Background
The transformation of normal cells into cancer cells is generally believed to rely on genetic or epigenetic changes in the genome. These changes often activate oncogenes and/or inactivate tumor suppressor genes. The recently described
N-Myc Downstream Regulated Gene 2 (
NDRG2) is down-regulated in a variety of human tumors, including colorectal, liver and thyroid cancers as well as glioblastoma [
1‐
5]. Furthermore, elevated expression of
NDRG2 in tumors correlates with an improved prognosis in gastric cancer, high-grade glioma and hepatocellular carcinomas [
2,
6,
7]. Several
in vitro studies have demonstrated a reduced cell growth when NDRG2 was over-expressed in cell-lines lacking endogenous expression [
5,
8]. These data suggest that
NDRG2 might have a role in suppressing carcinogenesis.
NDRG2 is one of four members belonging to the
NDRG family. In contrast to
NDRG1, however,
NDRG2 does not appear to be down-regulated by N-myc
in vivo [
9]. All four NDRG family members possess an α/β-hydrolase fold domain but whether or not they have enzymatic activity is presently unclear.
Northern blot analysis was previously used to establish the mRNA expression pattern of
NDRG2 in normal tissues and it has been shown to have the highest expression in brain, heart, skeletal muscle, kidney and liver and lowest expression in tissues such as colon, spleen, placenta and lung [
5,
10,
11]. How
NDRG2 is regulated is not well understood, but recently it was shown that Myc, via Miz-1, can repress
NDRG2 expression [
12] and that the Wilms' tumor gene 1 protein as well as p53 can induce expression of
NDRG2 [
13,
14]. Furthermore, studies have shown that
NDRG2 transcription can be subjected to an epigenetic repression through promoter methylation, leading to a decreased expression [
15,
16]. Finally, NDRG2 protein contains several phosphorylation sites [
17] suggesting a completely separate regulatory mechanism via a phosphorylation-dephosphorylation cycle.
Based on our current understanding, NDRG2 is a candidate tumor suppressor gene. To further understand how NDRG2 is involved in carcinogenesis, we need to know more about the regulation, distribution and function of NDRG2. In this paper, we present a broad profile of NDRG2 expression in human cancers with focus on breast cancer. Our data from 19 different cancers indicates that NDRG2 is up-regulated in around 8% of all tumors examined, unaltered in 62% of all cases and that approximately a third of the tumor samples show down-regulation of NDRG2 compared to the corresponding normal tissue. By quantifying the level of NDRG2 mRNA in 35 breast cancer samples, we observed a statistically significant down-regulation of NDRG2 in tumor samples compared to normal tissue. We also quantified the level of MYC in the same breast cancer samples to investigate whether or not a high level of MYC expression could be responsible for the observed NDRG2 down-regulation. In our hands, we observed a positive correlation, indicating that MYC and NDRG2 are co-regulated. Thus, another regulatory pathway than repression by Myc appears to reduce NDRG2 mRNA levels in breast cancer.
Methods
Cancer Profiling Array II
The Cancer Profiling Array II (CPA II, Clontech) was hybridised with 50 ng of radioactively labelled NDRG2 probe according to the manufacturer's instructions. The 460 bp NDRG2 probe was generated by PCR using the primers 5'CTCACTCTGTGGAGACACCAT3' and 5'GGGTGATATCACCTCCACGCT3'. The hybridised array was exposed to a phosphorimaging screen for 24 hours and the intensity of each spot was quantified using ImageQuant (Molecular Dynamics). The CPA II consists of paired cDNA samples generated from the total RNA of normal and tumor tissue. Because the array is normalised for several housekeeping genes, quantification of the hybridisation signal provides an estimate of relative transcript abundance. Furthermore, the CPA II was hybridised with a human ubiquitin control cDNA probe provided by the manufacturer and exposed for 48 hours. All samples showed uniform expression of ubiquitin confirming the normalisation by the manufacturer (data not shown).
TissueScan mRNA quantification
Breast and thyroid cancer TissueScan qPCR Arrays (Origene Technologies, Rockville MD) were used to quantify and normalise the level of
NDRG2 and
MYC to the expression level of
β-actin. The TissueScan Arrays consist of pre-normalized cDNA from both normal and tumor tissues. The commercial supplier has confirmed that the removal of human tissue was performed with the approval of an Institutional Review Board ethics committee within each medical center. The medical centers from which the tissue was banked represented major academic medical centers within the US. Tissue and data collection conform to Federal requirements and informed consent was obtained from all human subjects. According to the manufacturer, samples were selected based on tumor content (minimally 50% tumor) as determined by microscopic pathology analysis. Based on a pathology report the tumors were partly classified, following the tumor-node-metastasis (TNM) staging system, according to their size/extent (TI-TIVC). The normal samples were taken from patients diagnosed with cancer, but the tissues were harvested from normal regions. The quantification was performed on an ABI 7300 sequence detector as previously described [
1] or as recommended by the manufacturer. The primers and probes for
NDRG2 have previously been described [
1]. Primers and probes for
β-actin (part no. 4310881) and
MYC (assay ID Hs00153408 m1) were obtained from Applied Biosystems.
In a validation experiment using a control sample, a 2-fold dilution series was produced and assayed for
NDRG2, MYC and
β-actin expression as described in the comparative C
t method [
18]. When C
t values were plotted against log dilution it was shown that the assays were quantitative over a range of 128-fold dilution for the
NDRG2/β-actin assay and over a 512-fold dilution for the
MYC/β-actin assay, using a threshold of 0.2 for
β-actin, 0.07 for
NDRG2 and 0.15 for
MYC.
NDRG2, MYC and β-actin mRNAs were quantified in separate wells in triplicate. The standard deviation of the same sample (internal and positive control) in separate experiments was 20%, 10% and 20% for NDRG2, MYC and β-actin respectively in thyroid tissue, and 27%, 10% and 16% for NDRG2, MYC and β-actin respectively in breast tissue, indicating a day-to-day variation of the assay. Negative controls (without conversion of RNA into cDNA) and positive controls were included in all sets.
For both thyroid and breast cancer samples, some data were excluded either due to C
t values being too low and outside the standard curve or because the mRNA levels were below the detection level. In general, the levels of both
NDRG2 and
MYC were lower in thyroid tissue than breast tissue samples compared to
β-actin, which was equally expressed in both tissues. The distribution of gender and age among cases whose tissue was included in the final analysis is presented in Tables
1 and
2.
Table 1
Characteristics of cases with thyroid gland cancer
n
| 4 | 9 | 3 | 6 | 2 | 1 |
Females
| 4 | 7 | 2 | 3 | 1 | - |
Males
| - | 2 | 1 | 3 | 1 | 1 |
Mean age
| 47.3 | 40.3 | 59.3 | 65.5 | 48.5 | 80.0 |
SD
| 16.3 | 15.0 | 3.2 | 9.4 | 5.0 | - |
Table 2
Characteristics of cases with breast cancer
n
| 7 | 10 | 13 | 7 | 8 | 3 |
Females
| 7 | 10 | 13 | 7 | 8 | 3 |
Males
| - | - | - | - | - | - |
Mean age
| 43.1 | 63.3 | 58.2 | 53.6 | 55.3 | 57.7 |
SD
| 9.6 | 10.6 | 13.9 | 6.0 | 13.9 | 9.6 |
Statistical analysis
GraphPad Prism 4 was used for all the statistic calculations and p values < 0.05 were considered significant for all tests. The paired two-tailed t-test was used for comparisons of the CPA data and p-values were subsequently corrected for type I errors using the Benjamini-Hochberg method. For the quantitative RT-PCR data for the breast and thyroid tissues, the Mann-Whitney test was used to compare mean values between normal and tumor samples, since the variances in those samples were different. The non-parametric Kruskal-Wallis test was used on the different tumor stages TI through TIIIC groups, since the variances in the groups were different. For correlation analyses we used Spearman's correlation test due to the number of samples and since the data did not follow a Gaussian distribution.
Discussion
In order to understand the initiation and progression of cancer, it is important to identify the genes and mechanisms involved. In this paper, we have examined the expression of
NDRG2, a recently discovered gene with a proposed tumor suppressor activity, using a commercially available cancer profiling array covering 19 different human cancer forms. One third of all tumor samples analysed showed down-regulation of
NDRG2 mRNA levels by 2-fold or more compared to corresponding normal tissue, suggesting that either the regulators of
NDRG2 are generally affected or that
NDRG2 is often inactivated in tumors. Several tumor types (cervix, colon, testis and thyroid) from the CPA showed a statistically significant reduction of
NDRG2 mRNA when analysed separately (t-test), but when corrected together for type I errors (Benjamini-Hochberg correction) covering the whole array, none of them retained significance. However, analysing both colon and thyroid cancer in a more quantitative manner and using larger sample sets, we and others have demonstrated a clear and significant decrease in
NDRG2 mRNA levels in both cancer types [
1,
3]. Similarly, we observed lower levels in all 3 liver cancer samples, which was recently confirmed in a larger study [
2]. From the CPA results, we would also like to highlight tissues such as cervix and testis as candidate tissues with a potential decrease of
NDRG2 mRNA between normal and tumor tissues. The sample size for each cancer type on the CPA was between 3 to 10 paired normal and tumor samples. A conclusion from the CPA is therefore not straightforward and we only used the array data as a survey-type study to identify candidate tissues for subsequent quantification studies.
In the case of both breast and thyroid cancer, we analysed 5-7 normal and around 40 tumor samples by quantitative RT-PCR. All samples showed a uniform level of β-actin mRNA, indicating equivalent amounts of mRNA in each sample. However, in some cases the
NDRG2 levels were too low (outside the standard curve) or even undetectable. In the case of thyroid cancer, 2 normal and the majority of tumor samples (34) were discarded due to too low or undetectable mRNA levels. Thus, our statistical analysis was based on 3 normal and 9 tumor samples. Although we observed a significant difference between normal and tumor tissues with regard to
NDRG2 mRNA levels, we are aware that it is not a satisfying sample size to draw any conclusions from, but as mentioned previously our data are consistent with recent findings by others [
3].
Using quantitative RT-PCR, we demonstrated that
NDRG2 mRNA levels were statistically significantly reduced in breast cancer tissue when compared to normal tissue. Liu
et al. have previously reported by semi-quantitative PCR that there was a reduction in
NDRG2 mRNA levels in 5 out of 21 breast cancer samples tested, compared to normal tissue [
16]. However, our results are based on quantitative real-time PCR and a slightly larger sample set (n = 35). Differences at the mRNA level for
NDRG2 are likely to be reflected at the protein level [
12] and our results are in agreement with other studies showing that NDRG2 protein is reduced in breast cancer [
19].
Two alternative protein isoforms of NDRG2 have been described. In the long form of NDRG2, a short stretch of 14 amino acids is inserted N-terminal to the conserved α/β-hydrolase fold domain, as a result of alternative splicing [
11,
20]. It would be interesting in future studies to examine whether the different splice forms of
NDRG2 are under differential regulation, which could explain the mixed expression results observed in some cancer types and shed light on the possible function of this protein.
In 2006 it was shown that the Myc oncoprotein could interact with the
NDRG2 promoter via Miz-1 and that the level of Myc is inversely correlated with the level of NDRG2 in colon cancer cell-lines undergoing differentiation [
12]. Since
MYC is known to be either over-expressed or amplified in many human cancers [
21], it could be responsible for the decreased levels of
NDRG2 mRNA. Zhao
et al. demonstrated that there is an inverse correlation between the levels of
NDRG2 and
MYC mRNA in thyroid cancer [
3]. Whether this also applies to other cancers is not known, and we therefore quantified
MYC and
NDRG2 mRNA in a set of breast cancer samples. However, we observed only a weak correlation between
MYC and
NDRG2 mRNA expression in normal and tumor samples from breast tissue, indicating that
MYC and
NDRG2 are co-regulated. This makes it unlikely that Myc acts as a repressor for
NDRG2 gene expression in breast cancer. One possibility is that Myc acts as a transcriptional activator for
NDRG2 by recruiting co-factors other than Miz-1 [
22]. Alternatively, the observed correlation of
MYC and
NDRG2 mRNA expression in breast tissue may result from a shared regulatory mechanism of these genes in the normal and tumor samples tested.
In summary, we have demonstrated that
NDRG2 mRNA levels are significantly reduced in thyroid and breast cancer, and that
NDRG2 mRNA levels will be interesting to quantify in cervix and testis cancer. Overexpression of NDRG2 protein in cancer cell-lines results in a marked reduction in cell proliferation, although the precise mechanisms are unclear [
5,
8,
23]. Since studies in other cancer types demonstrate that high
NDRG2 mRNA expression correlates with improved prognosis [
2,
6,
7], future studies could be aimed at developing therapeutic approaches to increase NDRG2 expression in tumor tissue.
Competing interests
The authors declare that they have no competing interests.
Authors' contributions
CM and AL conceived the idea of the study. RL and AL carried out the Cancer Profiling Array analysis. AL made all the quantitative RT-PCR, supervised by JB. CM and AL drafted the manuscript. LV and JB contributed to the manuscript and gave statistical advice. All authors contributed to interpretation and discussion of the results and read and approved the final version.