Skip Navigation


Carcinogenesis Advance Access originally published online on March 26, 2007
Carcinogenesis 2007 28(9):2028-2035; doi:10.1093/carcin/bgm066
This Article
Right arrow Abstract Freely available
Right arrow FREE Full Text (PDF) Freely available
Right arrow Supplementary Data
Right arrow All Versions of this Article:
28/9/2028    most recent
bgm066v1
Right arrow Alert me when this article is cited
Right arrow Alert me if a correction is posted
Services
Right arrow Email this article to a friend
Right arrow Similar articles in this journal
Right arrow Similar articles in ISI Web of Science
Right arrow Similar articles in PubMed
Right arrow Alert me to new issues of the journal
Right arrow Add to My Personal Archive
Right arrow Download to citation manager
Right arrow Search for citing articles in:
ISI Web of Science (3)
Right arrowRequest Permissions
Google Scholar
Right arrow Articles by Chandramouli, A.
Right arrow Articles by Nelson, M. A.
Right arrow Search for Related Content
PubMed
Right arrow PubMed Citation
Right arrow Articles by Chandramouli, A.
Right arrow Articles by Nelson, M. A.
Social Bookmarking
Add to CiteULike   Add to Connotea   Add to Del.icio.us  
What's this?

© The Author 2007. Published by Oxford University Press. All rights reserved. For Permissions, please email: journals.permissions@oxfordjournals.org

Haploinsufficiency of the cdc2l gene contributes to skin cancer development in mice

Anupama Chandramouli{dagger}, Jiaqi Shi{dagger}, Yongmei Feng, Hana Holubec1, Renée M.Shanas, Achyut K. Bhattacharyya, Wenxin Zheng and Mark A. Nelson*

Department of Pathology, Arizona Cancer Center, University of Arizona, 1501 North Campbell Avenue, LSN 550, Tucson, AZ, USA
1 Department of Cell Biology and Anatomy, College of Medicine, University of Arizona, Tucson, AZ, USA

* To whom correspondence should be addressed. Tel: +520 626 2619; Fax: +520 626 8864; Email: mnelson{at}azcc.arizona.edu


   Abstract
Top
 Abstract
Introduction
 Materials and methods
 Results
 Discussion
 Supplementary material
 References
 
The Cdc2L gene encodes for the cyclin-dependent kinase 11 (CDK11) protein. Loss of one allele of Cdc2L and reduced CDK11 expression has been observed in several cancers, implicating its association with carcinogenesis. To directly investigate the role of CDK11 in carcinogenesis, we first generated cdc2l haploinsufficient mice by gene trap technology and then studied the susceptibility of these gene-trapped (cdc2lGT) mice to chemical-mediated skin carcinogenesis in the 7,12-dimethylbenz[a]anthracene (DMBA)/12-O-tetradecanoylphorbol-13-acetate (TPA)-induced two-stage skin carcinogenesis model. Wild-type and cdc2lGT mice were subjected to a single topical application of initiation by DMBA and promotion twice a week for 19 weeks with TPA. At 19 weeks, 70% of the cdc2lGT mice and 60% of the cdc2l+/+ mice developed benign papillomas. However, there was an overall 3-fold increase in the average number of tumors per mouse observed in cdc2lGT mice as compared with cdc2l+/+ mice. There was also an increased frequency of larger papillomas in cdc2lGT mice. By using the polymerase chain reaction–restriction fragment length polymorphism assay, we found A to T transversion mutations at the 61st codon of H-ras gene in the papilloma tissue of both cdc2lGT mice and cdc2l+/+ mice. Ki-67 staining revealed increased proliferation in the papillomas of cdc2lGT (77.75%) as compared with cdc2l+/+ (30.84%) tumors. These studies are the first to show that loss of one allele of cdc2l gene, encoding CDK11, facilitates DMBA/TPA-induced skin carcinogenesis in vivo.

Abbreviations: DMBA, 7,12-dimethylbenz[a]anthracene; ES, embryonic stem; PCR, polymerase chain reaction; PI, proliferative index; RT, reverse transcription; TPA, 12-O-tetradecanoylphorbol-13-acetate


   Introduction
Top
 Abstract
 Introduction
Materials and methods
 Results
 Discussion
 Supplementary material
 References
 
Cyclin-dependent kinase 11 (CDK11) (formerly known as PITSLRE) protein is a member of the extended family of p34cdc2-related kinases. It is becoming increasingly clear that in addition to controlling the cell cycle, cyclin-dependent kinases have other functions within the cell. Recent studies indicate that the p110 isoform of CDK11 (CDK11p110) may be involved in RNA processing or transcription by virtue of the fact that CDK11 co-immunoprecipitates and/or co-purifies with multiple transcriptional elongation factors including ELL1, TFIIF, TFIIS and FACT (1,2). Cyclin L1 and, more recently, cyclin L2 are regulatory partners of the CDK11 isoforms (3). Cyclin L is an arginine/ serine domain protein that may function in pre-mRNA splicing (3). The CDK11p110 localizes to the nucleoplasm and nuclear speckle-staining regions (i.e. splicing factor centers) (1,2,4). In addition, CDK11p110 also interacts with the general pre-mRNA splicing factors RNPS1 and 9G8 (both of which are classified as splice-site recognition proteins), RNA polymerase II and casein kinase 2 (1,5,6).

The major CDK11p110 isoform is encoded by two human genes Cdc2L1 and Cdc2L2. These genes also contain the open reading frame for a smaller isoform, CDK11p58, which is generated by an internal ribosome entry site sequence found in most species. CDK11p58 is a mitosis-specific isoform, which exclusively functions in the G2 and M phases of the cell cycle. By means of RNA interference, it has been demonstrated that CDK11p58, but not the CDK11p110, isoform is required for mitotic spindle formation (7).

In regards to apoptosis, increased expression of CDK11 reduces cell growth due to apoptosis (8). In addition, our group and others have shown that the CDK11p110 isoform and the CDK11p58 isoform are cleaved by caspases to generate a smaller 46–50 kDa protein, CDK11p46, that contains the catalytic portion of the protein (9–11). Generation of this smaller CDK11p46 protein can be triggered by Fas, tumor necrosis factor-{alpha} or staurosporine. Caspase inhibitors can modulate the kinase activity of the caspase-processed protein CDK11p46 (9). Over-expression of CDK11p46 can inhibit cell growth and induce apoptosis, whereas tumor cells with abnormal CDK11p46 protein levels appear to be more resistant to Fas and/or staurosporine-induced apoptosis (9,12).

The human Cdc2L gene locus, encoding CDK11, maps to chromosome band region 1p36, a chromosomal region frequently deleted in many cancers such as childhood sinus tumors, neuroblastomas and lymphomas (13–15) as well as melanoma (16). Reduced mRNA levels of CDK11 have also been reported in expression profile studies of breast cancer, lymphoma, neuroblastoma and melanoma (17,18). In regards to melanoma, we have documented the deletion or translocation of Cdc2L in melanoma cell lines and reduced protein levels in surgical malignant melanoma specimens (19).

Homozygous deletion of cdc2l gene results in early embryonic lethality in mice due to apoptosis of the blastocyst cells between 3.5 and 4 days of postcoitus whereas heterozygous mice appear to develop normally (20). Molecular analysis of tumors (specifically pheochromocytomas and melanomas) derived from Pten+/– Ink4a/Arf –/– mice revealed loss of CDK11/PITSLRE relative to control DNA (21). In addition, loss of CDK11 expression was associated with concurrent expression of activated H-ras and dominant negative mutant p53 genes in skin tumors derived from H-ras/p53 transgenic mice (22). However, despite these observations suggesting a role of CDK11 in neoplastic transformation and tumor progression, whether or not a deficiency in CDK11 has a direct role in tumorigenesis and progression in vivo has not yet been investigated.

Gene trapping has gained prominence in recent years as a method for inactivating a given gene for in vivo studies (23–25). In the present study, we utilized gene trap technology to generate haploinsufficient cdc2l mice with reduced CDK11 protein levels and a well-characterized mouse model of chemically induced multi-step skin carcinogenesis to explore the role of CDK11 in skin carcinogenesis. Our study shows, for the first time, that reduced gene expression of cdc2l gene in haploinsufficient cdc2lGT mice has a significant impact on skin tumor development in a mouse model of chemically induced skin carcinogenesis.


   Materials and methods
Top
 Abstract
 Introduction
 Materials and methods
Results
 Discussion
 Supplementary material
 References
 
Generation of cdc2l gene-trapped (cdc2lGT) mice
An E14 embryonic stem (ES) cell line, XG836, was selected from the Baygenomics Gene Trap Resource Database (http://baygenomics.ucsf.edu). This cell line originated from the 129/sv background and contains an insertion of a secretory gene trap vector pGT1.8TM within intron 12 of the cdc2l gene. The vector contains a splice-acceptor sequence upstream of a selection marker/reporter ß-Geo gene (a fusion of ß-galactosidase and neomycin phosphotransferase). The insertion of this vector into the mammalian genome generates a fusion transcript containing sequences from the exons 5' to the insertion site spliced into the ß-Geo reporter sequence (Figure 1A). The selected XG836 ES cells were injected into appropriately maintained blastocysts from C57BL/6J mice and then transferred to pseudopregnant recipient females. The resulting founder chimeric male (a mixture of C57BL/6J:129/sv strains) was further bred onto C57BL/6J females to purify the strain. Mice resulting after multiple rounds of breeding contain one wild-type allele and one allele harboring the gene trap cdc2l-ß-Geo transcriptional fusion and were thus named as cdc2lGT mice (Figure 1A).


Figure 1
View larger version (39K):
[in this window]
[in a new window]
[Download PowerPoint slide]
  		Fig. 1. Characterization of cdc21GT mice. (A) Schematic representation of normal and gene trapped fusion products of cdc21 gene in mouse. The cdc21 gene contains 20 exons resulting in a normal transcript of 2590 bp. Primers P1 and P2 are designed in exons 12 and 13 respectively. Position 1435 on the transcript represents the position at which exons 12 and 13 join after splicing (above). Insertion of the ß-Geo coding sequence within intron 12 of the cdc21 gene results in a cdc21-ß-Geo fusion transcript. The fusion transcript is terminated prematurely at position 1461 (below). Empty and hashed boxes represent exons and ß-Geo coding sequence respectively. Introns are shown as solid lines. (B) Genotyping of cdc21+/+ and cdc21GT C57BL/6 mice. Genomic DNA from mouse tail was used in PCR analysis for ß-Geo and Tcrd genes. (C) Real-time RT-PCR analysis comparing CDK11 expression in cdc21+/+ and cdc21GT mice. The cdc21 transcript levels are normalized to mouse ß-actin mRNA. (D) CDK11 protein levels from the skin tissues of CDK11p130 protein was detected by GN1 antibody (dilution 1:1000). Mouse {alpha}-tubulin is used as loading control

 
Genotyping of cdc2lGT mice
All mice were genotyped by polymerase chain reaction (PCR) amplification of genomic DNA using primers that recognize the ß-Geo gene (5'-CAAATGGCGATTACCGTTGA-3' and 5'-TGCCCGTCATAGCCGAATA-3'). As an internal control, the tcrd gene (5'-CAAATGTTGCTTGTCTGGTG-3' and 5'-GTCAGTCGAGTGCACAGTTT-3') was used. Genomic DNA was isolated from rodent tails using the DNeasy Blood and Tissue Kit from Qiagen (Valencia, CA). The PCR cycling conditions were 95°C for 2.5 min, followed by 30 cycles of 94°C for 1 min, 60°C for 45 s and 72°C for 1 min. Subsequently, a final extension at 72°C for 10 min was conducted. The PCR products were analyzed on a 1% agarose gel.

Animals and tumor induction protocol
All experimental procedures on mice were performed as per the regulations of Institutional Animal Care and Use Committees abiding by policies set by University of Arizona. All experimental mice were maintained in a protective environment and handled under sterile conditions in a laminar hood. Mice were shaved on their dorsal part (back) using surgical clippers 2 days prior to the start of the experiment. For the skin carcinogenesis study, four treatment groups were set up each containing 10 cdc2l+/+ and 10 cdc2lGT mice as detailed in Figure 2A. In 7,12-dimethylbenz[a]anthracene (DMBA)/12-O-tetradecanoylphorbol-13-acetate (TPA) group (Group IV), carcinogenesis was initiated once on Day 1 with one dose of 5 µg/200 µl DMBA (prepared in acetone) followed by multiple doses of promotion starting a week later with 5 µg/200 µl TPA (prepared in acetone) administered twice every week for 19 weeks. Three control groups were also established: acetone/acetone (Group I), DMBA/acetone (Group II) and acetone/TPA (Group III). DMBA and TPA were purchased from Sigma Chemicals Co. (St Louis, MO). Stock solutions were prepared and diluted in acetone (Mallinckrodt Chemicals, Phillipsburg, NJ). All mice were given sterilized food and water ad libitum. During the carcinogen treatments, mice were weighed periodically and tumor sizes (diameter) were measured using calipers. The measured tumor diameters were converted to tumor volume using the following formula as described previously (26):

Formula


Figure 2
View larger version (30K):
[in this window]
[in a new window]
[Download PowerPoint slide]
  		Fig. 2. DMBA/TPA induced skin papillomas in cdc21+/+ and cdc21GT mice. (A) Four carcinogen treatment groups were set up. Each treatment group contained 10 mice each of cdc21+/+ and cdc21GT genotypes. The experiment was started at Day 1 (two days after the dorsal part of the mouse skin was shaved) with a single dose of either DMBA or vehicle (Acetone) and carried over for 19 weeks with multiple applications of TPA or vehicle (Acetone). Concentrations of carcinogen are described in Materials and Methods. (B) Tumor Multiplicity. The total number of skin tumors were counted every week for 19 weeks for cdc21+/+ and cdc21GT mice treated as in Group IV. Skin tumor multiplicity was calculated as the average number of papillomas per mouse for cdc21+/+ and cdc21GT mice treated as in Group IV (p < 0.05). (C) CDK 11 protein expression in cdc21+/+ and cdc21GT mice. Western analysis comparing mouse CDK11p130 protein expression in normal appearing skin (S) or skin papillomas (T) derived from cdc21+/+ and cdc21GT mice treated as in Group IV. Protein lysates were obtained as described in Materials and Methods. Presence of CDK11p130 protein was detected by GN1 antibody (dilution 1:1000). Mouse {alpha}-tubulin is used as loading control.

 
Histopathology
The experiment was terminated at week 19. All mice were necropsied. The dorsal skin tissue including tumors were dissected out from euthanized animals and fixed in 10% neutral-buffered formalin (VWR, West Chester, PA) for 24 h and embedded in paraffin. At the same time, frozen tissue sections and solid tumors (wherever possible) were also harvested in liquid nitrogen for various molecular analyses. In addition, 2–3 mm sections from DMBA/TPA (Group IV)-treated skin, as well as cleanly dissected portions of solid tumors, wherever possible, were collected for H-ras PCR–restriction fragment length polymorphism analysis. Serial sections (5 µm thick) of paraffin-embedded tissues were prepared and processed in graded ethanol solutions and xylene, purchased from Sigma Chemicals Co. Paraffinized tissue sections were stained with hematoxylin and eosin (Richard Allan Scientific, Kalamazoo, MI) and subjected to review by a pathologist (Dr A.K.Bhattacharyya).

Immunohistochemistry
Cleaved caspase-3 immunohistochemistry.
Apoptosis was assessed by cleaved caspase-3 staining as described previously (27). Skin tumors were fixed, paraffin embedded and sectioned as described above. These sections were deparaffinized in xylene, followed by rehydration in graded series of ethanol and ending with water immersion. Antigen retrieval was performed by microwave exposure in sodium citrate buffer (2.1 g/l, pH 6.1). Endogenous peroxidase blocking was performed with 3% H2O2 in methanol and sections were blocked with 1.5% normal goat serum (Vector Laboratories, Burlingame, CA). Then sections were incubated with a rabbit polyclonal anti-cleaved caspase-3 antibody followed by a biotinylated secondary antibody (Vector Laboratories). Sections were then treated with Vectastain Elite ABC Reagent, used according to the manufacturer's instructions (Vector Laboratories), diaminobenzidine activated with H2O2 followed by hematoxylin counterstain. Cleaved caspase-3 expression was evaluated in the tumors directly as well as on adjacent skin. Six mice each of cdc2l+/+ and cdc2lGT genetic backgrounds, treated with DMBA/TPA (Group IV), were used in this study.

Ki-67 immunohistochemistry.
Cell proliferation was assessed by immunohistochemistry staining for Ki-67 antigen using anti-mouse Ki-67 antibody (Novacastro NCL ki-67p raised in rabbit; dilution 1:500) and developed using the biotin–streptavidin complex method. Probing was performed by the Tissue and Cellular/Molecular Analysis Shared Service Core Facility at the Arizona Cancer Center in the Discovery® XT Automated IHC System (Ventana Molecular Discovery Systems, Tucson, AZ). To improve antigen detection, sections were subjected to antigen retrieval prior to staining as described above. Proliferation was quantified under a 40x light objective and was expressed as a proliferative index (PI) score that was determined as the mean percentage of nuclei staining positive for Ki-67 antibody in 200 cells at 40x magnification. Two mice each of cdc2l+/+ and cdc2lGT genetic backgrounds, treated with DMBA/TPA (Group IV), were used in this study. Three to four fields of view for each animal were included in the calculation for PI score.

Western analysis
Tissue protein extracts were prepared from frozen mouse skin tissue minced in 500 µl of tissue lysis buffer [20 mM Tris pH 7.5, 150 mM NaCl, 1 mM ethylenediaminetetraacetic acid, 1% NP-40, 0.5% sodium deoxycholate, 1% protease inhibitor cocktail (Sigma), 1 mM phenyl methyl sulfonyl fluoride, 1 mM sodium orthovanadate]. Following lysis, the samples were centrifuged at 13 000g for 30 min at 4°C and the protein content was estimated using bicinchonic acid assay (Pierce, Rockland, IL). Equal amounts of protein were resolved by electrophoresis through a 10% sodium dodecyl sulfate–polyacrylamide gel and then transferred to polyvinylidene difluoride membrane (Bio-Rad, Philadelphia, PA). The membrane was blocked with 5% non-fat dry milk solution prepared in 1x phosphate-buffered saline and then probed with anti-CDK11 GN1 primary antibody. A secondary probe with horseradish peroxidase-labeled antibody (Sigma Chemicals Co.) was detected by enhanced chemiluminescence (GE Healthcare Bio-Sciences Corp., Piscataway, NJ). Anti-mouse {alpha}-tubulin antibody was used as loading control.

Antibodies
Anti-CDK11 GN1 rabbit polyclonal antibody recognizes the amino acids 341–413 of CDK11p110 isoform (12). Anti-mouse {alpha}-tubulin monoclonal antibody raised in chicken was purchased from EMD Biosciences (San Diego, CA). Anti-cleaved caspase-3 (asp175) and anti-ki67 antibodies for immunohistochemistry were purchased from Cell Signaling Technology (Danvers, MA) and Novacastro/Vector Laboratories, respectively.

Quantitative real time RT–PCR analysis of cdc2l gene
Total RNA was extracted from frozen skin tissue of mice using RNeasy Mini Kit (Qiagen). Reverse transcription (RT) was performed with 2 µg of total RNA using the cDNA iScript Kit (Bio-Rad). Two microliters of cDNA was further used to perform real time RT–PCR using iQTM SYBR® Green Supermix (Bio-Rad) and amplified in a PerkinElmer Life Sciences Prism 5700 Sequence Detection System according to the manufacturer's instructions. Dissociation curves were analyzed to determine the specificity of the amplicon. Each sample was amplified in triplicates. Threshold cycle (Ct) during the exponential phase of amplification was determined by real time monitoring of fluorescent emission. Mouse ß-actin was used as a control gene. The primers for the mouse ß-actin and cdc2l genes (5'-CCTAGCACCATGAAGATCAAG-3' and 5'-ATCGTACTCCTGCTTGCTG-3'; 5'-GGGTGGTCTACAGAGCAAAG-3' and 5'-TGAGGATGGTGTTGATTTC-3', respectively) were purchased from Sigma Genosys. The cdc2l mRNA levels were represented as absolute number of copies normalized against ß-actin mRNA. Difference in amplification was calculated as 1/(Ct cdc2l – Ct ß-actin).

Mutation analysis of H-ras gene
Genomic DNA was extracted from skin tissue as well as cleanly dissected portions of solid tumors wherever possible from DMBA/TPA (Group IV)-treated mice using DNeasy Blood and Tissue Kit as per manufacturer's guidelines (Qiagen). Genomic DNA extracted from tails (for genotyping) was used as a negative control. A 208 bp region from exon 2 of H-ras gene was amplified using primers described by Nagase et al. (28). The amplified product was purified using GFX columns to remove any residual buffer and 4 µg of DNA was further digested using restriction enzyme Xba I at 37°C for 3 h and electrophoresed on 4% Nu-sieve 3:1 agarose gel (BioWhittaker Molecular Applications, Rockland, ME) at 80 V for 90 min. Direct DNA sequencing of PCR product was used to verify H-ras codon 61 mutation in one animal (# 105) using primers as described previously.

Epidermal thickness measurements
Serial sections of paraffin-embedded mouse skin tissue were observed under the microscope at 40x magnification. The overall view at 40x magnification was standardized to a diameter of 500 µm. From this, relative skin epidermal thickness measurements were taken as a series of non-overlapping fields of view by spanning the complete skin sections from one end to the other. Four to five complete tissue sections each of three to four mice were used for this study. supplementary Figure 1 (available at Carcinogenesis Online) shows a typical skin section at 4x magnification with highly varying epidermal thickness. Thickness means ± standard deviations are represented in Table II.

Statistical analyses
PIs from KI-67 staining and epidermal thickness measurements were analyzed by two sample t-test and considered significant at P < 0.001. KI-67 immunohistochemistry data were compared in cdc2lGT versus cdc2l+/+ mice. Epidermal thickness measurements were compared between different carcinogen treatment groups as well as in cdc2lGT versus cdc2l+/+ in DMBA/acetone and DMBA/TPA treatment groups. Differences in tumor volumes between cdc2lGT and cdc2l+/+ mice were analyzed by chi-square test and considered significant at P < 0.05. Data for tumor multiplicity were analyzed by the Mann–Whitney U-test for each week of carcinogen treatment and considered significant at P < 0.05.


   Results
Top
 Abstract
 Introduction
 Materials and methods
 Results
Discussion
 Supplementary material
 References
 
Genotyping and characterization of gene-trapped cdc2l mice (cdc2lGt)
Haploinsufficient cdc2l (cdc2lGT) mice were generated by using gene trap technology as described in Materials and methods. Briefly, an ES cell line (XG836), harboring an insertion of ß-Geo gene, within the intron 12 of cdc2l gene (as assessed by 5'RACE on ES cell RNA) was obtained from BayGenomics (data not shown). These ES cells were injected into blastocysts and further implanted onto pseudopregnant C57BL/6J female mice. The resulting founder chimera contained a disruption in the normal message of cdc2l gene due to the formation of a cdc2l-ß-Geo fusion transcript (Figure 1A). All mice were genotyped by PCR amplification of genomic tail DNA using primers specific for ß-Geo and tcrd genes as detailed in Materials and methods. Wild-type cdc2l+/+ mice generated a single 200 bp product representing the internal control tcrd gene, whereas the gene-trapped (cdc2lGT) mice showed the presence of an additional product (581 bp) representing the ß-Geo gene insertion (Figure 1B). We verified that this insertion leads to decreased levels of CDK11 mRNA and protein by quantitative real time RT–PCR and Western analyses (Figure 1C and 1D).

To further confirm the insertional gene trap specifically in the context of the cdc2l gene, we intended to amplify the region containing the ß-Geo insertion within the processed cdc2l transcript by RT–PCR amplification of cDNA from cdc2lGT mice. We used primers P1 and P2 (designed in exons 12 and 13 of the cdc2l gene, respectively) as well as primers P1 and P3 (primer P3 is designed in the ß-Geo coding sequence) to amplify the wild-type and fusion transcripts, respectively, from cDNA synthesized from cdc2lGT mice (Figure 1). Primers P1 and P3 did not yield any detectable product in all cdc2lGT mice (data not shown). Interestingly, primers P1 and P2 yielded a single product representing the normal transcript (125 bp) and not the larger fusion transcript (>500 bp) in all cdc2lGT mice (data not shown). This suggests that cdc2lGT mice contain one functional copy of the cdc2l transcript, whereas the other copy is not functional (Figure 1A). The non-functional cdc2l-ß-Geo fusion transcript appears to generate a premature stop codon (UGA) at position 1461 (analyzed in silico on National Center for Biotechnology Information Open Reading Frame Finder (NCBI ORF Finder). Taken together, PCR genotyping, quantitative real time RT–PCR and Western analyses indicate that the gene trap mutagenesis generated an effective null allele of the mouse cdc2l gene. Hereafter, we use cdc2lGT to denote this insertional mutation.

Disruption of one allele of cdc2l gene results in increased skin papilloma number and growth upon DMBA/TPA-mediated carcinogenesis
Next, we wanted to gain insight into the physiologic role of CDK11 in skin carcinogenesis. For this purpose, we investigated whether loss of CDK11 affected the development of skin papillomas induced by the classical skin carcinogenesis protocol. We studied the susceptibility of wild-type (cdc2l+/+) and gene-trapped (cdc2lGT) mice to chemical carcinogenesis in the DMBA/TPA-induced two-stage skin carcinogenesis model. Four carcinogen treatment groups were set up as detailed in Materials and methods (Figure 2A). The mice were genotyped by PCR analysis prior to the start of the study (supplementary Figure 2A, available at Carcinogenesis Online). Body weights of both cdc2l+/+ and cdc2lGT mice were not significantly different prior to the start of, or at any time during the course of, this study, suggesting that level of DMBA dosing was not toxic to the animals (supplementary Figure 2B, available at Carcinogenesis Online).

Both cdc2l+/+ and cdc2lGT mice started developing papillomas as early as week 10 of DMBA/TPA treatment (Group IV). By week 18 of the study, 50% of cdc2lGT as well as cdc2l+/+ mice developed at least one papilloma (tumor latency). In contrast, no tumors were observed in mice treated with acetone/acetone (Group I) or acetone/TPA (Group III). In the group treated with DMBA/acetone (Group II), a single cdc2lGT mouse developed four papillomas of more than 1 mm3 in size at week 14. Because the tumor burden became excessive in the DMBA/TPA treatment group, we terminated the experiment at week 19. The incidence of papillomas (number of animals with at least one papilloma) after DMBA/TPA treatment for cdc2lGT and cdc2l+/+ mice was similar (70 and 60%, respectively). Starting at week 14, we observed a marked difference in average number of papillomas and the number of papillomas per mouse in cdc2lGT mice relative to cdc2l+/+ mice (P < 0.05). By week 19 of the study, the average number of papillomas and number of papillomas per mouse (tumor multiplicity) in cdc2lGT mice was more than double than that in cdc2l+/+ mice (fold difference cdc2lGT:cdc2l+/+ = 2.7-fold) (Figure 2B). In addition, we also observed a higher frequency of larger tumors (1–5 and >5 mm3) in cdc2lGT mice as compared with cdc2l+/+ mice; P < 0.05 (Table I).


View this table:
[in this window]
[in a new window]

  		Table I. Increased incidence of tumors by size in DMBA/TPA-treated cdc2lGT mice

 
Gene-trapped cdc2lGT mice have reduced CDK11 protein levels
The major p110 isoform of CDK11 protein in humans corresponds to a 130 kDa protein in mice, which we hereby refer to as CDK11p130. By Western analysis, we observed that the protein levels of CDK11p130 were reduced in the skin of cdc2lGT mice as compared with wild-type cdc2l+/+ mice when treated with acetone/acetone (Group I) (Figure 2C, Lanes 1, 2). For DMBA/TPA (Group IV)-treated mice, although CDK11p130 protein was present in the normal skin tissue (S) from both cdc2l+/+ (Figure 2C, Lanes 3, 5) and cdc2lGT (Figure 2C, Lanes 7, 9) mice, the level of CDK11p130 protein was markedly suppressed in the adjacent papilloma tissue (T) of mice irrespective of their genetic backgrounds compared with vehicle (i.e. acetone/acetone)-treated mice (Figure 2C; cdc2l+/+: Lanes 4, 6; cdc2lGT: Lanes 8, 10). Interestingly, we observed the presence of a prominent smaller isoform of CDK11, 120 kDa in size (CDK11p120), in the DMBA/TPA-induced papilloma tissue from wild-type cdc2l+/+ mice in Group IV (Figure 2C, Lanes 4, 6). Taken together, these results suggest that there is a reduction in CDK11p130 protein levels associated with DMBA/TPA exposure and that the carcinogen treatments led to the generation of a smaller CDK11p120 isoform in the papilloma tissue of cdc2l+/+ mice.

Harvey ras (H-ras) mutational analysis
Two distinct molecular characteristics associated with DMBA-initiated skin tumorigenesis are (i) oncogenic activation exclusively occurs in the ras family of proto-oncogenes and (ii) the mutations are almost exclusively an A to T transversion (61CAA to 61CTA) at the middle adenine within codon 61 of the H-ras gene that creates an Xba I restriction site (Figure 3A) (29–31). We amplified the complete 208 bp exon 2 region of H-ras gene by PCR using genomic DNA extracted from papilloma tissue of cdc2l+/+ and cdc2lGT mice (Materials and methods). The primers were designed as described previously (28). We then performed restriction fragment length polymorphism analysis for Xba I restriction site on these PCR amplified fragments to look for A to T transversion mutations in codon 61 of the H-ras gene (Figure 3B). We observed that the frequency of H-ras codon 61 mutations was high (three out of three cdc2l+/+ mice and three out of three cdc2lGT, i.e. 100%) in papilloma tissue regardless of the cdc2l allele status. In addition, direct sequencing verified A to T transversion in one of the alleles of the H-ras gene (Figure 3C).


Figure 3
View larger version (30K):
[in this window]
[in a new window]
[Download PowerPoint slide]
  		Fig. 3. Mutational status of H-ras oncogene. (A) Schematic representation of method for detecting the CAA-to-CTA mutation at codon 61 of the H-ras gene. Forward and Reverse primers amplify the exon 2 (208 bp) of H-ras gene. Codon 61 is shown as an arrowhead. Treatment with DMBA and TPA results in a CAA-to-CTA mutation which is susceptible to Xba I restriction digestion. (B) Detection of the codon 61 CAA-to-CTA mutation by PCR-RFLP method. Genomic DNA isolated from mouse tail or papilloma (tumor) tissues were subjected to PCR amplification and subsequent digestion with restriction enzyme Xba I as described in Materials and Methods. (C) Sequence verification of PCR amplified Exon 2 region. Genomic DNA from papilloma tissue of mouse #105 was directly sequenced. Xba I restriction site is shown within the box. Arrowhead represents the dominant A-to-T transversion mutation.

 
Assessment of tumor growth and proliferation
Histological analysis did not reveal any marked differences between papillomas in cdc2lGT and wild-type mice (Figure 4A, 4B). Benign tumors displayed dyskeratosis, nuclear pleomorphism and an increase in the nuclear to cytoplasmic (N:C) ratio. There was evidence in one tumor (derived from a DMBA/TPA-treated cdc2lGT mouse) of progression toward squamous cell carcinoma. This carcinoma was graded as highly advanced, characterized by the presence of extensive and multiple foci of invasion into the lower dermal layers of the skin (data not shown).


Figure 4
View larger version (123K):
[in this window]
[in a new window]
[Download PowerPoint slide]
  		Fig. 4. Effect of cdc21 deficiency on DMBA/TPA induced hyperplasia and proliferation. (A, B) Histopathology of H & E stained formalin-fixed papillomas. Representative papillomas are shown for cdc2l+/+ control mice and cdc2lGT mice respectively. (C, D) Immunohistochemical analysis of serial sections prepared from formalin-fixed papillomas. Representative papillomas from cdc2l+/+ control mice and cdc2lGT mice show the expression of Ki-67 (proliferation). All images have been taken at 400X magnification. (E, F) Proliferation in mouse skin epidermis. Representative skin tissue sections from cdc2l+/+ and cdc2lGT mice respectively shown the expression of Ki-67 for proliferation. (G, H) H & E stained formalin-fixed mouse skin epidermis. Representative skin sections show epidermal layers from cdc2l+/+ and cdc2lGT mice respectively. All images have been taken at 400X magnification.

 
Positive staining for Ki-67 was evident in papillomas as a diffuse nuclear staining with accentuation of the nucleoli and chromosomes in proliferating cells. Staining was performed as described in Materials and methods. The PI was significantly higher in papillomas from DMBA/TPA (Group IV)-treated cdc2lGT mice (77.75 ± 3.96%) as compared with papillomas from cdc2l+/+ mice (30.84 ± 7.53%); P < 0.001 (Figure 4C, 4D). Apoptosis was assessed by staining for cleaved caspase-3 as described in Materials and methods. We did not observe any difference in the apoptotic index between cdc2lGT mice and cdc2l+/+ mice (data not shown).

The degree of hyperplasia was analyzed by estimating the epidermal thickness of tissue sections. Epidermal thickness of mice treated with various carcinogens was significantly higher as compared with vehicle-treated mice, irrespective of the genetic status of cdc2l gene (Table II). Interestingly, cdc2lGT mice treated with DMBA/acetone (cdc2lGT: 26.17 ± 24.03 µm versus cdc2l+/+: 10.0 ± 0.00 µm) as well as DMBA/TPA (cdc2lGT: 84.86 ± 42.53 µm versus cdc2l+/+: 62.73 ± 35.97 µm) showed a significant increase in epidermal thickness as compared with cdc2l+/+ mice from the same treatments, suggesting that DMBA treatment may influence cdc2l gene expression directly to affect proliferation (Table II). A representation of skin sections from different carcinogen treatment groups is shown in supplementary Figure 3 (available at Carcinogenesis Online) This result is in accordance with Ki-67 staining indicating higher degree of proliferation in skin tissue sections from cdc2lGT mice as compared with cdc2l+/+ mice (Figure 4E and F). Reference hematoxylin- and eosin-stained sections of the same animals are shown in Figure 4G and H.


View this table:
[in this window]
[in a new window]

  		Table II. Elevated evidence of epidermal skin thickness in carcinogen-treated mice

 

   Discussion
Top
 Abstract
 Introduction
 Materials and methods
 Results
 Discussion
Supplementary material
 References
 
There is evidence that decreased expression of CDK11 encoded by the cdc2l gene locus is associated with the development of various malignancies in mice (21,22). However, no information is available, to date, to assess whether modulation of the cdc2l expression may directly affect tumorigenesis and/or progression in vivo. Chemically induced skin carcinogenesis studies continue to provide valuable information concerning the impact of genetic alterations on multi-step carcinogenesis and progression in genetically modified mice (32–35). In the present study, we addressed the role of CDK11 in skin tumor initiation and promotion using haploinsufficient cdc2l mice. These mice express a significantly lower level of CDK11 protein in their skin compared with their wild-type littermates. We now report that in parallel with the decreased CDK11 expression, there was an increase in the multiplicity of benign papilloma lesions in cdc2lGT compared with cdc2l+/+ mice upon topical treatment with the tumor initiator DMBA followed by repeated applications of the tumor promoter TPA (Figure 2B). These results indicate, for the first time, that loss of cdc2l gene expression and reduced CDK11 levels may enhance cutaneous tumor promotion toward malignancy.

Gene trap technology has been invaluable in developing mutant mouse models for inactivating key genes in recent years (23–25). In the present study, we have carefully characterized our gene-trapped cdc2lGT mice. Primers P1 and P2, designed in exons 12 and 13, respectively (Figure 1A), generated a single product of 125 bp in size by RT–PCR analysis of cDNA from cdc2lGT mice. However, the larger transcript representing cdc2l-ß-Geo fusion could not be detected. Also, primers P1 and P3, designed to detect a 275 bp product within the fusion region, failed to generate any product. Additionally, the fusion sequence generates a premature stop codon (UGA) at position 1461 suggesting that perhaps the fusion transcript remains undetectable by RT–PCR due to degradation of the null allele. Together with genotyping, real time RT–PCR and Western analyses, our results confirm that the cdc2lGT mice contain only one functional copy of the cdc2l gene. This is in accordance with previous studies showing that homozygous deletion of the cdc2l gene causes early embryonic lethality, whereas heterozygote mice develop normally (20).

We further evaluated the H-ras mutational status in papillomas from cdc2lGT and cdc2l+/+ mice. The present results show high frequency of A to T transversion mutations in codon 61 of H-ras gene in these mice irrespective of their genotype (Figure 3B). This observation is consistent with previous observations on H-ras oncogene activation upon exposure to DMBA/TPA treatments (29,30). These data suggest that the mechanisms by which mutations are introduced into the H-ras gene by DMBA does not differ between cdc2lGT and wild-type mice.

We also investigated the level of CDK11 in adjacent skin tissues and papillomas arising in cdc2lGT and cdc2l+/+ mice after DMBA/TPA treatment. As expected, we observed significantly reduced levels of CDK11 protein in the cdc2lGT mice in both adjacent skin and papillomas after DMBA/TPA treatment as compared with animals without treatment. Interestingly, whereas CDK11 was present in the adjacent skin of wild-type mice, an abundant smaller CDK11p120 isoform was observed in the papillomas derived from these cdc2l+/+ mice treated with DMBA/TPA (Figure 2C). It is plausible that DMBA and subsequent H-ras oncogene activation may lead to alternative promoter usage within the cdc2l gene resulting in the generation of the smaller isoform. Studies by Francone and Mezquita indicate that the presence of alternate CDK11 promoters and exons may play a role in the diversification of CDK11 transcripts during chicken testis development and upon testicular regression by diethylstilbestrol treatment (36). Another possibility is that DMBA may induce mutations within the cdc2l gene in wild-type animals leading to a truncated, smaller CDK11 protein. We are presently pursuing these possibilities.

To gain insight into the mechanism by which loss of cdc2l might contribute to tumor development, we evaluate cell proliferation and apoptosis in DMBA/TPA-induced tumors. We observed an increase in proliferation in tumors derived from the cdc2lGT mice compared with their wild-type littermates (Figure 4). This result is consistent with our observation of increased tumor multiplicity in cdc2lGT mice as compared with cdc2l+/+ mice. We also observed that the thickness of the epidermal skin tissue in cdc2lGT mice treated with DMBA/acetone or DMBA/TPA was higher than that in cdc2l+/+ mice (Table II and supplementary Figure 3, available at Carcinogenesis Online). However, promotion with multiple doses of TPA alone (acetone/TPA) in cdc2lGT and cdc2l+/+ mice showed similar degree of increase in epidermal thickness, suggesting that exposure to DMBA, and not TPA, might be responsible for increase in proliferation of cells in cdc2lGT mice. CDK11 isoforms are implicated in RNA processing and transcription, apoptosis and mitosis; all events that can contribute to cancer. Exactly which function of CDK11 contributes to tumor promotion will require further study.

In conclusion, this is the first in vivo study using the two-stage skin chemical carcinogenesis protocol to directly address the role of haploinsufficiency of cdc2l gene in tumor development. The loss of CDK11 protein levels markedly increases skin tumor multiplicity. These data suggest that loss of cdc2l, encoding CDK11, plays a role in regulating tumor promotion, possibly by facilitating the proliferation of the initiated cells in the tumor microenvironment and consequently facilitating tumor development.


   Supplementary material
Top
 Abstract
 Introduction
 Materials and methods
 Results
 Discussion
 Supplementary material
References
 
Supplementary figures 1–3 can be found at http://carcin.oxfordjournals.org/


   Footnotes
 
{dagger} These authors contributed equally to this study. Back


   Acknowledgments
 
This work was supported by National Institutes of Health Grant CA70145. We are grateful to Dr John W.B.Hershey (U.C.Davis) for assistance in generating the cdc2lGT mice. We also thank Dr G.Tim Bowden and Eva Sikorski for assistance with these studies. We also thank Jim Averill and the Tissue and Cellular/Molecular Analysis Shared Service Core Facility (Arizona Cancer Center, National Institutes of Health Grant CA23074) for immunohistochemistry analysis.

Conflict of Interest Statement: None declared.


   References
Top
 Abstract
 Introduction
 Materials and methods
 Results
 Discussion
 Supplementary material
 References
 

  1. Loyer P, et al. The RNP protein, RNPS1, associates with specific isoforms of the p34cdc2-related PITSLRE protein kinase in vivo. J. Cell Sci. (1998) 111:495–506.[Abstract]
  2. Trembley JH, et al. PITSLRE p110 protein kinases associate with transcription complexes and affect their activity. J. Biol. Chem. (2002) 277:2589–2596.[Abstract/Free Full Text]
  3. Dickinson LA, et al. Cyclin L is an RS domain protein involved in pre-mRNA splicing. J. Biol. Chem. (2002) 277:25465–25473.[Abstract/Free Full Text]
  4. Hu D, et al. CDK11 complexes promote pre-mRNA splicing. J. Biol. Chem. (2003) 278:8623–8629.[Abstract/Free Full Text]
  5. Mayeda A, et al. Purification and characterization of human RNPS1: a general activator of pre-mRNA splicing. EMBO J. (1999) 18:4560–4570.[CrossRef][Web of Science][Medline]
  6. Trembley JH, et al. Casein kinase 2 interacts with cyclin-dependent kinase 11 (CDK11) in vivo and phosphorylates both the RNA polymerase II carboxyl-terminal domain and CDK11 in vitro. J. Biol. Chem. (2003) 278:2265–2270.[Abstract/Free Full Text]
  7. Petretti C, et al. The PITSLRE/CDK11p58 protein kinase promotes centrosome maturation and bipolar spindle formation. EMBO Rep. (2006) 7:418–424.[Web of Science][Medline]
  8. Lahti JM, et al. PITSLRE protein kinase activity is associated with apoptosis. Mol. Cell. Biol. (1995) 15:1–11.[Abstract]
  9. Ariza ME, et al. Fas-induced apoptosis in human malignant melanoma cell lines is associated with the activation of the p34(cdc2)-related PITSLRE protein kinases. J. Biol. Chem. (1999) 274:28505–28513.[Abstract/Free Full Text]
  10. Beyaert R, et al. Cleavage of PITSLRE kinases by ICE/CASP-1 and CPP32/CASP-3 during apoptosis induced by tumor necrosis factor. J. Biol. Chem. (1997) 272:11694–11697.[Abstract/Free Full Text]
  11. Tang D, et al. Phosphorylation of PITSLRE p110 isoforms accompanies their processing by caspases during Fas-mediated cell death. J. Biol. Chem. (1998) 273:16601–16607.[Abstract/Free Full Text]
  12. Shi J, et al. The p34cdc2-related cyclin-dependent kinase 11 interacts with the p47 subunit of eukaryotic initiation factor 3 during apoptosis. J. Biol. Chem. (2003) 278:5062–5071.[Abstract/Free Full Text]
  13. Perlman EJ, et al. Deletion of 1p36 in childhood endodermal sinus tumors by two-color fluorescence in situ hybridization: a pediatric oncology group study. Genes Chromosomes Cancer (1996) 16:15–20.[CrossRef][Web of Science][Medline]
  14. Lahti JM, et al. Alterations in the PITSLRE protein kinase gene complex on chromosome 1p36 in childhood neuroblastoma. Nat. Genet. (1994) 7:370–375.[CrossRef][Web of Science][Medline]
  15. Dave BJ, et al. Deletion of cell division cycle 2-like 1 gene locus on 1p36 in non-Hodgkin lymphoma. Cancer Genet. Cytogenet. (1999) 108:120–126.[CrossRef][Web of Science][Medline]
  16. Chin L, et al. Malignant melanoma: genetics and therapeutics in the genomic era. Genes Dev. (2006) 20:2149–2182.[Abstract/Free Full Text]
  17. Wang Q, et al. Integrative genomics identifies distinct molecular classes of neuroblastoma and shows that multiple genes are targeted by regional alterations in DNA copy number. Cancer Res. (2006) 66:6050–6062.[Abstract/Free Full Text]
  18. Barrett T, et al. NCBI GEO: mining millions of expression profiles—database and tools. Nucleic Acids Res. (2005) 33:D562–D566.[Abstract/Free Full Text]
  19. Nelson MA, et al. Abnormalities in the p34cdc2-related PITSLRE protein kinase gene complex (CDC2L) on chromosome band 1p36 in melanoma. Cancer Genet. Cytogenet. (1999) 108:91–99.[CrossRef][Web of Science][Medline]
  20. Li T, et al. Failure to proliferate and mitotic arrest of CDK11(p110/p58)-null mutant mice at the blastocyst stage of embryonic cell development. Mol. Cell. Biol. (2004) 24:3188–3197.[Abstract/Free Full Text]
  21. You MJ, et al. Genetic analysis of Pten and Ink4a/Arf interactions in the suppression of tumorigenesis in mice. Proc. Natl Acad. Sci. USA (2002) 99:1455–1460.[Abstract/Free Full Text]
  22. Zhang Z, et al. Induction of invasive mouse skin carcinomas in transgenic mice with mutations in both H-ras and p53. Mol. Cancer Res. (2005) 3:563–574.[Abstract/Free Full Text]
  23. Stanford WL, et al. Gene-trap mutagenesis: past, present and beyond. Nat. Rev. Genet. (2001) 2:756–768.[CrossRef][Web of Science][Medline]
  24. Stryke D, et al. BayGenomics: a resource of insertional mutations in mouse embryonic stem cells. Nucleic Acids Res. (2003) 31:278–281.[Abstract/Free Full Text]
  25. Skarnes WC. Two ways to trap a gene in mice. Proc. Natl Acad. Sci. USA (2005) 102:13001–13002.[Free Full Text]
  26. Steel GG. The heyday of cell population kinetics: insights from the 1960s and 1970s. Semin. Radiat. Oncol. (1993) 3:78–83.[CrossRef][Medline]
  27. Yerushalmi HF, et al. Role of polyamines in arginine-dependent colon carcinogenesis in Apc(Min) (/+) mice. Mol. Carcinog. (2006) 45:764–773.[CrossRef][Web of Science][Medline]
  28. Nagase H, et al. Allele-specific Hras mutations and genetic alterations at tumor susceptibility loci in skin carcinomas from interspecific hybrid mice. Cancer Res. (2003) 63:4849–4853.[Abstract/Free Full Text]
  29. Balmain A, et al. Activation of the mouse cellular Harvey-ras gene in chemically induced benign skin papillomas. Nature (1984) 307:658–660.[CrossRef][Medline]
  30. Quintanilla M, et al. Carcinogen-specific mutation and amplification of Ha-ras during mouse skin carcinogenesis. Nature (1986) 322:78–80.[Medline]
  31. Tang MS, et al. Both (+/-)syn- and (+/-)anti-7,12-dimethylbenz[a]anthracene-3,4-diol-1,2-epoxides initiate tumors in mouse skin that possess -CAA- to -CTA- mutations at Codon 61 of c-H-ras. Cancer Res. (2000) 60:5688–5695.[Abstract/Free Full Text]
  32. Formisano P, et al. Raised expression of the antiapoptotic protein ped/pea-15 increases susceptibility to chemically induced skin tumor development. Oncogene (2005) 24:7012–7021.[CrossRef][Web of Science][Medline]
  33. Hsieh YH, et al. Papilloma development is delayed in osteopontin-null mice: implicating an antiapoptosis role for osteopontin. Cancer Res. (2006) 66:7119–7127.[Abstract/Free Full Text]
  34. Nicol CJ, et al. PPARgamma influences susceptibility to DMBA-induced mammary, ovarian and skin carcinogenesis. Carcinogenesis (2004) 25:1747–1755.[Abstract/Free Full Text]
  35. Suzuki JS, et al. Metallothionein deficiency enhances skin carcinogenesis induced by 7,12-dimethylbenz[a]anthracene and 12-O-tetradecanoylphorbol-13-acetate in metallothionein-null mice. Carcinogenesis (2003) 24:1123–1132.[Abstract/Free Full Text]
  36. Francone VP, et al. Diversification of CDK11 transcripts during chicken testis development and regression. Mol. Reprod. Dev. (2005) 72:273–280.[CrossRef][Web of Science][Medline]
Received January 16, 2007; revised March 2, 2007; accepted March 13, 2007.


Add to CiteULike CiteULike   Add to Connotea Connotea   Add to Del.icio.us Del.icio.us    What's this?


This article has been cited by other articles:


Home page
 J. Cell Sci.Home page
D. Hu, M. Valentine, V. J. Kidd, and J. M. Lahti
CDK11p58 is required for the maintenance of sister chromatid cohesion
J. Cell Sci., July 15, 2007; 120(14): 2424 - 2434.
[Abstract] [Full Text] [PDF]

This Article
Right arrow Abstract Freely available
Right arrow FREE Full Text (PDF) Freely available
Right arrow Supplementary Data
Right arrow All Versions of this Article:
28/9/2028    most recent
bgm066v1
Right arrow Alert me when this article is cited
Right arrow Alert me if a correction is posted
Services
Right arrow Email this article to a friend
Right arrow Similar articles in this journal
Right arrow Similar articles in ISI Web of Science
Right arrow Similar articles in PubMed
Right arrow Alert me to new issues of the journal
Right arrow Add to My Personal Archive
Right arrow Download to citation manager
Right arrow Search for citing articles in:
ISI Web of Science (3)
Right arrowRequest Permissions
Google Scholar
Right arrow Articles by Chandramouli, A.
Right arrow Articles by Nelson, M. A.
Right arrow Search for Related Content
PubMed
Right arrow PubMed Citation
Right arrow Articles by Chandramouli, A.
Right arrow Articles by Nelson, M. A.
Social Bookmarking
Add to CiteULike   Add to Connotea   Add to Del.icio.us  
What's this?