilocytic astrocytoma (I) and diffuse astrocytoma (II) tumors. In accord with this, high-grade glioma cell lines exhibited significantly increased expression of β-catenin and moderate expression in their low-grade counterpart.

Gliomas are characterized by a resistance to chemotherapy and radiotherapy despite aggressive treatment, and a high morbidity and mortality. This is particularly true in the elderly population that has the greatest incidence rate of GBM, which additionally appears to be on the rise [82, 83]. Despite several available treatments in the form of surgery, chemotherapy and radiotherapy, patient survival rates for primary malignant brain tumors remain extremely low, only 12 to 15 months following a GBM diagnosis that can fall to 4 to 5 months in elderly patients [84– 86]. Our present study suggests that glioma progression associates with dysregulation of CLDN1 and CLDN5 genes. The discovery of CLDN alterations in GBM provides a pathophysiologic link between TJ proteins and glioma progression; whether or not one can take advantage of this to find novel therapeutics is an avenue worth pursuing.

5. Conclusion

The key finding in the current study is that down regulated levels of CLDN1 and CLDN5 associated with the progression of malignant gliomas, being more pronounced with increasing glioma grade. Furthermore, adheren junction protein β-catenin was significantly elevated; proportionally increasing from low- to high-grade tumors. In accord with this data from human tumors sample resections, results from immortal tumor cell lines provided similar changes that likewise associated with glioma aggressiveness. Further studies are required to assess the molecular mechanisms underpinning these changes in expression, and whether normalization of CLDN1, CLDN5 and β-catenin protein levels in gliomas might improve their responsiveness to radio- and chemotherapy.

Figure 12.

Immunofluorescence images of tight junction proteins CLDN1 and CLDN5 in microvessels of high grade glioma, GBM, and normal brain tissue.

List of Abbreviations

AEG-1

astrocyte elevated gene-1

APC

Adenomatous polyposis coli

BBB

Blood-brain barrier

CLDN

Claudin

EEAT2

Anti-Excitatory Amino Acid Transporter 2

GBM

Glioblastoma multiforme

GSK

Glycogen synthase kinase 3β

LEF-1/TCF

Lymphoid enhancer factor-1/T-cell factor

TJ

Tight Junction