Differential Expression of CHL1 Gene during Development of Major Human Cancers

Abstract

Background

CHL1 gene (also known as CALL) on 3p26.3 encodes a one-pass trans-membrane cell adhesion molecule (CAM). Previously CAMs of this type, including L1, were shown to be involved in cancer growth and metastasis.

Methodology/Principal Findings

We used Clontech Cancer Profiling Arrays (19 different types of cancers, 395 samples) to analyze expression of the CHL1 gene. The results were further validated by RT-qPCR for breast, renal and lung cancer. Cancer Profiling Arrays revealed differential expression of the gene: down-regulation/silencing in a majority of primary tumors and up-regulation associated with invasive/metastatic growth. Frequent down-regulation (>40% of cases) was detected in 11 types of cancer (breast, kidney, rectum, colon, thyroid, stomach, skin, small intestine, bladder, vulva and pancreatic cancer) and frequent up-regulation (>40% of cases) – in 5 types (lung, ovary, uterus, liver and trachea) of cancer. Using real-time quantitative PCR (RT-qPCR) we found that CHL1 expression was decreased in 61% of breast, 60% of lung, 87% of clear cell and 89% papillary renal cancer specimens (P<0.03 for all the cases). There was a higher frequency of CHL1 mRNA decrease in lung squamous cell carcinoma compared to adenocarcinoma (81% vs. 38%, P = 0.02) without association with tumor progression.

Conclusions/Significance

Our results suggested that CHL1 is involved in the development of different human cancers. Initially, during the primary tumor growth CHL1 could act as a putative tumor suppressor and is silenced to facilitate in situ tumor growth for 11 cancer types. We also suggested that re-expression of the gene on the edge of tumor mass might promote local invasive growth and enable further metastatic spread in ovary, colon and breast cancer. Our data also supported the role of CHL1 as a potentially novel specific biomarker in the early pathogenesis of two major histological types of renal cancer.

About the Authors:

Vera N. Senchenko

Affiliation: Laboratory of Structural and Functional Genomics, Engelhardt Institute of Molecular Biology, Russian Academy of Sciences, Moscow, Russia

George S. Krasnov

Affiliation: Laboratory of Structural and Functional Genomics, Engelhardt Institute of Molecular Biology, Russian Academy of Sciences, Moscow, Russia

Alexey A. Dmitriev

Affiliation: Laboratory of Structural and Functional Genomics, Engelhardt Institute of Molecular Biology, Russian Academy of Sciences, Moscow, Russia

Anna V. Kudryavtseva

Affiliation: Laboratory of Structural and Functional Genomics, Engelhardt Institute of Molecular Biology, Russian Academy of Sciences, Moscow, Russia

Ekaterina A. Anedchenko

Affiliation: Laboratory of Structural and Functional Genomics, Engelhardt Institute of Molecular Biology, Russian Academy of Sciences, Moscow, Russia

Eleonora A. Braga

Affiliation: Laboratory of Molecular Diagnosis, Russian State Genetics Center GosNIIgenetika, Moscow, Russia

Irina V. Pronina

Affiliation: Laboratory of Molecular Diagnosis, Russian State Genetics Center GosNIIgenetika, Moscow, Russia

Tatiana T. Kondratieva

Affiliation: Blokhin Cancer Research Center, Russian Academy of Medical Sciences, Moscow, Russia

Sergey V. Ivanov

Affiliation: Cardiothoracic Surgery Department, NYU Langone Medical Center, New York, New York, United States of America

Eugene R. Zabarovsky

* E-mail: [email protected]

Affiliations Laboratory of Structural and Functional Genomics, Engelhardt Institute of Molecular Biology, Russian Academy of Sciences, Moscow, Russia, Department of Microbiology, Tumor and Cell Biology, Department of Clinical Science and Education, Sodersjukhuset, Karolinska Institute, Stockholm, Sweden

Michael I. Lerman

Affiliation: Hematology Branch, National Heart, Lung, and Blood Institute, National Institutes of Health, Bethesda, Maryland, United States of America

Introduction

Cancer-associated genes fall into two main categories: cancer-causing genes that drive malignant transformation and maintain tumor growth, and cancer progression genes that orchestrate local invasion and further spread of metastatic cells and growth of distant metastases [1], [2], [3]. The CHL1 gene – close homolog of L1, also known as CALL - cell adhesion L1-like (GenBank Accession No. NM_006614.2) encodes a one-pass trans-membrane cell adhesion molecule (CAM) capable of both homotypic and heterotypic binding. The protein encoded by this gene is a member of the L1 gene family of neural cell adhesion molecules. It is a neural recognition molecule that may be involved in signal transduction pathways. CHL1 is expressed in normal tissues besides the brain and is expressed in a variety of human cancer cell lines and primary tumor tissues [4], [5]. It was also shown that the gene is involved in general cognitive activities (g/IQ) [6], [7] and some neurological diseases (i.e. schizophrenia [8]). The deletion of one copy of this gene may be responsible for mental defects in patients with 3p- syndrome. Recently several CAMs including L1 were shown to be involved in cancer growth and metastasis [9], [10]. CHL1 is located at 3p26, a region that is shown to harbor a candidate for prostate cancer susceptibility in Finnish prostate cancer families, although no mutations were detected in the coding part of the gene [11]. Thus, these reports suggest that CHL1 plays a role in cancer development [12], not only in neuronal activities. Previously, in collaboration with Dr. Helen S. Smith, we performed a deletion mapping of the short arm of chromosome 3 on a panel of breast cancers and delineated three regions as harboring breast cancer candidate tumor suppressor genes (TSG), namely, 3p24-26, 3p21-22, and 3p12-13 [13], [14], [15], [16]. Then we cloned the CHL1 (CALL) gene in 1997/1998 and analyzed its expression in mouse development and performed extensive bioinformatics analysis [5].

Here we provided a comprehensive study of CHL1 mRNA expression using two methods. Qualitative analysis was performed using Clontech Cancer Profiling Arrays, and further real time quantitative PCR (RT-qPCR) was employed for validation of the microarray data for three major cancer types: non-small cell lung cancer (NSCLC), breast cancer (BC), and renal cell carcinomas (RCC). Our results suggested a dual role of the CHL1 in tumorigenesis: it may contribute to initial tumor growth and then to progression and finally tumor spread/metastasis. The data further supported the role of CHL1 as a potentially novel specific biomarker in the early pathogenesis of two major histological types of renal cancer.

The work is dedicated to the memory of Dr. Helen S. Smith.

Results

In silico analysis of CHL1 expression in normal and tumor tissues

The vast public expression databases allow to detect and quantify the expression of most if not all known RefSeq genes (∼20,000) in normal and tumor tissues. We used several public web-based servers to analyze mouse and human CHL1 expression [17], [18], [19], [20]. The data shows that CHL1 is expressed in many normal adult and fetal tissues besides the brain and peripheral nervous system [17], [19]. Variable expression was seen in many tumors; it was especially high in a melanoma cell line G361. According to Oncomine [18] preliminary data based on microarray analysis, the CHL1 expression also varies in several major cancer types – renal [21], [22], cervical [23], colon [24], [25], ovary [26], lung [27], [28], stomach [29] and breast [30], [31] cancer. The Oncomine also showed co-expression of CHL1 with another known cancer metastasis-associated gene, lysyl oxidase (LOX) [32] in metastatic melanoma.

Investigation of CHL1 expression with Cancer Profiling Arrays

We used Cancer Profiling Arrays I and II (Clontech) to test the CHL1 expression in a large sample of human primary tumors including breast, lung, kidney, ovary, colon, stomach and others (Fig. 1). Only 395 samples of 486, including 90 metastatic tumors and 12 metastases were informative. We first showed that the change of CHL1 expression in all studied tumors compared to the matched non-cancerous (normal) tissues was statistically significant (P<0.05, Fisher exact test or χ² criteria).

[Figure omitted. See PDF.]

Figure 1. Results of microarray analysis of CHL1 mRNA level in different types of cancer.

Abbreviations used: ADC - adenocarcinoma, ASC - adenosquamous carcinoma, BAC - bronchiolo-alveolar adenocarcinoma, C - carcinoma, CAC - cystadenocarcinoma, CC-ADC - clear cell adenocarcinoma, EDST - endodermal sinus tumor, ENB - epithelial nephroblastoma, ESS - endometrial stromal sarcoma, FAC - follicular adenocarcinoma, FS - fibrosarcoma, I-DC - infiltrating ductal carcinoma, I-IDC - infiltrating intraductal carcinoma, I-LC - infiltrating lobular carcinoma, LC - lobular carcinoma , LDC - mixed lobular-ductal carcinoma, LM - leiomyoma, M - malignant melanoma, MAC - mucinous adenocarcinoma, MBC - mucinous borderline carcinoma, MC - medullary carcinoma, MMMT - malignant mixed Mullerian tumor, NI-IDC - noninfiltrating intraductal carcinoma, PAC - papillary adenocarcinoma, PC - papillary carcinoma, PSC - papillary serous carcinoma, PSCA - papillary serous cystadenoma, PSCAC - papillary serous cystadenocarcinoma, RCC - renal cell carcinoma, S - seminoma, SC - serous carcinoma, SCAC - serous cystadenocarcinoma, SCC - squamous cell carcinoma, TAC - tubular adenocarcinoma, TC - tubular carcinoma, TCC - transitional cell carcinoma, UBT - uterus benign tumor. Asterisks (*) show samples with metastases. **G361 – a melanoma cell line. The boxed samples indicate a matched normal (left) - primary tumor (bottom right) pair with an associated metastatic sample (upper right corner of a box). T - tumor samples; N - matched normal control specimens.

https://doi.org/10.1371/journal.pone.0015612.g001

Down-regulation.

As demonstrated by Cancer Profiling Arrays data in Figures 1 and 2, a high percentage of patients displayed a down-regulation of CHL1 expression in breast, kidney, rectum, colon, thyroid, stomach, skin, small intestine, bladder, vulva and pancreatic cancer. The results of the microarray data analysis were presented for 11 types of cancer in Table 1. In total, a statistically significant decrease of CHL1 expression was found in breast cancer - 71% (45 of 63 cases), colon - 48% (23 of 48), rectum - 50% (14 of 28), thyroid - 69% (11 of 16), kidney - 75% (21 of 28) and small intestine – 67% (6 of 9) cancers. Importantly, a statistically significant increase of down-regulation frequency was shown in samples with metastases compared to samples without metastases in colon (83% vs. 36%, P = 0.01) and rectum (75% vs. 31%, P = 0.05) cancers. The same tendency was found in ovary cancer (60% vs. 19%, P = 0.17).

[Figure omitted. See PDF.]

Figure 2. Distribution of primary tumors with different CHL1 mRNA level alterations.

Fraction of tumors with CHL1 up-regulation is showed with red, down-regulation – green, mRNA level retention – yellow. Data revealed with Clontech Microarray analysis. Asterisks (*) show statistically significant differences between frequencies of CHL1 expression changes with up- and down-regulation.

https://doi.org/10.1371/journal.pone.0015612.g002

[Figure omitted. See PDF.]

Table 1. Summary of CHL1 expression in 11 tumor types as measured with the Clontech Cancer Profiling Arrays I and II.

https://doi.org/10.1371/journal.pone.0015612.t001

Up-regulation.

The CHL1 up-regulation (frequency from 20% to 100%) was found in lung, ovary, uterus, liver, skin, prostate, stomach, cervix and trachea cancers. However, the increase of the CHL1 mRNA level was statistically significant only in lung cancer −64% (16 of 25, P<0.01). The majority of such cases (14 of 22, P<0.01) were found in different histotypes of NSCLC (ADC, BAC, SCC) at the Stage I. Also we observed several cases of CHL1 up-regulation in metastatic tumors (stomach, lung, trachea, ovary and uterus, Table 1). Thus, the cases with CHL1 up-regulation could serve as examples of CHL1 involvement both in initial and possibly in further progression and invasive tumor growth.

Deregulation.

In uterus and ovary cancer the frequency of up- and down-regulation was close (41% and 30%, 46% and 27%, respectively). In ovary cancer the down-regulation was a prevalent event (52%) in samples without metastases, on the contrary the up-regulation was prevalent (60%) in the group of metastatic tumors. In stomach cancer a statistically significant change of CHL1 expression (both up- and down-regulation) was shown in the group with metastases relative to the group without metastases (88% vs. 45%, P = 0.02).

Metastases.

We observed re-expression of the CHL1 in 4 of 12 metastases (first coordinate) along with low CHL1 mRNA level in primary tumor (second coordinate): in ovary (24K/24L), colon (14O/14P, 14V/14W) and breast (4I/4J, Fig. 1, Array I). In addition, we also found silencing of the gene expression in both metastasis and primary tumors, for example, breast cancer (4G/4H, 4K/4L).

The CHL1 expression in breast, lung and renal cancer tissues studied using RT-qPCR

The CHL1 mRNA content was decreased in the majority of studied tumor samples compared to normal samples but in some tumor samples the CHL1 expression was up-regulated (Fig. 3. A, B and C).

[Figure omitted. See PDF.]

Figure 3. Expression profiling of CHL1 in primary breast, lung and renal tumors.

A. The relative CHL1 mRNA level (R) in breast cancer (BC). N₀ – without metastases, N_1–2 – metastases in regional lymph nodes. Samples #1, 2 (Stage I), #3-22 (Stage II), #23 (Stage IV); samples #3-9 (Grade 1), #10-21 (Grade 2). B. The relative CHL1 mRNA level (R) in lung cancer (NSCLC). SCC – lung squamous cell carcinomas, ADC – lung adenocarcinomas, N – normal samples from cancer free healthy donors; N₀ – without metastases, N_1–2 – metastases in regional lymph nodes; I, II and III – Stages. C. The relative CHL1 mRNA level (R) in renal cancer (RCC). CC-RCC – clear cell renal carcinomas, pRCC – papillary renal carcinomas; N₀ – without metastases, N_1–2 – metastases in regional lymph nodes; I, II and III – Stages.

https://doi.org/10.1371/journal.pone.0015612.g003

Breast cancer (BC).

We found that the CHL1 mRNA level was decreased in 61% (14 of 23, P<0.03), increased in 22% (5 of 23) and not changed in 17% (4 of 23) of samples. Maximal decrease of the CHL1 mRNA level was 20-fold, maximal increase was 34-fold. There was no evident correlation between the change of the CHL1 expression and the tumor progression (Fig. 3 A).

Non-small cell lung cancer (NSCLC).

The CHL1 mRNA level was decreased in 60% (18 of 30, P<0.02) and was normal in 33% (10 of 30), i.e. less than 2-fold changes. The decrease or increase of the mRNA level was detected neither in lung non-cancerous (normal) matched tissues nor in tissues from cancer free healthy donors. However, for two histological subtypes of NSCLC (ADC and SCC) the frequency of the mRNA changes was different. Down-regulation was observed in 38% (5 of 14) of ADC samples. The increase of the CHL1 mRNA (7-fold) was detected only in one ADC sample. On the contrary, in SCC samples the CHL1 expression was significantly decreased in 81%, (13 of 16, P<0.02). LD (level of mRNA decrease) varied from 2 to 100-fold in ADC and from 2 to 44-fold in SCC. There was a more significant increase of FD (frequency of mRNA decreases) values in SCC as compared to ADC (81% vs. 38%, P = 0.02) without noticeable association with tumor progression (Fig. 3, B and Table 2).

[Figure omitted. See PDF.]

Table 2. The frequency (FD) and the level (LD_av) of the CHL1 mRNA decrease in lung cancer (ADC and SCC types).

https://doi.org/10.1371/journal.pone.0015612.t002

Clear cell renal cell carcinoma (CC-RCC), papillary renal cell carcinoma (pRCC) and renal carcinoma cell lines.

A significant decrease (from 3 to 302-fold) of CHL1 mRNA was detected in 87% (26 of 30, P<0.01) of CC-RCC specimens and 89% (8 of 9, P<0.02) of pRCC specimens with LD_av (geometric mean of LD) equal to 18 and 19-fold respectively (Fig. 3, C). Therefore we could conclude that the frequency and the average level of the CHL1 expression decrease were similar for two major histological types of renal cancer, CC-RCC and pRCC. The LD_av value was significant in all RCC tumors at all developmental stages independent of metastasis presence (Table 3). In CC-RCC with or without metastases, the FD and LD_av values were similar.

[Figure omitted. See PDF.]

Table 3. The frequency (FD) and the level (LD_av) of the CHL1 mRNA decrease in renal cancer (CC-RCC).

https://doi.org/10.1371/journal.pone.0015612.t003

The estimates of the CHL1 mRNA levels in seven renal cancer cell lines revealed strong down-regulation of this gene: 80-fold (Caki2, KRC/Y), about 1000-fold (TK164) and total silencing (TK10, KH39, HN4, Caki1, Fig. 4).

[Figure omitted. See PDF.]

Figure 4. The relative expression levels of CHL1 in CC-RCC cell lines.

The mRNA level of the target gene was normalized to the reference genes RPN1 and GUSB.

https://doi.org/10.1371/journal.pone.0015612.g004

Comparison of microarrays and RT-qPCR data for breast, renal and lung cancer

The microarray data for 61 BC, 23 RCC and 25 NSCLC samples were compared with RT-qPCR data for 23 BC, 30 CC-RCC and 30 NSCLC samples. In our array study a significant CHL1 down-regulation was shown for most of RCC samples; up-regulation was observed in 3 cases only. Our results also showed the down-regulation of CHL1 in most of BC samples independent of the presence of metastases and up-regulation in 7 tumors only. Almost the same results were obtained using RT-qPCR. There were similarities between array and quantitative data for renal and breast cancer (Table 4).

[Figure omitted. See PDF.]

Table 4. The CHL1 mRNA changes using microarrays and RT-qPCR.

https://doi.org/10.1371/journal.pone.0015612.t004

Cancer Profiling Arrays I and II include a very heterogeneous group of lung cancer with different histological subtypes: BAC, ADC, SCC, carcinoid with Stage I and II, only two metastatic tumors and limited number of specimens of each subtype. Overall, there were 15 SCC and 5 ADC which we could compare with RT-qPCR data. We found an up-regulation in 11 SCC (6 of 11 were identified as Stage I) and 4 ADC (3 ADC were Stage I as well); down-regulation in 2 SCC (13%) and 1 ADC (25%, Fig. 1). Recently, we showed an up-regulation of several TSG on 3p in lung ADC at Stage I. These tumors were characterized with high differentiation grade [33]. On the other hand, according to the RT-qPCR data frequency of decreased expression was 38% (5 of 14 cases) in ADC and 81% (13 of 16 cases, P<0.02) in SCC (see Table 2). An up-regulation was detected only in 7% (1 of 14 cases) ADC and never in SCC.

Discussion

CHL1, located at 3p26.1, belongs to the family of cell adhesion molecules (CAM) – cell surface proteins mediating cell-cell and cell-matrix interactions. Alterations in CAM expression (including CHL1) and functions have been implicated in development of different tumor types, for example, melanoma [34], ovary [9], [35], prostate [11] and colon cancer [36]. According to [9], the evaluation of LOH patterns in serous epithelial ovarian cancer (EOC) suggested that CHL1 is a tumor suppressor candidate (TSG). The studies published by us and other authors (see Introduction) suggested that the CHL1 gene could be one of the putative tumor suppressor genes localized on human chromosome 3 [12]. However, overexpression of CHL1 was observed in serous EOC samples [9]. Furthermore, L1 CAM overexpression in malignant melanoma was shown to be associated with metastases [34].

According to Oncomine preliminary microarray expression data [17] along with the prevalent CHL1 down-regulation in several tumors (RCC, lung SCC, colon ADC), the overexpression of CHL1 was found in melanoma. The differential expression was observed in lung ADC [26], cervical [22] and breast [29], [30] cancer.

Based on this data, we hypothesized that CHL1 and other recognition receptors of this type might have dual roles in cancer: in early pre-invasive growth they could serve as TSG and are silenced; later at invasion and metastasis stages these genes might be re-expressed on the edge of the tumor to drive local invasion and enable metastatic spread.

This hypothesis was analyzed in the current study using a combination of preliminary expression screening in 19 different types of epithelial tumors with commercial microarrays (altogether 395 informative samples, Table 1) and evaluation of the CHL1 mRNA expression in primary tumors using RT-qPCR. This method is widely used to corroborate disease-associated expression signatures derived from microarrays. Furthermore, this technology is well suited for translating microarrays data into accurate and quantitative, clinically useful assays [37].

We showed here that the expression of CHL1 was deregulated in major epithelial malignancies (76%, P<0.01, including 54% down-regulation cases according to microarray data). Statistically significant FD values were shown for breast, colon, rectum, thyroid, kidney and small intestine cancer (Table 1). For three social significant/important cancer types – breast, kidney and lung microarray data were validated by RT-qPCR. There was a good concordance between data of two methods for kidney and breast cancer. According to Oncomine microarray data the significant decrease of CHL1 expression level in CC-RCC samples was shown as well.

Clontech Microarray (overexpression in 64% lung cancer samples) and RT-qPCR (down-regulation in 38% of ADC and in 81% of SCC samples) were not in concord because different cancer subtypes were present in studied microarray samplings. The disagreement between arrays and RT-qPCR data for NSCLC could also result from non-homogeneous samples with different content of normal cells as well as limited number of specimens and may be statistically non-significant. Although these data are not statistically valid they could reflect important trends and associations.

However there was a rather good agreement between our quantitative results for lung cancer and Oncomine data [17] for two major lung cancer histotypes – ADC and SCC.

It is important to note that Oncomine exploits microarrays based on completely different platform than Clontech Cancer Profiling Arrays. Traditional microarrays (Affymetrix, Agilent) contain a number of various gene probes immobilized on glass slides. Only one cDNA sample can be hybridized with the slide. On the contrary, Clontech Cancer Profiling Arrays contain a number of immobilized cDNA samples from various tumor and normal tissues. Oncomine includes traditional microarrays data making possible genome-wide analysis of a limited number of samples and Cancer Profiling Arrays enable analysis of one gene in many tumors in one experiment.

According to the Clontech Microarray data, the mRNA level increase was observed for several tumor types – uterus, ovary, colon, stomach, thyroid, lung, kidney, and trachea – mainly for non-metastatic tumors. However, also there were frequent cases of the CHL1 mRNA level increase in metastatic tumors, for example, in stomach and lung cancer.

Moreover in four metastases (4I, 14O, 14V, 24K) of 12 available for analysis cases (i.e. when a primary tumor and metastasis for the same patient were accessible) we detected an increased CHL1 expression in metastasis compared to primary tumor (ovary, colon and breast). Similar results were recently reported for the metastasis-associated gene lysyl oxidase (LOX), whose expression was associated both with tumor suppression and tumor progression depending on transformation status [32]. The overexpression of another cell adhesion molecule L1 gene was associated with metastasis in malignant melanoma [34].

Cancer is a fatal disease whereby invasive local tumor growth and metastatic spread to distant vital organs resulting in dormant and/or active growth and inevitable death of patients. Contrary to previous models new evidence suggested that metastatic cells might be created already during initial growth of a primary local tumor. These cells then succeed in cell migration/invasion, embolization, survival in the circulation, arrest in a distant capillary bed, and extravasation into and multiplication within the distant organ parenchyma. Failure at any of these steps could block the entire metastatic process and may lead to “dormant cancer cells and dormant micrometastases”. Surgical removal of the primary tumor might then lead to active growth [38]. Because tumor spread is responsible for the majority of deaths of cancer patients, the development of therapeutic agents that inhibit tumor metastasis is of paramount importance [39], [40], [41], [42], [43], [44], [45], [46], [47].

One of us predicted previously [5] that the cytoplasmic end of CHL1 protein might interact with the cytoskeleton and might induce/regulate filopodia formation driving tumor cell migration and invasion [41], [45], [46]. CHL1 behavior in cancer is thus strikingly similar to L1 [10], [40] and LOX which both work through the actin network.

This study suggested that CHL1 might contribute to cancer invasive growth and metastasis. It might act either as a tumor suppressor (early growth) or oncogene (invasive and metastatic growth, Fig. 1, Table 1). CHL1 therefore could belong to the new rapidly growing category of cancer genes that may function either as TSGs or oncogenes [32], [41], [43], [46], [47], [48]. During initial growth CHL1 is not expressed (silenced) in tumor cells to facilitate in situ tumor growth. Re-expression of CHL1 on the edge of the tumor mass and around tumor vessels could promote migration and local invasive growth and furthermore allow initiating the metastatic process. Thus, our results along with the findings that CHL1 was a mutated candidate cancer-associated gene in colon cancer [1] suggested that this type of recognition receptors may indeed have dual roles in carcinogenesis. The mutations discovered in the extra-cellular part of CHL1 could afford a therapeutic antibody to selectively treat patients [1]. This validates CHL1 as a novel target for personalized immune intervention in cancers expressing mutated CHL1. New therapeutic small inhibitors directed at CHL1 could be effective in restraining new tumor formation from dormant micrometastases.

Our results indicated that the CHL1 gene could be important for the development of major human cancers, and also allowed to suggest a hypothesis on a probable dual role of CHL1, although only for three types of cancer (ovary, colon and breast) supportive data were thus far obtained. A frequent decrease of an expression level was prevalent for 11 of 19 tumor types and statistically significant for breast, colon, rectum, thyroid, kidney and small intestine cancer.

Our data also supported the role of CHL1 as a potentially novel biomarker in the early pathogenesis of two major histological types of renal cancer both CC-RCC and pRCC. Results derived with 7 RCC cell lines suggested them as a potential model system for study of methylation role in CHL1 silencing.

Materials and Methods

Cancer profiling arrays analysis

Cancer Profiling Arrays I and II (154 and 241 samples respectively, overall 19 different types of cancers namely, breast, kidney, rectum, colon, stomach, skin, thyroid, small intestine, bladder, vulva, pancreas, prostate, cervix, testis, lung, ovary, uterus, liver, trachea) purchased from BD Biosciences Clontech (Palo Alto, CA), were used to analyze the expression of the CHL1 gene in normal and tumor tissues. Full sample information for Array I and II is presented in Clontech Catalog: No. 7841-1 and No. 631777 respectively (see “Arrays Information S1”).

We analyzed only informative samples with clear ratio of normal-tumor spots intensity. The information for samples of the Cancer Profiling Array I is presented below.

1. Breast. Most of the tumors are infiltrating ductal (DC), intraductal (IC) and lobular (LC) carcinomas. Stage I: 2Q, 2R, 2S, 2T, 2U, 4S, 4D, 4F. Stage II: 2W, 4E, 4L, 4N. Stage III: 4A, 4B, 4H, 4J, 4M. Coordinates of 18 metastatic (m) tumors are 2B, 2C, 2D, 2E, 2H, 2I, 2J, 2M, 2N, 2P, 4H, 4J, 4L, 4O, 4P, 4R, 4S, 4U. Coordinates of metastases are 4G, 4I, 4K.

2. Uterus. Most of the tumors are adenocarcinomas (ADC). Stage I: 8C, 8F, 8H, 8I, 8J, 8K, 8L, 8M, 8N, 8O, 8P, 8Q, 8R, 8S, 8U, 8X, 8Y, 8BB, 8CC, 8DD, 8EE, 8FF, 10A, 10B, 10C. Stage II: 8T. Coordinates of metastatic tumors 8W and 8AA, Stage III. Coordinates of metastases are 8V, 8Z.

3. Colon. All tumors are ADC. Stage I: 14L. Stage II: 14M, 14P, 14Q, 14AA. Stage III: 14S, 14U, 14V, 14Y, 14BB. Other samples have no information about the Stage. Coordinates of 9 metastatic tumors are 14E, 14N, 14P, 14U, 14W, 14Y, 14CC, 16A, 16C. Coordinates of metastases are 14O, 14T, 14V, 14X.

4. Stomach. Most of the tumors are ADC. There is no information about the Stage. Coordinates of 11 metastatic tumors are 20A, 20B, 20E, 20F, 20H, 20I, 20K, 20S, 20T, 20V, 20X.

5. Ovary. Stage I: 24B, 24D, 24E. Stage II: 24F. Stage III: 24A, 24G, 24H, 24J, 24L. Most tumors are ADC. Coordinates of metastatic tumors are 24J, 24L, 24M, 24N. Coordinates of metastases are 24I, 24K.

6. Cervix. 24X (adenosquamous carcinoma).

7. Lung. Stage I: 28E (squamous cell carcinoma, SCC), 28F (carcinoid), 28H (SCC), 28I (ADC), 28J (ADC), 28K (SCC), 28L (bronchiolo-alveolar adenocarcinoma, BAC), 28M (SCC), 28N (ADC). Unknown Stage: 28A (SCC), 28B (BAC), 28C (SCC), 28D (SCC), 28G (SCC), 28O (m, ADC), 28P (m, BAC), 28Q (SCC), 28R (carcinoid), 28S (ASC), 28T (SCC), 28U (SCC).

8. Kidney. Stage III: 32D (carcinoid). Unknown Stage: 32A (clear cell renal cell carcinoma, CC-RCC), 32B (RCC), 32C (RCC), 32E (RCC), 32F (transitional cell carcinoma), 32G (RCC), 32H (m, RCC), 32I (oncocytoma), 32J (RCC), 32K (RCC), 32L (m, RCC), 32M (RCC), 32N (m, RCC), 32O (RCC), 32P (RCC), 32Q (RCC), 32R (RCC), 32S (RCC), 32T (RCC).

9. Rectum. All tumors are ADC. Stage I: 36G. Stage II: 36J, 36F. Stage III: 36C, 36H, 36I, 36L. Coordinates of 6 metastatic tumors are 36B, 36E, 36L, 36M, 36Q, 36R. Coordinate of metastasis is 36K.

10. Small intestine. 36Y (m, ADC), 36Z (ADC).

11. Thyroid gland. All tumors are papillary ADC, Stage II: 40D. Stage III: 40C, 40E.

12. Prostate. All tumors are ADC. Stage I: 40M.

13. Pancreas. Unknown Stage: 40U (ADC).

The information for samples of the Cancer Profiling Array II is shown below.

1. Breast. Stage I: 6E (DC), 6G (mucinous ADC), 6M (DC). Stage II: 6I (DC), 6K (m, DC), 6L (DC), 6N (DC). Stage III: 6F (m, DC), 6H (m, DC), 6J (m, LC).

2. Uterus. Stage I: 6T (ADC), 6Z (SCC), 6BB (ADC). Stage II: 6X (SCC), 6AA (ADC). Unknown Stage: 6V (SCC).

3. Ovary. Stage I: 10E (papillary serous cystadenoma), 10K (mucinous cystadenocarcinoma). Stage II: 10I (ADC), 10J (leiomyoma). Stage III: 10G (ADC), 10H (CC-ADC), 10M (serous surface papillary carcinoma), 10N (papillary serous cystadenoma). Stage IV: 10L (ADC). Unknown Stage: 10F (leiomyoma).

4. Cervix. Stage I: 10Z (ADC), 10AA (SCC), 10BB (SCC). Stage II: 10V (SCC). Stage III: 10S (m, SCC).

5. Colon. Stage I: 14E (tubulovillous adenoma, other tumors are ADC), 14F. Stage II: 14L. Stage III: 14H (m), 14I (m), 14J, 14K (m), 14M, 14N (m). Stage IV: 14G (m).

6. Rectum. All tumors are ADC. Stage II: 14X, 14Y, 14Z. Stage III: 14T, 14V (m), 14W (m), 14AA (m), 14BB (m). Stage IV: 14S (m), 14U (m).

7. Stomach. Most of the tumors are ADC. Stage II: 18E. Stage III: 18M (m). Stage IV: 18J (m), 18K (m). T3N1Mx: 18F (m).

8. Thyroid gland. Stage II: 18X (papillary adenocarcinoma, PAC), 18Y (PAC), 18AA (PAC), 18BB (follicular ADC). Stage III: 18S (m, ADC), 18W (m, PAC). Stage IV: 18T (m, medullary carcinoma), 18Z (m, PAC). T3N0M0: 18U (PAC).

9. Lung. Stage I: 22I (SCC), 22J (SCC), 22N (SCC). Stage II: 22L (m).

10. Testis. All tumors are seminomas. Stage I: 22V, 22X. Stage II: 22T. Stage III: 22BB (m). Stage IV: 22AA (m).

11. Kidney. Stage II: 26E (RCC), 26M (epithelial nephroblastoma). Stage III: 26H (RCC), 26I (transitional cell carcinoma, TCC), 26J (clear cell ADC), 26K (m, TCC). Stage IV: 26G (m, RCC), 26L (m, RCC). Unknown Stage: 26F (oxyphilic adenoma). T3aNxM0: 26N.

12. Skin. Most of the tumors are melanomas. Stage I: 26S, 26BB. Stage II: 26Y (SCC), 26Z (SCC). Stage III: 26T, 26U, 26W, 26AA. Stage IV: 26V. Unknown Stage: 26X.

13. Bladder. All the tumors are transitional cell carcinomas: Stage II: 30E, 30H. Stage III: 30F, 30I. Stage IV: 30G.

14. Trachea. T4N1M0: 30L (m); T4N0M0: 30M. T3N1M0: 30N (m).

15. Small intestine. Most of the tumors are ADC. Stage II: 30T, 30V. Stage III: 30S (m), 30U, 30X (m). Stage IV: 30Y (m). Unknown Stage: 30W (carcinoid).

16. Vulva. All tumors are SCC. Stage II: 34E, 34F, 34G, 34H. T2aN1M0: 34I (m).

17. Liver. T3N0M0: 34L. T2N0M0: 34M. T2N0M0: 34N.

18. Pancreas. Stage I: 34S (ADC), 34T (ADC). Stage II: 34V (SCC). Stage IV: 34W (m, ADC). T3NxM0: 34Y (cancer of caput of pancreas). Unknown Stage: 34X (carcinoid).

19. Prostate. All tumors are ADC. Stage I: 38E. Unknown Stage: 38F, 38G, 38H.

Totally, there were 90 metastatic samples (with lymph node metastases) and 12 metastases at distant sites.

The arrays were hybridized with specific CHL1 probes labeled with ³²P-α-deoxycytidine triphosphate according to manufacturer's protocol (Clontech). The images were obtained using the Packard Cyclone Storage Phosphor System (PerkinElmer, Shelton, CT).

Tissues specimens and cell lines for RT-qPCR analysis

We analyzed cDNA samples isolated from 30 NSCLC specimens including 16 SCC and 14 ADC; 39 renal cancer specimens including 30 CC-RCC and 9 papillary renal cell carcinomas (pRCC) and 23 breast cancer primary tumors including 19 ductal BC, one lobular BC and 3 mixed lobular-ductal BC. Accordingly, we used 29, 39 and 23 samples from adjacent morphologically normal lung, renal and breast tissues as controls. Six samples of normal lung tissues from healthy donors were used as additional controls. Paired tumor and non-cancerous (normal) matched tissues were received from Blokhin Cancer Research Center, Russian Academy of Medical Sciences. All tumors were classified according to the International TNM Classification system [49], [50]. These samples were obtained from patients after surgical resection of primary lung, renal and breast cancer prior radiation or chemotherapy and stored in liquid nitrogen. The diagnosis was verified by histopathology and only samples containing 70–80% or more tumor cells were used in the study. The samples were collected in accordance to the guidelines issued by the Ethics Committee of Blokhin Cancer Research Center, Russian Academy of Medical Sciences (Moscow). All patients gave written informed consent that is available upon request. The Ethics Committee of Blokhin Cancer Research Center, Russian Academy of Medical Sciences specifically approved this study. The study was done in accordance with the principles outlined in the Declaration of Helsinki.

Clear cell renal cancer cell lines TK10, TK164, KRC/Y, KH39, HN4, Caki1and Caki2 from MTC collection were also analyzed in the study. Mix of six samples of normal lung tissues from healthy donors were used as controls.

RNA and cDNA preparation for RT-qPCR

Total RNA was isolated from tumor and matched normal tissues, renal cancer cell lines using RNeasy Mini Kit (Qiagen, Netherlands) according to the manufacturer's instructions. Purified RNA was quantified using NanoDrop® ND-1000 spectrophotometer (NanoDrop Technologies Inc., USA), and the quality was determined by Bioanalyzer 2100 (Agilent Technologies, USA). All RNA samples were treated with DNase I and cDNA was synthesized using MMLV reverse transcriptase and random hexamers according to standard manufacturer's protocol (Fermentas, Lithuania).

Analysis of CHL1 expression by RT-qPCR

RT-qPCR was performed essentially as described previously [51] in total volume of 25 µl in triplicate. The sequences of primers (F and R) and TaqMan probes (Z) are shown in Table 5. Final concentrations of primers and probes for target and reference genes were as follow: CHL1 primers – 300 nM, probe – 400 nM; GAPDH primers – 300 nM, probe – 150 nM, RPN1 primers – 300 nM, probe – 150 nM, GUSB primers – 300 nM, probe – 250 nM, B2M primers and probe – 300 nM.

[Figure omitted. See PDF.]

Table 5. Primers and probes sequences for CHL1 and reference genes for RT-qPCR.

https://doi.org/10.1371/journal.pone.0015612.t005

The thermocycler conditions were 10 min at 95°C, then 45 two-step cycles 15 s at 95°C and 60 s at 60°C (ABI 7000 PRISM Sequence Detection System, Applied Biosystems). The sequences of the amplicons were verified by sequencing in a 3730 DNA Analyzer (Applied Biosystems).

RT-qPCR data were analyzed using the relative quantification method, or -method [52], [53] based on target and reference gene, in tumor (T) and normal (N) samples comparison. Relative mRNA level (R) is calculated by formula: , where n is mRNA copy number in tumor sample relative normal sample. . Here is E – reaction efficiency, – quantification cycle, , effective quantification cycle.

Finally, .

The reaction efficiencies (E) were calculated as described [33] and their values were 85±9% (lung), 93±11% (breast), 87±10% (renal) for CHL1; 89±8% (lung) for GAPDH; 86±11% (breast) for B2M; 87±9% (renal) for RPN1; 81±9% (renal) for GUSB in tumor and normal samples. The level of mRNA expression change was calculated as 1/R and reflected the n-fold factor by which the mRNA content decreased or increased in the tumor compared to normal tissue.

All preliminary validation steps were performed according to Gene quantification resource (www.gene-quantification.de). As reference genes, we used GAPDH and RPN1 for NSCLC, B2M for BC, and GUSB and RPN1 for RCC. The relative inner variability between GAPDH, B2M, GUSB and RPN1 mRNA levels was not higher than 2-fold in tumor (T) and normal (N) tissues, therefore 2-fold and more mRNA changes for CHL1 were considered as significant for all types of studied tumors.

All calculations performed using our program ATG (Analysis of Transcription of Genes) compatible with Relative Quantification (RQ) software (Applied Biosystems). The program is designed for mathematical processing of RT-qPCR data in paired normal-tumor samples. The general features of AEGIS consist of estimating reaction efficiencies with three different methods (including calculating efficiency from kinetic curves) and taking it into account for further data analysis, using the relative comparative method [52] with several reference genes and estimating their variability. Also this program allows checking adjusted normal samples variability and using as paired normal sample for each tumor sample as one or several normal samples for all tumor samples.

Statistical analysis

For validation of CHL1 expression differences between normal and tumor samples and also between tumor samples with and without metastases we applied Fisher exact test and χ² criterion. Nonparametric Wilcoxon test was used for analyzing RT-qPCR data to compare mRNA content differences of the CHL1 gene and reference genes for the same samples. We used nonparametric Kruskal-Wallis and Mann-Whitney tests for the rank order differences between average relative levels of mRNA in different groups of samples. P-values<0.05 were considered statistically significant. All statistical procedures were performed using BioStat software [54].

Supporting Information

[Figure omitted. See PDF.]

Arrays Information S1.

This archive contains detailed information about localization and clinical characteristics of all the tumors (histological type, tumor size, stage, presence of metastases etc.).

https://doi.org/10.1371/journal.pone.0015612.s001

(DOC)

Acknowledgments

Authors thank R. Nalewaik for technical assistance.

Author Contributions

Conceived and designed the experiments: MIL VNS ERZ. Performed the experiments: AVK EAA IVP SVI. Analyzed the data: VNS GSK AAD ERZ. Contributed reagents/materials/analysis tools: VNS AVK EAA EAB TTK. Wrote the manuscript: MIL VNS GSK AAD ERZ.

Citation: Senchenko VN, Krasnov GS, Dmitriev AA, Kudryavtseva AV, Anedchenko EA, Braga EA, et al. (2011) Differential Expression of CHL1 Gene during Development of Major Human Cancers. PLoS ONE 6(3): e15612. https://doi.org/10.1371/journal.pone.0015612

References

1. Sjoblom T, Jones S, Wood LD, Parsons DW, Lin J, et al. (2006) The consensus coding sequences of human breast and colorectal cancers. Science 314: 268–274.T. SjoblomS. JonesLD WoodDW ParsonsJ. Lin2006The consensus coding sequences of human breast and colorectal cancers.Science314268274

2. Greenman C, Stephens P, Smith R, Dalgliesh GL, Hunter C, et al. (2007) Patterns of somatic mutation in human cancer genomes. Nature 446: 153–158.C. GreenmanP. StephensR. SmithGL DalglieshC. Hunter2007Patterns of somatic mutation in human cancer genomes.Nature446153158

3. Ley TJ, Mardis ER, Ding L, Fulton B, McLellan MD, et al. (2008) DNA sequencing of a cytogenetically normal acute myeloid leukaemia genome. Nature 456: 66–72.TJ LeyER MardisL. DingB. FultonMD McLellan2008DNA sequencing of a cytogenetically normal acute myeloid leukaemia genome.Nature4566672

4. Ross DT, Scherf U, Eisen MB, Perou CM, Rees C, et al. (2000) Systematic variation in gene expression patterns in human cancer cell lines. Nat Genet 24: 227–235.DT RossU. ScherfMB EisenCM PerouC. Rees2000Systematic variation in gene expression patterns in human cancer cell lines.Nat Genet24227235

5. Wei MH, Karavanova I, Ivanov SV, Popescu NC, Keck CL, et al. (1998) In silico-initiated cloning and molecular characterization of a novel human member of the L1 gene family of neural cell adhesion molecules. Hum Genet 103: 355–364.MH WeiI. KaravanovaSV IvanovNC PopescuCL Keck1998In silico-initiated cloning and molecular characterization of a novel human member of the L1 gene family of neural cell adhesion molecules.Hum Genet103355364

6. Angeloni D, Lindor NM, Pack S, Latif F, Wei MH, et al. (1999) CALL gene is haploinsufficient in a 3p- syndrome patient. Am J Med Genet 86: 482–485.D. AngeloniNM LindorS. PackF. LatifMH Wei1999CALL gene is haploinsufficient in a 3p- syndrome patient.Am J Med Genet86482485

7. Frints SG, Marynen P, Hartmann D, Fryns JP, Steyaert J, et al. (2003) CALL interrupted in a patient with non-specific mental retardation: gene dosage-dependent alteration of murine brain development and behavior. Hum Mol Genet 12: 1463–1474.SG FrintsP. MarynenD. HartmannJP FrynsJ. Steyaert2003CALL interrupted in a patient with non-specific mental retardation: gene dosage-dependent alteration of murine brain development and behavior.Hum Mol Genet1214631474

8. Sakurai K, Migita O, Toru M, Arinami T (2002) An association between a missense polymorphism in the close homologue of L1 (CHL1, CALL) gene and schizophrenia. Mol Psychiatry 7: 412–415.K. SakuraiO. MigitaM. ToruT. Arinami2002An association between a missense polymorphism in the close homologue of L1 (CHL1, CALL) gene and schizophrenia.Mol Psychiatry7412415

9. Manderson EN, Birch AH, Shen Z, Mes-Masson AM, Provencher D, et al. (2009) Molecular genetic analysis of a cell adhesion molecule with homology to L1CAM, contactin 6, and contactin 4 candidate chromosome 3p26pter tumor suppressor genes in ovarian cancer. Int J Gynecol Cancer 19: 513–525.EN MandersonAH BirchZ. ShenAM Mes-MassonD. Provencher2009Molecular genetic analysis of a cell adhesion molecule with homology to L1CAM, contactin 6, and contactin 4 candidate chromosome 3p26pter tumor suppressor genes in ovarian cancer.Int J Gynecol Cancer19513525

10. Stoeck A, Schlich S, Issa Y, Gschwend V, Wenger T, et al. (2006) L1 on ovarian carcinoma cells is a binding partner for Neuropilin-1 on mesothelial cells. Cancer Lett 239: 212–226.A. StoeckS. SchlichY. IssaV. GschwendT. Wenger2006L1 on ovarian carcinoma cells is a binding partner for Neuropilin-1 on mesothelial cells.Cancer Lett239212226

11. Rokman A, Baffoe-Bonnie AB, Gillanders E, Fredriksson H, Autio V, et al. (2005) Hereditary prostate cancer in Finland: fine-mapping validates 3p26 as a major predisposition locus. Hum Genet 116: 43–50.A. RokmanAB Baffoe-BonnieE. GillandersH. FredrikssonV. Autio2005Hereditary prostate cancer in Finland: fine-mapping validates 3p26 as a major predisposition locus.Hum Genet1164350

12. Qin YR, Fu L, Sham PC, Kwong DL, Zhu CL, et al. (2008) Single-nucleotide polymorphism-mass array reveals commonly deleted regions at 3p22 and 3p14.2 associate with poor clinical outcome in esophageal squamous cell carcinoma. Int J Cancer 123: 826–830.YR QinL. FuPC ShamDL KwongCL Zhu2008Single-nucleotide polymorphism-mass array reveals commonly deleted regions at 3p22 and 3p14.2 associate with poor clinical outcome in esophageal squamous cell carcinoma.Int J Cancer123826830

13. Chen LC, Matsumura K, Deng G, Kurisu W, Ljung BM, et al. (1994) Deletion of two separate regions on chromosome 3p in breast cancers. Cancer Res 54: 3021–3024.LC ChenK. MatsumuraG. DengW. KurisuBM Ljung1994Deletion of two separate regions on chromosome 3p in breast cancers.Cancer Res5430213024

14. Wei MH, Latif F, Bader S, Kashuba V, Chen JY, et al. (1996) Construction of a 600-kilobase cosmid clone contig and generation of a transcriptional map surrounding the lung cancer tumor suppressor gene (TSG) locus on human chromosome 3p21.3: progress toward the isolation of a lung cancer TSG. Cancer Res 56: 1487–1492.MH WeiF. LatifS. BaderV. KashubaJY Chen1996Construction of a 600-kilobase cosmid clone contig and generation of a transcriptional map surrounding the lung cancer tumor suppressor gene (TSG) locus on human chromosome 3p21.3: progress toward the isolation of a lung cancer TSG.Cancer Res5614871492

15. Protopopov A, Kashuba V, Zabarovska VI, Muravenko OV, Lerman MI, et al. (2003) An integrated physical and gene map of the 3.5-Mb chromosome 3p21.3 (AP20) region implicated in major human epithelial malignancies. Cancer Res 63: 404–412.A. ProtopopovV. KashubaVI ZabarovskaOV MuravenkoMI Lerman2003An integrated physical and gene map of the 3.5-Mb chromosome 3p21.3 (AP20) region implicated in major human epithelial malignancies.Cancer Res63404412

16. Senchenko VN, Liu J, Loginov W, Bazov I, Angeloni D, et al. (2004) Discovery of frequent homozygous deletions in chromosome 3p21.3 LUCA and AP20 regions in renal, lung and breast carcinomas. Oncogene 23: 5719–5728.VN SenchenkoJ. LiuW. LoginovI. BazovD. Angeloni2004Discovery of frequent homozygous deletions in chromosome 3p21.3 LUCA and AP20 regions in renal, lung and breast carcinomas.Oncogene2357195728

17. Wu C, Orozco C, Boyer J, Leglise M, Goodale J, et al. (2009) BioGPS: an extensible and customizable portal for querying and organizing gene annotation resources. Genome Biol 10: R130.C. WuC. OrozcoJ. BoyerM. LegliseJ. Goodale2009BioGPS: an extensible and customizable portal for querying and organizing gene annotation resources.Genome Biol10R130

18. Rhodes DR, Yu J, Shanker K, Deshpande N, Varambally R, et al. (2004) ONCOMINE: a cancer microarray database and integrated data-mining platform. Neoplasia 6: 1–6.DR RhodesJ. YuK. ShankerN. DeshpandeR. Varambally2004ONCOMINE: a cancer microarray database and integrated data-mining platform.Neoplasia616

19. Safran M, Solomon I, Shmueli O, Lapidot M, Shen-Orr S, et al. (2002) GeneCards 2002: towards a complete, object-oriented, human gene compendium. Bioinformatics 18: 1542–1543.M. SafranI. SolomonO. ShmueliM. LapidotS. Shen-Orr2002GeneCards 2002: towards a complete, object-oriented, human gene compendium.Bioinformatics1815421543

20. Sausville EA, Holbeck SL (2004) Transcription profiling of gene expression in drug discovery and development: the NCI experience. Eur J Cancer 40: 2544–2549.EA SausvilleSL Holbeck2004Transcription profiling of gene expression in drug discovery and development: the NCI experience.Eur J Cancer4025442549

21. Yusenko MV, Kuiper RP, Boethe T, Ljungberg B, van Kessel AG, et al. (2009) High-resolution DNA copy number and gene expression analyses distinguish chromophobe renal cell carcinomas and renal oncocytomas. BMC Cancer 9: 152.MV YusenkoRP KuiperT. BoetheB. LjungbergAG van Kessel2009High-resolution DNA copy number and gene expression analyses distinguish chromophobe renal cell carcinomas and renal oncocytomas.BMC Cancer9152

22. Beroukhim R, Brunet JP, Di Napoli A, Mertz KD, Seeley A, et al. (2009) Patterns of gene expression and copy-number alterations in von-hippel lindau disease-associated and sporadic clear cell carcinoma of the kidney. Cancer Res 69: 4674–4681.R. BeroukhimJP BrunetA. Di NapoliKD MertzA. Seeley2009Patterns of gene expression and copy-number alterations in von-hippel lindau disease-associated and sporadic clear cell carcinoma of the kidney.Cancer Res6946744681

23. Pyeon D, Newton MA, Lambert PF, den Boon JA, Sengupta S, et al. (2007) Fundamental differences in cell cycle deregulation in human papillomavirus-positive and human papillomavirus-negative head/neck and cervical cancers. Cancer Res 67: 4605–4619.D. PyeonMA NewtonPF LambertJA den BoonS. Sengupta2007Fundamental differences in cell cycle deregulation in human papillomavirus-positive and human papillomavirus-negative head/neck and cervical cancers.Cancer Res6746054619

24. Kaiser S, Park YK, Franklin JL, Halberg RB, Yu M, et al. (2007) Transcriptional recapitulation and subversion of embryonic colon development by mouse colon tumor models and human colon cancer. Genome Biol 8: R131.S. KaiserYK ParkJL FranklinRB HalbergM. Yu2007Transcriptional recapitulation and subversion of embryonic colon development by mouse colon tumor models and human colon cancer.Genome Biol8R131

25. Ki DH, Jeung HC, Park CH, Kang SH, Lee GY, et al. (2007) Whole genome analysis for liver metastasis gene signatures in colorectal cancer. Int J Cancer 121: 2005–2012.DH KiHC JeungCH ParkSH KangGY Lee2007Whole genome analysis for liver metastasis gene signatures in colorectal cancer.Int J Cancer12120052012

26. Hendrix ND, Wu R, Kuick R, Schwartz DR, Fearon ER, et al. (2006) Fibroblast growth factor 9 has oncogenic activity and is a downstream target of Wnt signaling in ovarian endometrioid adenocarcinomas. Cancer Res 66: 1354–1362.ND HendrixR. WuR. KuickDR SchwartzER Fearon2006Fibroblast growth factor 9 has oncogenic activity and is a downstream target of Wnt signaling in ovarian endometrioid adenocarcinomas.Cancer Res6613541362

27. Bhattacharjee A, Richards WG, Staunton J, Li C, Monti S, et al. (2001) Classification of human lung carcinomas by mRNA expression profiling reveals distinct adenocarcinoma subclasses. Proc Natl Acad Sci U S A 98: 13790–13795.A. BhattacharjeeWG RichardsJ. StauntonC. LiS. Monti2001Classification of human lung carcinomas by mRNA expression profiling reveals distinct adenocarcinoma subclasses.Proc Natl Acad Sci U S A981379013795

28. Talbot SG, Estilo C, Maghami E, Sarkaria IS, Pham DK, et al. (2005) Gene expression profiling allows distinction between primary and metastatic squamous cell carcinomas in the lung. Cancer Res 65: 3063–3071.SG TalbotC. EstiloE. MaghamiIS SarkariaDK Pham2005Gene expression profiling allows distinction between primary and metastatic squamous cell carcinomas in the lung.Cancer Res6530633071

29. Chen X, Leung SY, Yuen ST, Chu KM, Ji J, et al. (2003) Variation in gene expression patterns in human gastric cancers. Mol Biol Cell 14: 3208–3215.X. ChenSY LeungST YuenKM ChuJ. Ji2003Variation in gene expression patterns in human gastric cancers.Mol Biol Cell1432083215

30. Sorlie T, Tibshirani R, Parker J, Hastie T, Marron JS, et al. (2003) Repeated observation of breast tumor subtypes in independent gene expression data sets. Proc Natl Acad Sci U S A 100: 8418–8423.T. SorlieR. TibshiraniJ. ParkerT. HastieJS Marron2003Repeated observation of breast tumor subtypes in independent gene expression data sets.Proc Natl Acad Sci U S A10084188423

31. Radvanyi L, Singh-Sandhu D, Gallichan S, Lovitt C, Pedyczak A, et al. (2005) The gene associated with trichorhinophalangeal syndrome in humans is overexpressed in breast cancer. Proc Natl Acad Sci U S A 102: 11005–11010.L. RadvanyiD. Singh-SandhuS. GallichanC. LovittA. Pedyczak2005The gene associated with trichorhinophalangeal syndrome in humans is overexpressed in breast cancer.Proc Natl Acad Sci U S A1021100511010

32. Erler JT, Bennewith KL, Nicolau M, Dornhofer N, Kong C, et al. (2006) Lysyl oxidase is essential for hypoxia-induced metastasis. Nature 440: 1222–1226.JT ErlerKL BennewithM. NicolauN. DornhoferC. Kong2006Lysyl oxidase is essential for hypoxia-induced metastasis.Nature44012221226

33. Anedchenko EA, Dmitriev AA, Krasnov GS, Kondrat'eva TT, Kopantsev EP, et al. (2008) [Down-regulation of RBSP3/CTDSPL, NPRL2/G21, RASSF1A, ITGA9, HYAL1 and HYAL2 genes in non-small cell lung cancer]. Mol Biol (Mosk) 42: 965–976.EA AnedchenkoAA DmitrievGS KrasnovTT Kondrat'evaEP Kopantsev2008[Down-regulation of RBSP3/CTDSPL, NPRL2/G21, RASSF1A, ITGA9, HYAL1 and HYAL2 genes in non-small cell lung cancer].Mol Biol (Mosk)42965976

34. Thies A, Schachner M, Moll I, Berger J, Schulze HJ, et al. (2002) Overexpression of the cell adhesion molecule L1 is associated with metastasis in cutaneous malignant melanoma. Eur J Cancer 38: 1708–1716.A. ThiesM. SchachnerI. MollJ. BergerHJ Schulze2002Overexpression of the cell adhesion molecule L1 is associated with metastasis in cutaneous malignant melanoma.Eur J Cancer3817081716

35. Zecchini S, Bianchi M, Colombo N, Fasani R, Goisis G, et al. (2008) The differential role of L1 in ovarian carcinoma and normal ovarian surface epithelium. Cancer Res 68: 1110–1118.S. ZecchiniM. BianchiN. ColomboR. FasaniG. Goisis2008The differential role of L1 in ovarian carcinoma and normal ovarian surface epithelium.Cancer Res6811101118

36. Gavert N, Conacci-Sorrell M, Gast D, Schneider A, Altevogt P, et al. (2005) L1, a novel target of beta-catenin signaling, transforms cells and is expressed at the invasive front of colon cancers. J Cell Biol 168: 633–642.N. GavertM. Conacci-SorrellD. GastA. SchneiderP. Altevogt2005L1, a novel target of beta-catenin signaling, transforms cells and is expressed at the invasive front of colon cancers.J Cell Biol168633642

37. Murphy J, Bustin SA (2009) Reliability of real-time reverse-transcription PCR in clinical diagnostics: gold standard or substandard? Expert Rev Mol Diagn 9: 187–197.J. MurphySA Bustin2009Reliability of real-time reverse-transcription PCR in clinical diagnostics: gold standard or substandard?Expert Rev Mol Diagn9187197

38. Ding L, Ellis MJ, Li S, Larson DE, Chen K, et al. (2010) Genome remodelling in a basal-like breast cancer metastasis and xenograft. Nature 464: 999–1005.L. DingMJ EllisS. LiDE LarsonK. Chen2010Genome remodelling in a basal-like breast cancer metastasis and xenograft.Nature4649991005

39. Mareel M, Leroy A (2003) Clinical, cellular, and molecular aspects of cancer invasion. Physiological Reviews 83: 337–376.M. MareelA. Leroy2003Clinical, cellular, and molecular aspects of cancer invasion.Physiological Reviews83337376

40. Husemann Y, Geigl JB, Schubert F, Musiani P, Meyer M, et al. (2008) Systemic spread is an early step in breast cancer. Cancer Cell 13: 58–68.Y. HusemannJB GeiglF. SchubertP. MusianiM. Meyer2008Systemic spread is an early step in breast cancer.Cancer Cell135868

41. Pantel K, Alix-Panabieres C, Riethdorf S (2009) Cancer micrometastases. Nature Reviews Clinical Oncology 6: 339–351.K. PantelC. Alix-PanabieresS. Riethdorf2009Cancer micrometastases.Nature Reviews Clinical Oncology6339351

42. Entschladen F, Drell TLt, Lang K, Joseph J, Zaenker KS (2004) Tumour-cell migration, invasion, and metastasis: navigation by neurotransmitters. Lancet Oncol 5: 254–258.F. EntschladenTLt DrellK. LangJ. JosephKS Zaenker2004Tumour-cell migration, invasion, and metastasis: navigation by neurotransmitters.Lancet Oncol5254258

43. Silva JM, Ezhkova E, Silva J, Heart S, Castillo M, et al. (2009) Cyfip1 Is a Putative Invasion Suppressor in Epithelial Cancers. Cell 137: 1047–1061.JM SilvaE. EzhkovaJ. SilvaS. HeartM. Castillo2009Cyfip1 Is a Putative Invasion Suppressor in Epithelial Cancers.Cell13710471061

44. Strell C, Entschladen F (2008) Extravasation of leukocytes in comparison to tumor cells. Cell Commun Signal 6: 10.C. StrellF. Entschladen2008Extravasation of leukocytes in comparison to tumor cells.Cell Commun Signal610

45. Machesky LM, Tang HR (2009) Actin-based protrusions: promoters or inhibitors of cancer invasion? Cancer Cell 16: 5–7.LM MacheskyHR Tang2009Actin-based protrusions: promoters or inhibitors of cancer invasion?Cancer Cell1657

46. Maddaluno L, Verbrugge SE, Martinoli C, Matteoli G, Chiavelli A, et al. (2009) The adhesion molecule L1 regulates transendothelial migration and trafficking of dendritic cells. J Exp Med 206: 623–635.L. MaddalunoSE VerbruggeC. MartinoliG. MatteoliA. Chiavelli2009The adhesion molecule L1 regulates transendothelial migration and trafficking of dendritic cells.J Exp Med206623635

47. Sanada M, Suzuki T, Shih LY, Otsu M, Kato M, et al. (2009) Gain-of-function of mutated C-CBL tumour suppressor in myeloid neoplasms. Nature 460: 904–908.M. SanadaT. SuzukiLY ShihM. OtsuM. Kato2009Gain-of-function of mutated C-CBL tumour suppressor in myeloid neoplasms.Nature460904908

48. Kashuba VI, Li J, Wang F, Senchenko VN, Protopopov A, et al. (2004) RBSP3 (HYA22) is a tumor suppressor gene implicated in major epithelial malignancies. Proc Natl Acad Sci U S A 101: 4906–4911.VI KashubaJ. LiF. WangVN SenchenkoA. Protopopov2004RBSP3 (HYA22) is a tumor suppressor gene implicated in major epithelial malignancies.Proc Natl Acad Sci U S A10149064911

49. Spiessl B, International Union against Cancer (1989) TNM atlas : illustrated guide to the TNM/pTNM-classification of malignant tumours. Berlin, New York: Springer-Verlag. B. SpiesslInternational Union against Cancer1989TNM atlas : illustrated guide to the TNM/pTNM-classification of malignant tumoursBerlin, New YorkSpringer-Verlagxiv, 341

50. Travis WD, Colby TV, Corrin B, Shimosato Y, Brambilla E, et al. (1999) Histological Typing of Lung and Pleural Tumours (WHO. World Health Organization. International Histological Classification of Tumours). Berlin: Springer. 156 p.WD TravisTV ColbyB. CorrinY. ShimosatoE. Brambilla1999Histological Typing of Lung and Pleural Tumours (WHO. World Health Organization. International Histological Classification of Tumours)BerlinSpringer156

51. Senchenko VN, Anedchenko EA, Kondratieva TT, Krasnov GS, Dmitriev AA, et al. (2010) Simultaneous down-regulation of tumor suppressor genes RBSP3/CTDSPL, NPRL2/G21 and RASSF1A in primary non-small cell lung cancer. BMC Cancer 10: 75.VN SenchenkoEA AnedchenkoTT KondratievaGS KrasnovAA Dmitriev2010Simultaneous down-regulation of tumor suppressor genes RBSP3/CTDSPL, NPRL2/G21 and RASSF1A in primary non-small cell lung cancer.BMC Cancer1075

52. Livak KJ, Schmittgen TD (2001) Analysis of relative gene expression data using real-time quantitative PCR and the 2(-Delta Delta C(T)) Method. Methods 25: 402–408.KJ LivakTD Schmittgen2001Analysis of relative gene expression data using real-time quantitative PCR and the 2(-Delta Delta C(T)) Method.Methods25402408

53. Lindlof A, Brautigam M, Chawade A, Olsson O, Olsson B (2008) Evaluation of combining several statistical methods with a flexible cutoff for identifying differentially expressed genes in pairwise comparison of EST sets. Bioinform Biol Insights 2: 215–237.A. LindlofM. BrautigamA. ChawadeO. OlssonB. Olsson2008Evaluation of combining several statistical methods with a flexible cutoff for identifying differentially expressed genes in pairwise comparison of EST sets.Bioinform Biol Insights2215237

54. Glantz SA (2005) Primer of biostatistics. New York: McGraw-Hill Medical Pub. 520 p.SA Glantz2005Primer of biostatisticsNew YorkMcGraw-Hill Medical Pub520