| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
From the Department of Pathology,*Internal Oncology,Neuro-Oncology and Neurology,Erasmus Medical Center, Rotterdam, The Netherlands
 |
Abstract |
|---|
 Top Abstract Materials and MethodsResultsDiscussionReferences |
|---|
helix which is absent from l-CaD. h-CaD isoforms are restricted to fully differentiated smooth muscle cells (SMCs) and regulate the smooth muscle tone. l-CaD consists of at least four splicing variants (WI-38 l-CaDs I and II, Hela l-CaDs I and II) which are expressed via differential inclusion of the variable alternative spliced exons 1, 1' and 4 of the gene.7
The exons 1 and 1' encode the short amino terminus specific for Hela l-CaDs and WI-38 l-CaDs or h-CaD, respectively.7
The l-CaD isoforms are ubiquitously distributed in various cells and dedifferentiated SMCs. They play roles in the regulation of cell contractility, adhesion-dependent signaling, and cytoskeletal organization, influencing granule movement, hormone secretion, and reorganization of microfilaments during mitosis via mitosis-specific phosphorylation by cdc2 protein kinase.10-12
The distinct functions of different cell types must involve different isoforms of caldesmon. However, the expression of the various CALD1 splicing variants and protein isoforms has only been investigated in a limited selection of normal human tissues.8
In human aorta, all splicing variants of the gene have been investigated. The expression was restricted to h-CaD (exon 1, 3b, and 4) and WI-38 l-CaD II (exon 1').8 In glioma, microvascular proliferation or hyperplasia is a notorious event. Microvascular architecture and density in low-grade gliomas are similar to that in normal brain tissue. In anaplastic gliomas and glioblastomas however, microvascular hyperplasia such as glomeruloid and branching or sprouting proliferation, is a common event. Leakage of these vessels leads to perivascular edema and shows in neuroradiologic presentations of high-grade gliomas. The proliferated or hyperplastic vessels are dysfunctional in that there is a disruption of the blood-brain barrier.
In a previous study we found the low molecular isoform of caldesmon (l-CaD) in the cerebrospinal fluid (CSF) of glioma patients.13 It was noticed by immunohistochemistry on tissue sections of the very gliomas that the expression of caldesmon was restricted to the blood vessels while no immunopositivity was obtained in glial cells. In the present study, we further investigated the CALD1 splicing variants, focusing on l-CaD in tissue samples of 68 patients with gliomas. In order to localize the caldesmon protein in the tumors, immunohistochemistry was performed on tissue sections of the gliomas of the same patients from whom the CSF samples were used. The expression of caldesmon appeared to be restricted to the blood vessels and was not seen in glial cells. In addition, any possible or minor cellular components present in the normal controls and glioma samples were pre-screened by LCM/RT-PCR and immunohistochemistry. Missplicings in glioma microvessels are revealed by RT-PCR. The transcriptional changes consequently result in an up-regulated protein expression level. Alterations of splicing patterns could result from splice-site mutations via activation of cryptic splice-site usage.16,17 The phenotypic effects of such mutations on mRNA splicing often cause codon frame-shifts or single base substitution consequently resulting in premature termination codons.18,19 Such splice-site mutations account for at least 15% of point mutations causing disease in humans.16,20 To rule out the presence of splice-site mutations resulting in the missplicing events of the CALD1 gene in our tumor cases, the samples were scanned by coupled in vitro transcription/translation assay (IVTT, also known as the protein truncation test (PTT)). The principle of IVTT is based on targeting mutations that generate truncated proteins induced by premature translation termination.21 IVTT enables to pinpoint the site of a mutation, offers good sensitivity, and a low false-positive rate.22 Further tight junction proteins (occludin and ZO-1) were co-investigated in this study. Interestingly, the up-regulated expression of l-CaD resulting from CALD1 missplicing was coincident with the down-regulation of occludin and ZO-1, causing tight junction (TJ) breakdown of the endothelial cells in glioma microvasculature.
|
Materials and Methods |
|---|
|
TopAbstractMaterials and MethodsResultsDiscussionReferences |
|---|
The study was conducted on 68 snap-frozen specimens of glioma stored in the archives of the Department of Pathology, Erasmus Medical Center, Rotterdam, the Netherlands. Histopathological typing and grading of the tumors was performed on the corresponding paraffin sections based on the latest World Health Organization classification of tumors of the central nervous system.23 Frozen section screening was to get rid of those samples with massive necrosis, hemorrhage, and contaminating normal tissues. Finally, the tumors analyzed included 26 glioblastomas, 23 oligodendrogliomas (among which 18 anaplastic oligodendrogliomas), and 19 pilocytic astrocytomas. Six samples of white matter obtained from patients without neurological or systemic disease served as controls.
Light Microscopy/Immunohistochemistry
To determine the site and distribution of the l-CaD protein expression in the tumors and control samples, tissue sections were immunohistochemically stained with a monoclonal anti-l-CaD antibody at a 1:40 dilution (BD Biosciences). The immunohistochemical procedure was described previously.13
Pre-Screening CALD1 Expression in All of the Cellular Components Possibly Present in the Tissue Samples Used by Laser-Capture Micodissection (LCM)/RT-PCR and Immunohistochemistry
The main purpose for the pre-screening experiments was to determine whether unfractional samples can be used for analysis. The white matter controls predominantly consist of glial cells and blood vessels with minor blood cell components. In glioma, the major cellular components are neoplastic glial cells (glioma cells) and hyperplastic or proliferated microvessels with possible or minor contaminating cells such as inflammatory cells, fibroblasts, and leptomeningeal cells. Normal glial cells and normal vessels were captured from the control samples (4 cases), while glioma cells and glioma vessels were captured from glioma cases (10 cases), respectively. Fibroblasts and leptomeningeal cells were captured from normal dura and arachnoid (two each from autopsy cases), respectively. Since all kinds of inflammatory cells are derived from transmigration of leukocytes from blood vessels into the brain tissue, 20 normal blood samples were used for screening of possible CALD1 expression in leukocytes. LCM was performed by using a Robot Microbeam laser microscope as the manufacturer instructed (P.A.L.M, Microlaser Technologies, Bernried, Germany). Frozen sections for LCM were prepared by using Rnase-free conditions. The used frozen tissue blocks were sectioned at 5 µm in cryostat, mounted on non-coated clean glass sliders, and stored at –80°C until use. The staining procedures of the sections were mainly based on the published protocol at http://pathbox.wustl.edu/
tisscore/protocols.htm.24
A slight modification of the protocol was made by skipping "automation buffer" and using the stainer (HisGene, Arcturus) instead of Mayer’s hemotoxylin and eosin. For Robot Microbeam laser microdissection, the tissue area of interest was selected and positioned (Figure 1, A, C, and E)
, and cut out using a focused, pulsed laser beam. The dissected areas were collected in the cap of a microcentrifuge tube containing 18 µl Trizol (Invitrogen) via laser pressure catapulting. The object of interest was catapulted off the slide using a high-energy, defocused, short-duration laser pulse (Figure 1, B, D, and F)
. The cap with the procured tissues was immediately placed on a microcentrifuge tube containing 200 µl Trizol (Invitrogen) and lysed by mixing for further RNA isolation.
|
Total cellular RNA was extracted from the selected specimens using Trizol per the manufacturer’s protocol (Invitrogen). First-strand complementary DNA (cDNA) was generated with an oligo (dT)23 primer and DuraScript reverse transcriptase (Sigma). The resulting cDNA was amplified by PCR using CALD1 specific primer sets spanning the splice sites of this gene toward all of the four splicing variants of l-CaD as described elsewhere.7 A primer set for the CALD1, designed to amplify all of the splicing variants,25 was used for initial examining the microdissected samples (the product size for l-CaD isoforms is 744 bp). Glyceraldehyde 3-phosphate dehydrogenase (GAPD) fragment26 was prepared by PCR as an internal control. All of the primers used in this study were commercially synthesized (Invitrogen). The amount of each RNA sample used was selected on the basis of identical amounts of GAPD cDNA amplified from each sample. At a 2-minute initial denaturation at 94°C, amplification conditions were as follows: 94°C for 15 seconds, 66°C (60°C for GAPD) for 30 seconds and 68°C for 1 minute for 36 cycles (40 cycles for the microdissected samples), and final extension at 68°C for 5 minutes. GAPD transcripts were amplified by 30 cycles (36 cycles for the microdissected samples). Negative controls, consisting of one sample without reverse transcriptase and one without template, were included in each experiment. The PCR products were resolved by electrophoresis on a 1% agarose gel containing ethidium bromide, and viewed under ultraviolet illumination. At least three experiments were performed for reproducibility.
Immunoblotting
Total cellular protein was co-isolated and lysed during RNA isolation by Trizol according to the manufacturer’s protocol (Invitrogen). Protein concentrations in the extracts were determined by the BCA protein assay (Pierce Chemical Co). Protein samples (15 µg/lane) were separated on a 12% SDS-polyacrylamide gel and transferred to nitrocellulose. Electroblotting was performed on a nitrocellulose membrane in 25 mmol/L Tris and 192 mmol/L glycine containing 20% methanol. The membrane was then pre-treated with 5% skim milk in Tris-buffered saline with Tween (TBST) overnight at 4°C. The membrane was incubated with monoclonal anti-l-CaD antibody (BD Biosciences, dilution 1:1000) in TBST or anti-,ß-tubulin cocktail (Lab Vision, dilution 1:1000) in TBST for 1 hour at room temperature. The membrane was further incubated with horseradish peroxidase (HRP)-conjugated goat anti-mouse IgG antibody (Zymed Laboratory Inc., dilution 1:10,000) for 1 hour at room temperature. The peroxidase was finally activated with enhanced chemiluminescence (ECL kit, Amersham) and the immunoreactivity was visualized with Kodak X-Omat AR X-ray film. For quantitative analysis of l-CaD, the bands on films were scanned by using a UMAX PowerLook II scanner (Genomic Solutions, Huntingdon, UK) at 750 dpi. Densitometric analysis was performed with the ONE-Dscan software (Scanalytics, Billerica). Each experiment was repeated at least twice. The average change in band intensities was normalized against tubulin.
IVTT
The specimens used by a coupled in vitro transcription/translation assay (IVTT) were pre-treated by puromycin (200 µg/ml) (Sigma) for 6 hours at room temperature before RNA isolation. Puromycin is a translation inhibitor known to suppress nonsense-mediated mRNA decay (NMD), thus allowing mutation screening at the RNA level.27 The PCR products suitable for IVTT analysis were generated with sense primers containing a T7 RNA polymerase promoter, a start condon, and the Kozak translation initiation sequence.28 The transcripts were amplified by 40 cycles. Since the normal transcripts containing exons 1, 1 + 4, and 1' + 4 were not detected in the control white matter, the corresponding transcripts used in IVTT were retrieved from normal stomach, prostate, and thyroid, respectively. The PCR product (2.5 µl) amplified from each transcript was incubated in the TNT/T7 non-radioactive coupled transcription and translation system (TNT Quick for PCR DNA, Promega) for 90 minutes at 30°C in a total volume of 50 µl in the presence of biotinylated lysine residues. The biotinylated lysine residues are incorporated into nascent protein during translation, eliminating the need for labeling with [35S] methionine or other radioactive amino acids, and allowing non-radioactive detection of protein synthesized in vitro by binding either streptavidin-alkaline phosphatase (streptavidin-AP) or streptavidin-horseradish peroxidase (streptavidin-HRP). In our experiments, the in vitro synthesized proteins were size-fractionated on 12% SDS-polyacrylamide mini-gels. The fractionated proteins were electroblotted onto nitrocellulose at a constant voltage of 100V for 60 minutes. Streptavidin-HRP binding (dilution 1:20,000) was used for visualization with enhanced chemiluminescence (ECL kit, Amersham) as the protocol (Promega).
Expression of Occludin and ZO-1 in Glioma Microvasculature Detected by Immunofluorescence and Immunohistochemistry
Cryostat sections (5 µm) were cut onto non-coated microscope slides (Menzel-Glaser). Slides were fixed in 100% ice-acetone at room temperature for 10 minutes and air-dried. After washes with phosphate-buffered saline (PBS), indirect immunofluorescence was carried out by using the anti-occludin polyclonal antibody (Zymed Laboratories Inc., dilution 1:100) and anti-ZO-1 polyclonal antibody (Zymed Laboratories Inc., dilution 1:100) with incubation for 1 hour at room temperature. The FITC-conjugated swine anti-rabbit IgG (Dako, dilution 1:100) was used for 1 hour at room temperature for visualization. Slides were washed with PBS, mounted in imselmount (Klinipath), and covered with Pertex and covering glass. A CCD video camera (Leica) mounted on a Leica (DMRXA) fluorescence microscope was used to capture digital images on a Leica computer (Q550 CW) running the software (Leica CW 4000 FISH Version V 1.0).
For immunohistochemistry, 5-µm paraffin sections were cut onto aminopropyltriethoxysilane (APES)-coated glass slides (Knittel Glaser). Dewaxed sections were pre-treated in 0.1% pronase (Sigma) for 10 minutes at 37°C. After the pre-treatment, the slides were washed with PBS and incubated with anti-occludin polyclonal antibody (Zymed Laboratories Inc., dilution 1:100) for 2 hours at room temperature. The biotinylated goat anti-rabbit (Dako, dilution 1:200) was used for 1 hour at room temperature for visualization. Again the slides were washed in PBS and incubated with StreptABC complex/AP (Dako, dilution 1:100) for 1 hour at room temperature. After the slides were washed with PBS, enzyme detection was performed by using a solution of Tris-HCl (pH 8.0) containing 1% new fuchsin, 1% natriumnitriet, 0.03% napthol AS-MX phosphate (Sigma), and 0.025% levamisol (Acros) for 1 hour at room temperature. The slides were washed again with PBS, mounted in imselmount (Klininpath), and covered with Pertex and covering glass for examination under a light microscope.
|
Results |
|---|
|
TopAbstractMaterials and MethodsResultsDiscussionReferences |
|---|
Immunohistochemistry showed that the expression of the caldesmon protein was restricted to the blood vessels. The immunoreactivity for the caldesmon protein was stronger in the hyperplastic microvessels of the gliomas as compared to normal brain microvasculature (Figure 2, A to E)
. The enhanced immunoreactivity was confirmed by immunoblotting analysis showing a higher expression of total l-CaD in the tumor samples (see below). At the light microscopical level no immunopositivity was observed in neoplastic and normal glial cells in the specimens. Neither the neoplastic, nor the normal, glial cells captured by LCM showed expression of CALD1 transcripts, in contrast to the positive controls (tumor vessels) (Figures 1 and 3)
.
|
The microdissected microvessels from the normal controls showed expression of exon 1'(WI-38 II) (Figure 4)
. Fibroblasts express CALD1 with restriction to exon 1' and with immunopositivity of caldesmon (Figure 5)
. Further, no CALD1 expression was found in leptomeningeal cells by LCM/RT-PCR and immunohistochemistry (Figure 6)
. No CALD1 expression in leukocytes or inflammatory cells was detected (Figure 7)
.
|
|
|
|
The results of immunohistochemistry and the LCM data revealed that there was no expression of protein and transcript of the CALD1 gene in normal and neoplastic glial cells. In addition, the pre-screening experiments show that no differential expression patterns of CALD1 were detected in any of these cell types. Thus, it is feasible to use unfractional samples of control tissue (white matter) and glioma samples. Sixty-eight glioma samples were assessed by RT-PCR. The transcript containing exon 1' (752 bp, WI-38 l-CaD II) was invariably detected in all tumors and control samples, while the transcripts containing the exons 1 (731 bp, Hela l-CaD II), 1 + 4 (670 bp, Hela l-CaD I), and 1' + 4 (691 bp, WI-38 l-CaD I) were exclusively detected in the tumors (Figure 8
and Table 1
). The transcript of the ubiquitously expressed housekeeping gene GAPD (540 bp) was uniformly amplified in all of the used samples. The splicing patterns show various frequencies and combinations of the expressed transcripts (Table 1)
. The expression pattern of l-CaD in normal brain vessels is similar to that of the human normal aorta cDNA.8
The expression of the additional transcripts in the gliomas appears to be tightly linked with the presence of microvascular hyperplasia or proliferation in the sample examined. The samples with normal-looking microvessels retained normal expression patterns of CALD1. The findings in the microdissected vessels (4 of 6 controls and 10 of 68 gliomas) further confirmed the results from these unfractional samples. To determine whether the change in l-CaD transcription would be manifested as changes in the protein quantity, immunoblotting analysis was performed on total cell lysates of the glioma and control samples. Densitometric analysis revealed an average 3.6-fold enhancement of caldesmon protein levels in the tumor samples as compared to the normal controls (Figure 9)
, indicating that the transcriptional changes were translated into altered protein levels.
|
|
|
Mutation screening was carried out on 20 glioma cases by applying the IVTT to exon 1, exon 1', and exon 4, spanning all splice sites for each transcript. No mutations leading to premature protein terminations were observed in these target exons (Figure 10)
.
|
The expression of TJ proteins (occludin and ZO-1) was detected by immunohistochemistry (IH) and immunofluorescence (IF). Occludin and ZO-1 in our study showed similar expression patterns (data from ZO-1 not shown). In longitudinally sectioned normal blood vessels, the staining patterns for occludin and ZO-1 were predominantly axial and linear, with occasional anastomoses/bifurcations (Figure 11, A1 and A2)
. In transversely sectioned vessels, however, the staining revealed short, radial or near-radial, continuous staining bands (Figure 11, B1 and B2)
. In the proliferated and/or hyperplastic glioma microvessels abnormalities in expression of occludin and ZO1 were observed. The abnormalities include reduced clarity of TJ bands (Figure 11, C1 and C2)
, interruption or discontinuity (Figure 11, D1 and D2)
, beading or spot-like pattern of the TJ staining (Figure 11E)
, or diffuse intracytoplasmic staining or absence or complete loss of immunostaining (Figure 11F)
.
|
The up-regulation of l-CaD resulting from CALD1 missplicing in hyperplastic glioma microvasculature is coincident with down-regulation of occludin (Figure 12)
and ZO-1. In our study, strong immunopositivity of l-CaD in hyperplastic glioma microvasculature was always concomitant with complete or near-complete loss of occludin and ZO-1 expression.
|
|
Discussion |
|---|
|
TopAbstractMaterials and MethodsResultsDiscussionReferences |
|---|
Caldesmon is an actomyosin-associated cytoskeletal protein residing in contractile microfilaments (MFs).10 The actin cytoskeleton in eukaryotic cells operates as a tension-sensing molecular device and has an important regulatory role on cellular contractility and adhesion-dependent signaling via caldesmon modulation.40 The contractile functions are carried out by modulating cellular tensional integrity (tensegrity) without disrupting the cellular structural integrity.41 However, alterations of CALD1 expression and the molecular mechanisms leading to such alterations in the context of neoplastic angiogenesis remain unexplored. Based on our previous findings,13 we hypothesized that the enhanced protein expression in glioma cases is regulated at the transcriptional level, not just simply reflecting post-translational modifications of the protein. Definitely, the analysis of transcriptional changes could potentially lead to the identification of novel transcription-dependent and -independent molecular contributors to the process of neoplastic neovascularization. To address this issue, we investigated the CALD1 splicing transcripts and expression level of this protein. Our data indicate that the transcriptional changes of CALD1 in splicing variants are translated into altered protein levels, and both the expression of the splicing variants and the protein are up-regulated during the process of neo-angiogenesis in gliomas. In non-muscle cells, simulation of CALD1 overexpression by transfection of full-length l-CaD results in the inhibition of cell contractility and interferes with Rho A-mediated formation of stress fibers and focal adhesions.40 The co-activated caldesmon isoforms induce an overexpression of the protein and could have a synergetic effect on the cellular contractility of the vascular components, enhancing the permeability of microvessels, and consequently facilitate the extravasation and migration of cancer cells. Moreover, both an increase in cellular calcium concentration and caldesmon phosphorylation cause dissociation of caldesmon from actin and results in weakening most of caldesmon’s properties.10 Thus, it seems that caldesmon carries out a regulatory role in endothelial cell spread and elongation, two properties related to vascular development.42-44 There is evidence from in vitro experiments that the expression of CALD1 both at the transcriptional and translational level was markedly and reproducibly down-regulated during vascular tube formation.45 Taken together, caldesmon appears to be a target in a variety of signaling pathways that modulate its function and thereby its effect on cell contractility and adhesion-dependent signaling.40
Alternative splicing of specific pre-mRNAs is controlled by cis-acting regulatory elements (exonic splicing enhancers) and trans-acting factors (SR proteins).3 The alternative exons are commonly weakened as compared to constitutive exons by having suboptimal splice sites and length but are strengthened by the presence of exon enhancers that bind to putative splicing regulators46 and/or activation of cryptic splice sites by mutation.16 We did not trace any splice-site mutations in the examined cases. It appears that an alternative mechanism may alter the interaction between an exonic splice enhancer and mRNA splicing factors of the CALD1 via a signaling pathway modifying the splicing apparatus.2,47,48 The regulation of the transcriptional activation of CALD1 splicing variants could have far-reaching epigenetic effects on the development of molecular targeted anti-angiogenic therapies.
The tight junction is the most apical located element of the junctional complex in epithelial and endothelial cells.49 The TJ between endothelial cells of brain capillaries form the structural basis of the blood-brain barrier (BBB), which controls the exchange of molecules between blood and CNS. The endothelial cells of the normal blood-brain barrier possess tight junctions that maintain a severely restricted permeability. In the hyperplastic glioma microvasculature morphological alterations such as fenestrations, the increase of caveolae, pericyte detachment, and thickening and alteration of the extracellular matrix (ECM) exist.50-52 These changes result in a dramatic increase of the permeability of these vessels. Occudin and ZO-1 are well-studied TJ proteins. In particular, occludin was the first identified transmembrane protein that is exclusively localized at TJ.53 Therefore, occludin is considered a sensitive and reliable marker of TJs.54 The inverse relationship between the regulation of l-CaD and occludin or ZO-1 found in our study has not yet been reported so far. Our data clearly reveal the up-regulation of l-CaD resulting from CALD1 missplicing in hyperplastic glioma microvessels with a concomitant loss of TJ integrity of the endothelial cells. Although the molecular basis for this correlation remains unclear, our data are indicative of a specific contribution of CALD1 expression patterns to glioma angiogenesis and generate hypotheses regarding the mechanisms contributing to the dysfunctionality of glioma neovascularization.
|
|
Acknowledgements |
|---|
|
Footnotes |
|---|
Accepted for publication February 23, 2004.
|
References |
|---|
|
TopAbstractMaterials and MethodsResultsDiscussionReferences |
|---|
mRNA in hepatoma and digestive tract cancer. Int J Cancer 1997, 73:656-662[Medline]
This article has been cited by other articles:
 |
|
 |
|
  P.-P. Zheng, L.-A. Severijnen, M. van der Weiden, R. Willemsen, and J. M. Kros A crucial role of caldesmon in vascular development in vivo Cardiovasc Res, February 1, 2009; 81(2): 362 - 369. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 L. M. Hoffman, C. C. Jensen, S. Kloeker, C.-L. A. Wang, M. Yoshigi, and M. C. Beckerle Genetic ablation of zyxin causes Mena/VASP mislocalization, increased motility, and deficits in actin remodeling J. Cell Biol., February 27, 2006; 172(5): 771 - 782. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 M. E. Cuomo, A. Knebel, G. Platt, N. Morrice, P. Cohen, and S. Mittnacht Regulation of Microfilament Organization by Kaposi Sarcoma-associated Herpes Virus-cyclin{middle dot}CDK6 Phosphorylation of Caldesmon J. Biol. Chem., October 28, 2005; 280(43): 35844 - 35858. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 P.-P. Zheng, W. C. Hop, P. A.E. Sillevis Smitt, M. J. van den Bent, C. J.J. Avezaat, T. M. Luider, and J. M. Kros Low-Molecular Weight Caldesmon as a Potential Serum Marker for Glioma Clin. Cancer Res., June 15, 2005; 11(12): 4388 - 4392. [Abstract] [Full Text] [PDF]
|
|
| |||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
