| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
|
|
|
|||||||
|
|||||||||
| ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
Molecular & Cellular Proteomics 6:1039-1048, 2007.
© 2007 by The American Society for Biochemistry and Molecular Biology, Inc.
,
,**
From the Molecular Diagnostic Laboratory, Department of Clinical Biochemistry, Aarhus University Hospital, DK8200 Aarhus N, Denmark, ¶ Department of Medical Genetics, Biomedicum, University of Helsinki, FIN-00014 Helsinki, Finland, and || Department of Surgery, Aarhus University Hospital, DK8000 Aarhus C, Denmark
 |
ABSTRACT |
|---|
 TOP ABSTRACT EXPERIMENTAL PROCEDURESRESULTSDISCUSSIONREFERENCES |
|---|
25% of cases. Moreover we found that the overall Clusterin level in tumors often appeared to be lower than in normal mucosa due to the stromal compartment often being suppressed in tumors. Although Clusterin in normal neuroendocrine cells showed a basal localization, the localization in cancer cells was often apical and in some cases associated with apical secretion. Collectively our results indicate that Clusterin expression is very complex. We conclude that Clusterin expression is associated with neuroendocrine differentiation in normal epithelia and that the Clusterin observed in neoplastic cells is de novo synthesized. The cases with de novo synthesized Clusterin define a distinct subgroup of CRC that may be of clinical importance as anti-Clusterin therapeutics are now in clinical trials.
CLU has been reported to be a nine-exon gene encoding three protein isoforms with distinct subcellular locations and functions (1, 7, 8): a prosurvival secretory form of 80 kDa (gly co sylated, proteolytically cleaved) (9–11), a prosurvival cytoplasmic form of 60 kDa (non-gly co sylated) (3), and a proapo pto tic nuclear form of 50 kDa (non-gly co sylated) (12). It is unclear how the protein isoforms are produced from the CLU gene, and this remains a topic of debate. Current hypotheses focus on inclusion or exclusion of exon 2, which encodes an endoplasmic reticulum leader peptide. When included exon 2 targets the resulting protein for secretion (12, 13). Data from several public available database sources indicate the existence of other CLU mRNA variants, e.g. the National Center for Biotechnology Information (NCBI) reference sequence database (lists two CLU mRNA isoforms, neither has exon 2 excluded) and the Alternative Splicing Annotation Project database (predicts multiple alternative splicing CLU isoforms) (14). These other CLU transcripts could potentially be templates for the observed CLU proteins. Despite the publicly available data indicating the existence of multiple novel CLU transcripts, the task of evaluating the evidence for these transcripts and to validate their possible expression has not yet been addressed. Knowledge of all potential CLU transcripts is of utmost importance when designing oligos for silencing CLU translation using the novel antisense therapeutic strategy mentioned above.
Silencing CLU translation is of interest for the treatment of all cancers that overexpress CLU, including bladder, breast, lung, and potentially also colorectal cancer (CRC) (1). However, previous studies on CLU expression in CRC have disagreed on whether CLU is up-regulated (7, 15) or down-regulated by the cancer cells (16). There are many possible explanations for the inconsistent observations, e.g. analysis of different CLU variants caused by the use of different antibodies or by PCR assays targeting different isoforms. Another explanation could be the lack of knowledge of the cell types producing the observed CLU signal. The importance of the latter was exemplified in a recent study of prostate cancer (17) where immunohistochemistry (IHC) analysis revealed CLU to be almost undetectable in carcinoma cells despite being highly expressed in peritumoral fibroblasts, a conclusion unlikely to have been reached by applying crude sample techniques such as Western blotting or real time RT-PCR.
Another issue is the regulation of CLU expression. In neoplasia a well known mechanism of gene regulation involves genomic alterations. Genomic abnormalities of the short arm of chromosome 8, where the CLU locus is located (8p22), are commonly observed in CRC (18) and thus could potentially explain the reported CLU deregulation.
The objectives of the present study addressed the issues raised above and were as follows: 1) to identify potential CLU mRNA variants and validate their expression in colorectal tissue, 2) to investigate the expression levels of validated CLU transcripts in normal epithelia and cancer cells, 3) to resolve the controversy regarding the expression of CLU by neoplastic colorectal tissue, and (4) to investigate whether loss of genomic integrity at the CLU locus is associated with the deregulation of CLU observed in CRC.
|
EXPERIMENTAL PROCEDURES |
|---|
|
TOPABSTRACTEXPERIMENTAL PROCEDURESRESULTSDISCUSSIONREFERENCES |
|---|
Fresh frozen tissue from all 15 pairs of matched normal mucosa and adenocarcinoma were submitted to laser microdissection by which 100% pure cell subpopulations of normal epithelial and cancer cells were procured. DNA extracted from these pure cell subpopulations was used for single nucleotide polymorphism (SNP) array analysis (as published previously) (18) and for sequencing of the genomic CLU locus. RNA was extracted from five of the 15 pairs of laser-microdissected samples and used for quantitative real time RT-PCR analysis. Furthermore eight of the 15 sample pairs were also investigated by Western blot analysis. An independent set of 119 samples (17 normal mucosa and 102 adenocarcinomas) were used for transcriptional profiling as described previously (19).
Validation of CLU Transcript Variant Expression in Colorectal Tissues by RT-PCR and Sequencing of the PCR Products—
To validate the expression of predicted CLU transcript variants in colorectal tissues, cDNA was synthesized from 1 µg of total RNA from each of seven normal mucosa and two cancer samples. The oligo(dT)-initiated cDNA synthesis was performed using the Superscript II reverse transcriptase kit (Invitrogen). Variant-specific PCRs, using the cDNAs as templates, were performed using the Expand High Fidelity PCR System (Roche Applied Science). Variant-specific primer pairs were constructed by combining specific forward primers located in the unique exons 1 (1a, 1b, or 1c) with a common reverse primer located in exon 4 (see Fig. 1, A and B). Primer sequences are given in Supplemental Table 1. The resulting PCR products were excised from the agarose gel and purified using the QIAquick Gel Extraction kit (Qiagen, Valencia, CA). The excised products were then used as templates for direct bidirectional sequencing using the BigDye Terminator kit (Applied Biosystems, Foster City, CA) and an ABI 3100 Genetic Analyzer (Applied Biosystems). Sequencing was performed using the same primers as used for RT-PCR.
|
As a marker of neuroendocrine differentiation, we used a monoclonal mouse anti-human Chromogranin A antibody (LK2H10, ab8204) from Abcam. As loading control for Western blotting, we used a monoclonal mouse anti-human ß-actin antibody (catalog number A-1978, clone AC-15, Sigma-Aldrich).
Immunohistochemistry—
IHC was performed as described previously (20). In brief, paraffin was removed followed by blocking of endogenous peroxidase activity, heat-induced epitope demasking, and blocking to prevent unspecific binding. Then primary antibodies were applied at the following dilutions: CLU clone 41D, 1:600; CLU H-330, 1:100; CLU CLI-9, 1:100; CLU B-5, 1:100; and Chromogranin A (CgA), 1:10. Detection was performed using the Envision system (DakoCytomation, Glostrup, Denmark). Finally the sections were counterstained with hematoxylin.
Scoring of CLU IHC Staining—
The TMA and histological sections were reviewed independently by two observers (C. L. Andersen and K. Birkenkamp-Demtröder). The presence of cytoplasmic and secretory CLU was evaluated. To avoid misleading border staining, border areas were excluded from evaluation.
Criteria for scoring of cytoplasmic CLU were as follows. Only the epithelial cells of the normal mucosa sections were scored; likewise only the cancer cells were evaluated in the tumor sections. Sections were scored positive if immunoreactivity was seen in 10% or more of the target cells and negative if seen in less.
Criteria for scoring of secretory CLU were as follows. Only acini secretion was scored. Sections were scored positive if immunoreactivity was seen in 10% or more of the luminal acini and negative if seen in less.
Immunofluorescence Microscopy—
Immunofluorescence microscopy was performed as described previously (21). After deparaffinization, sections were incubated with a mixture of CLU H-330 rabbit polyclonal antibody (1:50 dilution) and monoclonal mouse anti-human CgA antibody (1:10 dilution). CgA is a well known marker of neuroendocrine differentiation (20). Bound primary antibodies were detected with AlexaFluor 488-conjugated goat anti-rabbit IgG highly cross-adsorbed (1:1,000; Molecular Probes Inc., Eugene, OR) and AlexaFluor 546-conjugated goat anti-mouse IgG1 (1:400; Molecular Probes Inc.). DNA was visualized by 4,6-diamidino-2-phen yl in dole (DAPI) (Sigma-Aldrich). Images were captured using a charge-coupled device camera (Quantix, Photometrics, Tucson, AZ) mounted on a Leica DMRXA epifluorescence microscope (Leica, Wetzlar, Germany) with appropriate filters and operated via SmartCapture software (Digital Scientific, Cambridge, UK).
Western Blotting—
Western blot analysis was performed as described previously (20). From eight pairs of adenocarcinoma and adjacent normal mucosa 10 µg of total protein were loaded on 12% NuPAGE gels (Invitrogen) and blotted to PVDF-plus membranes. After blocking with 5% skimmed milk, CLU protein was detected using three different specific anti-CLU antibodies at the following dilutions: Clone 41D, 1:500; CLI-9, 1:1500; and B-5, 1:200. As loading control, parallel immunoblotting using a ß-actin monoclonal antibody diluted to 0.05 µg/ml was performed.
Real Time RT-PCR—
Quantitative real time RT-PCR for the CLU mRNA variants was performed on an ABI PRISM® 7000 Sequence Detection System (Applied Biosystems) using the relevant (TaqMan or SYBR Green) Master Mix (Applied Biosystems). For normalization the gene Ubiquitin C (UBC) was used. We have demonstrated previously the suitability of UBC as a normalization gene for analysis of normal mucosa and colorectal cancer specimen sample sets (22). We applied laser microdissection to procure 100% pure normal epithelial and cancer cell populations from fresh frozen tumor and adjacent normal mucosa biopsies of five CRC patients (two stage I, two stage II, and one stage III). From each sample 10,000 cells were dissected. Total RNA was extracted from the pairs of matched normal epithelia and cancer cells using the RNeasy microkit, including on-column DNase digestion (Qiagen). RNA integrity was evaluated by Bioanalyzer analysis using Pico chips (Agilent Technologies, Palo Alto, CA). Only samples with two clear ribosomal bands were included.
The presence of the CLU mRNA variants were quantified using variant-specific primer/probe sets. The overall CLU level (the sum of all three variants) was measured using a primer/probe set targeting the boundary between exons 7 and 8, common to all variants. The sequences of primer and probes can be found in Supplemental Table 3. The specificity of each primer set was validated by Basic Local Alignment Search Tool (BLAST) searches (www.ncbi.nlm.nih.gov/BLAST/) and electronic PCR (29). Furthermore agarose gel electrophoresis confirmed the individual primer sets to generate only a single amplicon of the expected size. The normalization gene UBC was investigated using a SYBR Green assay. The primers for the UBC assay have been published previously (22).
Each measurement was performed in triplicate, and no-template controls were included for each assay. Relative expression values were obtained using a dilution curve consisting of four 10-fold serial dilution points. The dilution curve was created using a cDNA pool containing 2 µl of each of the test cDNAs.
Loss of Heterozygosity and DNA Copy Number Analysis Using Single Nucleotide Polymorphism Arrays—
Previously we have analyzed 15 pairs of matched germ line (blood) and laser-microdissected CRC samples by Affymetrix 10K XbaI SNP arrays (18). Briefly matched cancer and germ line DNA was extracted using the PUREGENE® DNA extraction system (Gentra Systems, Minneapolis, MN). The Single Primer Assay Protocol (labeling, hybridization to the 10K XbaI SNP array, washing, staining, and scanning) was performed according to the manufacturer's instructions (Affymetrix, Santa Clara, CA). Loss of heterozygosity (LOH) and DNA copy number information was extracted from the arrays as described previously (18).
Sequencing of the Genomic CLU Locus—
The integrity of the genomic CLU locus was analyzed by bidirectional sequencing using the BigDye Terminator kit (Applied Biosystems) and the ABI 3100 Genetic Analyzer (Applied Biosystems). Fragments representing each exon and the adjacent intron-exon boundaries were generated by PCR. The sense and antisense primers applied for generation of the PCR fragments were also used for sequencing. Primer details can be found in Supplemental Table 4. Importantly the splice junctions of all exons were included in the sequence analysis. Sequencing covered all 11 exons (1a, 1b, 1c, and 2–9) in Fig. 1.
Transcriptional Profiling—
The transcriptional profiling data used in the present study have been published previously by our group (19). In brief, total RNA from 17 normal mucosa samples and 102 colorectal adenocarcinomas (one Dukes A, 36 Dukes B, 64 Dukes C, and one Dukes D) were transcriptionally profiled using Affymetrix HG-U133A GeneChip arrays. The expression data were GC_RMA-normalized using ArrayAssist software (Stratagene, La Jolla, CA).
|
RESULTS |
|---|
|
TOPABSTRACTEXPERIMENTAL PROCEDURESRESULTSDISCUSSIONREFERENCES |
|---|
Using protein sequences based on the largest open reading frame, the subcellular localization of all three variants was predicted using PSORT II (23). Both CLU34 and CLU35 yield the same open reading frame starting in exon 2, right in front of an endoplasmic reticulum localization leader sequence; thus, the resulting proteins were predicted to be secreted. CLU36, on the other hand, was predicted to be localized to the nucleus.
All Three CLU mRNA Variants Are Expressed in Normal Colorectal Mucosa and Cancer—
Variant-specific RT-PCR followed by sequencing of the products confirmed the expression of all three variants in normal colorectal mucosa and cancer (Fig. 1B). We did not observe any additional bands in the three variant-specific RT-PCRs; thus we found no evidence of a skipping of exon 2 variant, which would have resulted in a second band.
Comparison of CLU mRNA Variant Expression Level in Normal Epithelial Cells and Colorectal Cancer Cells—
To investigate the expression pattern of the individual CLU variants in cancer cells relative to their normal counterparts, variant-specific quantitative real time RT-PCR was performed on matched samples of normal epithelial cells and cancer cells from five CRC patients. The analysis was performed on RNA from laser-microdissected 100% pure cell populations of normal epithelial cells and cancer cells. With the aim at hand, direct RNA isolation from the normal mucosa and tumor tissue biopsies would not have been meaningful as the cellular composition of the biopsies was very different (much more stromal tissue in the normal biopsies than in the tumor biopsies). The resulting real time RT-PCR data showed no trend in the expression of CLU34. In some patients the CLU34 expression level was higher in the cancer cells than in normal epithelia; in others it was lower or unchanged (Fig. 1C). In contrast CLU35 was significantly reduced in all cancer samples; and thus, the level of CLU35 appears to be associated with cancer development (Fig. 1C). CLU36 was expressed at levels so low in both the normal epithelial cells and cancer cells that quantification under the given conditions was futile. The overall CLU expression level (the sum of all three variants) followed a pattern very similar to that of CLU34 (Fig. 1C) indicating that the CLU34 variant constitutes the majority of the CLU transcripts found in colorectal tissue. Altogether these data suggest that the expression of the different CLU variants is regulated by different transcription factors or mechanisms and that only the expression of CLU35 seems to be uniformly associated with cancer development.
CLU Protein Expression in Normal Mucosa: Only a Small Fraction of the Epithelial Cells Express CLU—
Immunohistochemistry was used to study CLU protein expression in 17 normal colorectal mucosa specimens (15 conventional histological specimens and two TMA cores). The analysis revealed the vast majority of CLU proteins to be located in the stromal compartments and not in the epithelia (Fig. 2). Only a very limited number of epithelial cells (<1%) expressed CLU (Fig. 2, A and B). These cells were observed along the entire crypt axis with preponderance to a basal location. The subcellular localization of the CLU protein was basal. Stromal CLU protein was observed in the basal part of lamina propria (Fig. 2C) in a form that appeared to be protein aggregates. It is not clear which cells produced these aggregates. CLU protein was also found in the germinal centers of lymphoid nodules; the cells located there appeared to both express and secrete CLU protein (Fig. 2D). Also the parasympathetic ganglions of the Auerbach plexuses expressed CLU protein (Fig. 2E). Endothelial cells of vessel structures expressed CLU, and dense CLU protein aggregates were observed encircling the periphery of most arteries. Again it is not clear which cells produced these aggregates (Fig. 2F).
|
The Normal Epithelial Cells Expressing CLU Are Characterized by Neuroendocrine Differentiation—
Their low number and distribution along the crypt axis along with the subcellular localization of the CLU protein made us hypothesize that the CLU-positive epithelial cells were neuroendocrine epithelial cells. To test this hypothesis, we performed IHC with CLU and CgA, an often used marker of neuroendocrine differentiation (20), on consecutive sections of normal colon mucosa specimens from two individuals (Fig. 3, A and B). The results indicated that CLU was expressed by a subset of the CgA-positive neuroendocrine cells. If CLU is expressed by neuroendocrine cells, CLU would be expected to be expressed by cancers with neuroendocrine differentiation. Therefore, we performed CLU IHC on a colon carcinoid tumor, and indeed we found it to express large amounts of CLU (Fig. 3C) with a subcellular distribution similar to that observed in normal colon epithelial cells (Fig. 2B). To validate our hypothesis, we used immunofluorescence to co-localize CLU and CgA in normal colon mucosa specimens from two individuals. The co-localization analysis confirmed that CLU was expressed by a subset of the CgA-positive neuroendocrine cells (Fig. 3, D–G). We conclude that CLU protein is expressed only by a minor fraction (<1%) of the epithelial cells in the normal mucosa and that these cells constitute a fraction of the neuroendocrine cells.
|
|
|
|
Regulation of CLU Expression by Genomic Alterations at the CLU Locus—
A common mechanism to distort the transcriptional regulation of oncogenes and tumor suppressor genes involves alterations of their genomic loci. We therefore decided to investigate the integrity of the CLU locus in a set of colorectal cancers by LOH, DNA copy number, and sequencing analysis. Affymetrix 10K XbaI SNP arrays were used to extract genome-wide LOH and DNA copy number information from 15 pairs of matched germ line (blood) and laser-microdissected cancer samples. Analysis of the CLU locus at 8p21 (located 27.5 Mb from the p terminus of chromosome 8, according to the human March 2006 genome build) revealed 10 tumors (67%) showing simultaneous LOH and DNA copy number reduction, indicating loss of one CLU allele in these 10 tumors. To detect mutations in the remaining allele we sequenced the CLU locus in both germ line and cancer material from the 10 patients. The resulting germ line and cancer sequences were consistent with each other. No somatic mutations were identified, but the analysis revealed two SNPs, a G/C SNP in the non-coding region of exon 1b and a silent C/T SNP in exon 5. Most of the germ line samples were heterozygous for either one or both SNPs, and in all cases the matching cancer samples were homozygous, thus confirming LOH. We conclude that colorectal cancers often lose one CLU allele, whereas the integrity of the remaining allele apparently remains intact.
|
DISCUSSION |
|---|
|
TOPABSTRACTEXPERIMENTAL PROCEDURESRESULTSDISCUSSIONREFERENCES |
|---|
4,900 sequences in the CLU UniGene cluster. Moreover our experimental search for CLU mRNA variants expressed in colorectal tissues (illustrated in Fig. 1) showed no signs of a skipping of exon 2 variant. Possibly this variant is not expressed by colorectal tissue, or it could be an artifact of the MCF7 breast cancer cell line where it was discovered (12). Surprisingly sequence analysis of the CLU34 and CLU35 variants predicted them to yield one and the same secreted CLU protein. Although based on our data, we cannot determine whether the CLU34 and CLU35 variants give rise to the same protein, our real time RT-PCR data indicate that their expression levels are regulated by different factors/mechanisms and, thereby, that they have different biological functions. To explain our real time data we hypothesize that CLU34 and CLU35 are either 1) differentially expressed within the same cells or 2) expressed by different epithelial cell types. In line with the latter, it is interesting to note that disappearance of CLU35 expression coincides with disappearance of cells with neuroendocrine differentiation. Possibly the CLU35 variant is expressed predominantly by neuroendocrine cells. Sequence analysis of the CLU36 variant predicted its subcellular localization to be the nucleus. Interestingly our real time RT-PCR analysis indicated very low expression levels of this variant in both normal epithelia and cancer cells, implying no or at least very low nuclear levels of CLU protein, which was indeed what we observed by IHC.
Although some reports have indicated CLU to be up-regulated in neoplastic colorectal tissues (7, 15), others have indicated down-regulation (16). This apparent discrepancy is not unique for studies investigating CLU expression in CRC as it has also been reported for prostate cancer (1). Unawareness of the different mRNA and protein variants is, at least in part, a likely cause of the inconsistencies. Our novel knowledge of three unique transcript variants enabled us to address this problem. We performed our variant-specific real time RT-PCR on 100% pure normal epithelia and cancer cell subpopulations procured by laser microdissection. Thereby we ensured not only that we measured the variant of interest but also that we did so in the right cell population. This issue has only rarely been taken into account in previous studies on CLU expression. Interestingly only one of the variants showed a uniform dysregulated pattern in cancer cells: CLU35 was uniformly down-regulated in all five investigated cancer specimens. Based on these data, we hypothesize that either down-regulation of CLU35 is advantageous for the cancer cells, or the reduction in CLU35 is caused by the tumor-associated disappearance of the cell type expressing CLU35.
To characterize CLU expression in the normal mucosa we turned to IHC analysis. Unfortunately neither of the CLU antibodies we tested allowed us to distinguish the individual CLU protein isoforms: in Western analysis they all produced two bands of 40 and 60 kDa. However, an important finding of the IHC analysis was that the majority of CLU-expressing cells in the normal mucosa were located in the stromal compartment. CLU was observed in fibroblasts, endothelial cells, germinal center cells, and nerve cells and as protein aggregates in the basal part of lamina propria and encircling the periphery of arteries. Surprisingly we found that only a minority (<1%) of the epithelial cells expressed CLU. This finding was in contrast to a previous study on CLU expression in colorectal tissues that indicated expression of nuclear CLU in all normal epithelial cells and a shift in subcellular localization to the cytoplasm in tumors (7). Importantly our analysis included both the CLU antibodies (CLU-31D and H-330) applied in the aforementioned study. Despite this, we never observed nuclear expression in normal epithelial cells. For this discrepancy we have no explanation; however, our observations are supported by another study of CLU expression in normal and neoplastic colorectal tissue in which the authors did not observe any nuclear CLU either (15).
Our further characterization of the nature of the few CLU-positive epithelial cells led us to conclude yet another piece of novel information, namely that the CLU-positive cells constituted a subfraction of the rare neuroendocrine cells. This is very interesting as previous studies have shown CLU to be stored in secretory granules of other neuroendocrine secretory tissues and released upon stimulated exocytosis (26–28). This raises the possibility that the CLU protein we often observed in the basal part of the lamina propria was secreted by the CLU-expressing neuroendocrine cells. Many studies have indicated that CLU is a stress-related protein induced by many different stress factors including ionizing radiation, heat, cytokines, and chemotherapy (1, 2). Thus, stress could be the stimulus causing the neuroendocrine cells to release CLU into the extracellular space.
We also investigated a series of tumors by IHC. These tumors represented the full spectrum of CRC pathogenesis, including benign adenomas, adenocarcinomas, and metastases. In the majority of tumors (75%) the cancer cells did not express CLU. In the remaining 25%, we found CLU de novo synthesized (hence up-regulated) by non-neuroendocrine differentiated cancer cells. These 25% of cases may account for the few tumors that have expression values in the range of normal cells as seen in Fig. 5C. The de novo CLU synthesis was observed in tumors covering all stages of CRC pathogenesis from the earliest adenoma to advanced liver metastases. Interestingly the de novo synthesized CLU was often secreted.
Not surprisingly, the stromal compartments of the tumor specimens were often dramatically suppressed compared with normal mucosa. Hence in 75% of the tumor specimens without CLU-expressing cancer cells, the overall level of CLU appeared lower than in normal mucosa. This observation, which was confirmed both at the protein and mRNA level by Western blotting and array-based transcriptional profiling of complete biopsies, is a further explanation of the discrepancies among the previous studies on CLU expression in colorectal tissues. Collectively our data indicate that CLU expression is involved in tumorigenesis and progression of 25% of CRCs.
How CLU expression is regulated in normal colorectal epithelial cells and especially in cancer cells is unknown. However, a common mechanism to distort the transcriptional regulation of oncogenes and tumor suppressor genes involves alterations of their genomic loci. To investigate the integrity of the CLU locus, we performed LOH and DNA copy number analysis on a series of CRC specimens. Based on the CLU protein expression levels revealed by our IHC data we expected the majority of tumors to have two CLU alleles and up to 25% of tumors to show a slight amplification. Much to our surprise, we reached a completely different result, namely that the majority (67%) of tumors showed loss of one CLU allele. To investigate whether biallelic inactivation of the CLU locus was the explanation for CLU not being expressed in 75% of tumors, we next analyzed the integrity of the remaining alleles by sequencing. In all cases the integrity of the remaining allele was intact. Unfortunately our analyses did not reveal the LOH and copy number status of the CLU-expressing tumors. However, collectively our data revealed that all investigated tumors had at least one intact CLU allele and thus the potential to express CLU.
Seen from a clinical perspective the finding that CLU is de novo synthesized in 25% of CRCs is very interesting. It is well accepted that CLU up-regulation confers cytoprotective and prosurvival measures to the cancer cells and that this in turn facilitates tumor progression (1). Hence we suggest that the CRC patients with de novo CLU expression should be targets for treatment with the novel CLU antisense therapy possibly in conjunction with conventional treatment regimens. This suggestion is in line with 1) a recent review of the clinical implications of CLU expression that advocated silencing of CLU expression in all cancers that overexpress CLU (1) and 2) the multiple preclinical and clinical phase I trials that have shown CLU antisense therapy to sensitize prostate, breast, lung, bladder, and kidney cancer cells to conventional chemotherapeutics (1, 4–6). A recent CLU review has urged for caution not to target the anticancer/proapo pto tic nuclear CLU variant when using CLU antisense therapy (1). Our data indicate that this is not a major issue for the treatment of CRC as we did not find nuclear CLU expression. An important aspect of chemotherapy is to identify the patients likely to benefit from the treatment. In the present study we demonstrate that CRC patients can easily be subclassified according to CLU expression level using ordinary IHC methodology, thus pinpointing the patients likely to benefit from the antisense treatment.
In summary, we identified and validated the expression of three CLU transcript variants in colorectal tissues. Of these, the CLU35 variant was shown to be uniformly down-regulated by cancer cells. Furthermore we demonstrated that in the normal epithelia CLU protein was only expressed by rarely occurring neuroendocrine cells, whereas in 25% of tumors, CLU protein was de novo synthesized by non-neuroendocrine cancer cells. Collectively the data indicate that anti-CLU therapy may benefit as many as 25% of CRC patients.
|
ACKNOWLEDGMENTS |
|---|
|
FOOTNOTES |
|---|
Published, MCP Papers in Press, February 23, 2007, DOI 10.1074/mcp.M600261-MCP200
1 The abbreviations used are: CLU, Clusterin; CgA, Chromogranin A; CRC, colorectal cancer; DAPI, 4,6-diamidino-2-phen yl in dole; IF, immunofluorescence; IHC, immunohistochemistry; LOH, loss of heterozygosity; SNP, single nucleotide polymorphism; UBC, Ubiquitin C; TMA, tissue microarray; EST, expressed sequence tag. 
* This work was supported by funds from the Danish Research Council, the Lily Benthine Lunds Fond, the Desiree og Niels Ydes Fond, the Aase og Einar Danielsens Fond, the King Christian X's Fond, and the John and Birthe Meyer Foundation. The costs of publication of this article were defrayed in part by the payment of page charges. The article must therefore be hereby marked "advertisement" in accordance with 18 U.S.C. Section 1734 solely to indicate this fact.
S The on-line version of this article (available at http://www.mcponline.org) contains supplemental material.
To whom correspondence may be addressed: Molecular Diagnostic Laboratory, Dept. of Clinical Biochemistry, Aarhus University Hospital, Skejby, DK-8200 Aarhus N, Denmark. Tel.: 45-89495195; Fax: 45-89496018; E-mail: cla{at}ki.au.dk
** To whom correspondence may be addressed: Molecular Diagnostic Laboratory, Dept. of Clinical Biochemistry, Aarhus University Hospital, Skejby, DK8200 Aarhus N, Denmark. Tel.: 45-89495195; Fax: 45-89496018; E-mail: orntoft{at}ki.au.dk
|
REFERENCES |
|---|
|
TOPABSTRACTEXPERIMENTAL PROCEDURESRESULTSDISCUSSIONREFERENCES |
|---|
CiteULike 
Complore 
Connotea 
Del.icio.us 
Digg 
Reddit 
Technorati What's this?
This article has been cited by other articles:
|
|
 |
|
  K. Thorsen, K. D. Sorensen, A. S. Brems-Eskildsen, C. Modin, M. Gaustadnes, A.-M. K. Hein, M. Kruhoffer, S. Laurberg, M. Borre, K. Wang, et al. Alternative Splicing in Colon, Bladder, and Prostate Cancer Identified by Exon Array Analysis Mol. Cell. Proteomics, July 1, 2008; 7(7): 1214 - 1224. [Abstract] [Full Text] [PDF]
|
|
| ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
values of 0.899 (p < 0.0001) and 0.851 (p < 0.0001) for the cytoplasmic and secretory scores, respectively. The few cases showing disagreement were re-evaluated, and consensus was reached before interpretations were made. The results show a significant increase in frequency of cytoplasmic CLU expression in adenomas and adenocarcinomas (5% level, one-tailed Fisher exact test). The increase in frequency of acini secretion was not significant at the 5% level (one-tailed Fisher exact test) for either tumor category.
