| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
ORIGINAL ARTICLES |
From the Departments of Surgery (JCM, CJS), Molecular and Medical Genetics (JWH, JWD), and Clinical Epidemiology (SB), Samuel Lunenfeld Research Institute and Mount Sinai Hospital, University of Toronto, Toronto, Canada.
Correspondence: Address correspondence and reprint requests to: Carol J. Swallow, 600 University Avenue, Suite 1224, Toronto, Ontario, Canada M5G 1X5; Fax: 416-586-8392.
 |
ABSTRACT |
|---|
 TOP ABSTRACT INTRODUCTIONMETHODSRESULTSDISCUSSIONREFERENCES |
|---|
Methods: In this study, SAK expression was evaluated in a series of sporadic human colorectal cancer specimens (n = 74) and compared with that of PLK. Expression was assessed by reverse transcription-polymerase chain reaction.
Results: In the majority of cases, both SAK and PLK were more highly expressed in tumor tissue than in adjacent normal intestinal mucosa. Levels of SAK and PLK expression in tumor relative to paired normal mucosa correlated directly with patient age and with each other but did not correlate with tumor stage. These results suggest a mechanism for augmented disruption of mitotic regulation in older patients.
Conclusions: The polo family mitotic regulators SAK and PLK are both aberrantly expressed in colorectal cancer. The potential prognostic significance of SAK and PLK expression in colorectal cancer will be evaluated in the future.
Key Words: Polo kinase • Mitosis • Cell cycle • Colorectal cancer
|
INTRODUCTION |
|---|
|
TOPABSTRACTINTRODUCTIONMETHODSRESULTSDISCUSSIONREFERENCES |
|---|
PLK expression has been analyzed in four types of cancer: non-small-cell lung cancer (n = 111),9 squamous cell carcinoma of the head and neck (n = 89),10 esophageal cancer (n = 49), and gastric cancer (n = 75).11 Comparing tumor tissue with normal mucosa, PLK was overexpressed in the majority of cancers. Increased PLK expression correlated with adverse clinical outcome in the first three tumor types and was in fact an independent prognostic indicator in multivariate analysis. SAK expression has not previously been evaluated in a large series of tumors of any type.
The aim of this study was to characterize the expression of SAK and PLK in human colorectal cancers (CRCs). Expression was assessed in 74 cases of CRC by reverse transcription-polymerase chain reaction (RT-PCR), with tumor tissue compared with paired normal mucosa samples. Expression was also assessed in 15 cases of liver metastases derived from colorectal primary tumors and in adjacent normal liver tissue. SAK and PLK were expressed more highly in colorectal tumors (T) relative to normal mucosa (NM) in 66.2% and 77.1% of cases, respectively. Expression of both SAK and PLK in T:NM was directly correlated with the age of the patient. SAK and PLK expression were also correlated with each other, but not with tumor stage. CRC metastases to the liver did not show a significant difference in SAK or PLK gene expression when compared with primary tumors. These data indicate that SAK and PLK are both aberrantly expressed in the majority of human CRCs. Because polo-like kinases have been implicated in spindle assembly and temporal control of anaphase, the dysregulated expression observed in this study could be responsible for mitotic errors and contribute to the aneuploid phenotype commonly seen in CRCs, as well as in many other solid tumors.
|
METHODS |
|---|
|
TOPABSTRACTINTRODUCTIONMETHODSRESULTSDISCUSSIONREFERENCES |
|---|
Solutions
RPMI 1640 medium containing penicillin G (167 IU/ml) and streptomycin sulfate (100 µg/ml), and phosphate-buffered saline, pH 7.2, were obtained from the media preparation facility at Samuel Lunenfeld Research Institute, Mount Sinai Hospital (Toronto, Canada). Fetal bovine serum, qualified, came from Gibco BRL, Life Technologies.
Clinical Database
The following clinical variables were recorded in a prospective manner as gathered from clinical charts, radiology reports, and pathology reports: hospital medical record number, age at diagnosis, stage of disease according to the American Joint Committee on Cancer (AJCC) system, date of operation, location of tumor within the bowel, and presence of lymphovascular, venous, or perineural invasion. These were compiled into a database specific to either CRC or liver metastasis patients.
Tissue Samples
Primary colorectal adenocarcinomas resected sequentially by one surgeon (C.J.S.) were routinely banked. Tumor samples were obtained from the invasive edge to ensure viability. This was confirmed by examination of paraffin sections of the same invasive edge specimen processed simultaneously. Samples of normal mucosa were taken from the same operative specimen, several centimeters distant from the tumor, by dissecting along the muscularis mucosa layer. Hepatic metastases from CRC, resected sequentially by one surgeon (S.G.), were routinely banked. Samples of tumor and uninvolved normal liver were obtained from surgical specimens. Tissues were immediately flash-frozen in liquid nitrogen, then stored at -70 °C in either a colorectal tumor or liver metastasis bank.
Cell Lines and Culture
The HT 29 CRC cell line was obtained from the American Type Culture Collection (Rockville, MD). Cells were maintained at 
75% confluence in RPMI 1640 medium supplemented with 10% fetal bovine serum at 37°C in a 5% CO2 atmosphere.
RNA Extraction and Synthesis of Complementary DNA
Total RNA was extracted from tissue samples and from cultures of HT 29 cells by using Trizol reagent according to the manufacturer’s instructions. Cultured cells were approximately 75% confluent at the time of RNA extraction. RNA samples were stored at -70°C until ready for use.
Complementary DNA (cDNA) was synthesized in an 8-µl reaction mixture containing 400 ng of total cellular RNA, 1.6 µl 5 x First Strand Buffer, 1.6 µl of RNase-free dNTPs, .8 µl of .1 M dithiothreitol, .8 µl (40 ng) of random primers, .1 µl (3.86 U) of RNAguard, and .2 µl (40 U) of Moloney murine leukemia virus-RT. The mixture was incubated at 37°C for 60 minutes, denatured at 94°C for 4 minutes, and stored at -20°C. Before use as a template for the PCR, the mixture was diluted 1/5 with autoclaved deionized water.
PCR
Forward and reverse primers were designed to span intron/exon boundaries to prevent the amplification of any genomic DNA contaminating the cDNA mixture. The primers for SAK (1488F, 5'-AATCAAGCACTCTCCAATC-3'; 1665R, 5'-TGTGTCCTTCTGCAAATC-3') and porphobilinogen deaminase (PBGD) (283F, 5'-GAGAAGAATGAAGTGGACCT-3'; 536R, 5'-AGGTTTCCCCGAATACTCC-3') yield PCR products of 177 and 253 bp, respectively. The primers for PLK, which generate a fragment of 194 bp, were 1476F, 5'-CTTGATGAAGAAGATCACCC-3' and 1670R, 5'-TGGAAGAAGTTGATCTGGAC-3'.
PCR was performed in a 20-µl reaction mixture containing 1.6 µl of cDNA template, 2 µl of 10 x PCR buffer, 5.625 to 6.25 mM of MgCl2, .8 mM of dNTPs, .3 µl each of forward and reverse primers (each final .45 µM) and 1.25 U of AmpliTaq Gold polymerase. PCR was performed in a PTC-100TM (Manufacturer, City, ST) programmable thermal controller as follows: denaturation at 94°C for 10 minutes, then 25 to 36 cycles of denaturation at 94°C for 30 seconds, annealing at 52 to 53°C for 30 seconds, and extension at 72°C for 40 seconds. Reactions were then cooled to 4°C. PCR products were separated by electrophoresis on 4% to 20% tris base boric acid, EDTA (TBE) gels in a Novex (San Diego, CA) Xcell IITM apparatus at 100 V for 2.5 hours and visualized by ethidium bromide staining, and gels were photographed onto PolaPan 665 PNTM film (Polaroid, Cambridge, MA). Quantitation was performed by laser densitometry with a Computing DensitometerTM and ImageQuantTM software (Molecular Dynamics, Menlo Park, CA).
Criteria for Generating Valid PCR Results
Each sample was analyzed over a range of PCR cycles to ensure quantitative evaluation within the logarithmic phase of the PCR reaction. Only those reactions that had not plateaued were used in the analysis. For a sample to be valid, densitometry values for each gene must have increased by 50% over two cycles, and densitometry values <10 were discarded. To normalize for the amount of RNA and to control for the quality of the RNA samples, each reaction also contained primers for a constitutively expressed gene, PBGD, for comparison with the level of SAK or PLK expression.12 Also, the CRC cell line, HT 29, which was found to express both SAK and PLK, was used as a positive control in each set of PCR reactions.
For each sample of tumor or normal tissue, the ratio of expression of the gene of interest to that of PBGD was averaged over a range of cycles in the logarithmic phase of amplification, and this number was then normalized to the level of messenger RNA (mRNA) in the HT 29 cell line. The latter step acted as a control for day-to-day variability of the PCR and thus permitted comparison of SAK and PLK expression levels between reactions. For 13 PCR reactions performed on colorectal cDNA templates, the mean SAK:PBGD ratio for the HT 29 cell line was .661. The SE of this mean was .033, indicating minimal between-reaction variability and high precision. A T:NM gene expression ratio was then constructed for each case of CRC (Fig. 1). The majority of the samples were analyzed at least twice, yielding reproducible results. For each case of primary CRC, six RT-PCR values were calculated as listed in Table 1. RT-PCR densitometry values obtained from the analysis of liver tissue samples (metastatic tumor (M) and adjacent normal liver (NL)) were used to calculate four values called SAK NL, SAK M, PLK NL, and PLK M.
|
|
Analysis of Data
The data obtained consisted of the gene expression ratio (i.e., SAK:PBGD) and the ratio of gene expression in T relative to NM (T:NM). Data are expressed as mean ± SEM unless otherwise stated. Normal probability plots were constructed, and the Anderson-Darling normality test was applied. Where data were not normally distributed, a log transformation was applied. Statistical analyses were performed on both untransformed and transformed data with parametric tests where normally distributed, and nonparametric tests where not.
The Pearson coefficient was used to evaluate correlations between continuous variables. Student’s t-test was used to evaluate differences in means between two groups; one-way analysis of variance (ANOVA) was used when there were more than two groups. Nonparametric equivalents, which compare median values, were the Mann-Whitney U-test, the Wilcoxon signed rank test (for paired data), and the Kruskal-Wallis test. The 95% confidence interval (CI) for an individual observation was calculated as mean ± 1.96(SD). The 95% CI for the mean was calculated as mean ± 1.96(SD/
n), where n = number of cases in the cohort.
To examine the relationship between SAK and PLK expression among the CRC cases, a 95% confidence ellipse was constructed. The equation of this ellipse incorporates the mean and SD of the log of the T:NM ratio for each gene, as well as the coefficient for the correlation between all log T:NM ratios of SAK and PLK. All statistical tests assumed a level of significance of .05 and were performed with MINITABTM software for Windows 95 (Minitab, Inc., State College, PA).
|
RESULTS |
|---|
|
TOPABSTRACTINTRODUCTIONMETHODSRESULTSDISCUSSIONREFERENCES |
|---|
|
RT-PCR yielded valid results for both SAK and PLK expression in 67 of the 74 cases. In the remaining seven cases, valid results were obtained for only SAK in four and only PLK in three. In the cases in which valid results for one of the genes could not be obtained, gene expression was detected, but reactions failed to amplify PCR products in an exponential fashion such that the criteria for validity (as outlined in Methods) were met for only one cycle. No particular clinicopathologic property was apparent in these seven cases.
Liver metastases from a total of 15 patients, preserved in a liver metastasis bank from October 1997 to January 2000, were also analyzed by RT-PCR for expression of SAK and PLK. The mean age at resection of metastatic disease was 59.8 years (range, 33–77 years), and this group consisted of 11 men and 4 women. In one case, the colonic primary and the metastatic liver disease were banked simultaneously during the course of the same operation.
Expression of SAK in CRC
SAK mRNA levels were evaluated in 71 colorectal Ts and compared with paired NM specimens. SAK expression is measured relative to the expression of PBGD, a constitutively expressed gene. Log transformation of the SAK:PBGD ratio was performed to normalize the data and facilitate statistical analysis. To illustrate the range of SAK expression within each tissue type, histograms were constructed to show the frequency of all log values of the SAK:PBGD ratio generated from analysis of the NM specimens (Fig. 2a) and the colorectal T (Fig. 2b). Mean SAK expression and the 95% CI for an individual case (95% CI for xi) are shown. The log data fit a normal distribution (Anderson-Darling test of normality).
|
Three samples of NM showed very low SAK expression, with SAK:PBGD ratios outside the 95% CI. The only commonality among these was that all three were specimens from the sigmoid colon. However, when SAK expression levels in NM were grouped according to location within the bowel (right colon, other colon, or rectum), there was no difference in the log SAK:PBGD ratio among the three sites (P = .391, ANOVA; Table 3). Two tumor samples demonstrated very low expression of SAK (outside the 95% CI for xi). Both of these tumors were in male patients and were right-sided T3 lesions. One tumor showed very high SAK expression (SAK:PBGD = 2.8810), which was outside the 95% CI for xi.
|
A ratio of T:NM was generated for each individual patient. The distribution of all log values of the T:NM ratio for the 71 cases is shown in Fig. 2c. The mean SAK T:NM ratio was 1.312 ± .102. To aid in identifying the outlying cases, the log of the T:NM ratio for each individual case was plotted in Fig. 3a. Positive bars represent cases for which SAK gene expression was greater in the T relative to paired NM. In 66.2% of cases (47 of 71), expression was greater in T relative to NM. The 95% CI for the log SAK T:NM ratio for an individual case was -.3655 to .4580, and this defined three cases as outliers, two with low and one with high T:NM ratios.
|
In normal liver tissue, mean SAK expression was .298 ± .047; it was barely detectable in several of the specimens. The range of expression (plotted as log of SAK:PBGD) in the 15 normal liver samples is shown in Fig. 4a. In the 15 hepatic metastases, mean SAK expression was 1.153 ± .120 (log of SAK:PBGD shown in Fig. 4b).
|
To further illustrate the differences in SAK expression among the tissues studied, the mean value of log SAK and the associated 95% CI of the mean were plotted for each tissue (Fig. 5a). Although the 95% CI of the mean for log SAK NM overlaps with that for log SAK T, the 95% CI for the log T:NM ratio excludes 0. This confirms that the ratio of SAK expression in colorectal T relative to paired NM is significantly >1 (P < .05). SAK expression in normal liver tissue was significantly lower than in NM (median log SAK expression, -.6076 and .0091, respectively; P < .00001, Mann-Whitney U-test), consistent with greater expression of this mitotic regulator in a more rapidly proliferating cell population.
|
As a group, metastatic CRC lesions did not differ significantly in SAK expression when compared with primary tumors (log SAK, .033 ± .042 in liver metastases vs. .078 ± .022 in colorectal primary tumors, n = 15 and 71, respectively; P = .36, Student’s t-test), although the comparison between these tissues is limited by the fact that they were not paired. For one case, in which the patient underwent simultaneous resection of the primary tumor and metastatic liver lesion and a valid direct comparison could be made between the two, SAK expression in the liver metastasis was 1.4729 (log value, .1682), and in the primary colon tumor, it was 1.9249 (log value, .2844). The relationship between the paired metastatic and primary tumor was similar to that observed between the two tumor groups overall. Both NM and liver parenchyma from this individual, who was also much younger than the mean patient age, displayed extremely increased levels of SAK.
In 71 patients with colorectal primary tumors, SAK expression was analyzed for any association with the following clinicopathologic variables: age, sex, pathologic stage, or location of tumor. SAK expression in either NM, colorectal T tissue, or the ratio between the two (T:NM) did not vary significantly with patient sex (log SAK T:NM, .066 ± .038 for men and .053 ± .035 for women; P = .81, Student’s t-test) or pathologic stage (P = .493 for log SAK T:NM, ANOVA; Table 4). There was a trend toward increased SAK expression in more distal tumors, with rectal cancers showing the highest expression (P = .240, ANOVA); however, NM from the rectum also showed somewhat higher SAK expression, so the T:NM ratio was not significantly different among these locations (Table 3).
|
SAK expression in colorectal T relative to paired NM was higher in older patients (Table 5). Patients <60 years of age (n = 19) had a mean log SAK T:NM ratio of -.022 ± .040, whereas for those 60 years or older (n = 52), the value was .090 ± .031. This was a statistically significant difference (P = .035, Student’s t-test). It seemed that this increase in the log T:NM ratio with age was due to both a decrease in SAK expression in NM with age (log SAK, .061 ± .060 for age <60 years vs. .005 ± .030 for age 
60 years; P = .41, Student’s t-test) and an increase in expression in colorectal Ts with age (log SAK, .040 ± .054 for age <60 years vs. .092 ± .023 for age 60 years; P = .38, Student’s t-test).
|
Expression of PLK in CRC
PLK mRNA levels were measured in 70 colorectal specimens, evaluating both T tissue and paired NM. The expression of PLK in NM was quite homogeneous among all 70 samples evaluated, more so than for SAK. Log PLK values are shown as a histogram in Fig. 6a. Two cases were noted to be outliers on the basis of the log PLK expression lying outside the 95% CI for xi. Colorectal tumor specimens were more variable in their expression of PLK. The distribution of PLK expression within the 70 tumor samples is shown in Fig. 6b. Seven cases were observed to lie outside the 95% CI for xi. The mean T:NM ratio for PLK was 1.615 ± .125. Fig. 6c illustrates the frequency of all log T:NM values obtained for PLK expression in the 70 cases examined. The data describing PLK expression (log values for NM, T, and T:NM) all fit a normal distribution (Anderson-Darling normality test).
|
The ratio of PLK expression in the colorectal Ts relative to paired NM, for all 70 cases, is shown in Fig. 3b. The log of the T:NM ratio is plotted on the y-axis; positive bars represent cases for which PLK gene expression was greater in the T than in paired NM, which was 77.1% of cases (54 of 70). Five cases were identified as outliers relative to the 95% CI for an individual case.
For the 12 cases of normal liver tissue and hepatic metastases of CRC evaluated for PLK expression by RT-PCR, the distribution of PLK expression (relative to PBGD) is shown in Fig. 4c and 4d, respectively. Normal liver tissue expressed very little PLK, which was barely detectable even by this sensitive method (.2091 ± .0145). By contrast, PLK was readily detectable in liver metastases and had a fairly narrow range of expression in these samples (1.2025 ± .0791).
The mean level of expression of PLK in each tissue and the associated 95% CI of the mean are shown in Fig. 5b. This illustrates the differences in PLK expression in nonmalignant and malignant tissues, as well as in different types of benign tissue (colonic mucosa vs. liver parenchyma). The 95% CI of the mean log PLK T:NM does not include 0, indicating that the T:NM ratio is significantly >1 (P < .05). Thus, PLK expression is greater in CRC relative to paired NM. Consistent with previous studies, which have found PLK expression to be closely linked to the rate of proliferation of the tissue, expression was much lower in normal liver parenchyma when compared with NM (mean log PLK:PBGD, -.693 ± .033 and -.056 ± .023, n = 12 and 70, respectively; P < .00001, Student’s t-test). No difference in PLK expression was observed between primary T and hepatic metastases of CRC (median log PLK expression .0781 vs. .0939, n = 70 and 12, respectively; P = .778, Mann-Whitney U-test), although again these samples were not paired.
An examination of the relationship of PLK expression to the documented clinicopathologic variables revealed no difference in PLK expression in NM, colorectal T, or in T:NM ratio on the basis of sex (log PLK T:NM, .133 ± .030 for men and .145 ± .048 for women; P = .83, Student’s t-test) or stage (P = .703 for log PLK T:NM, ANOVA; Table 4). In contrast to SAK, PLK expression was statistically significantly lower in rectal tumors when compared with cancers of the proximal or more distal colon (P = .015 for log PLK T, ANOVA; Table 3). When cancers of these two regions (right-sided and other colon) were pooled, this difference remained significant (log PLK T, .116 ± .025 for colon T vs. -.042 ± .066 for rectal T, n = 54 and 16, respectively; P = .037, Student’s t-test). PLK expression in NM also decreased somewhat with the more distal location of specimens, so the T:NM ratio of PLK expression in rectal cancer was not significantly different from that of colonic specimens (Table 3).
There was a significant correlation between age at the time of resection and the log of the PLK T:NM ratio (Pearson coefficient, .287; P = .016). To investigate whether an age-dependent increase in PLK expression in T or a decrease in NM was responsible, PLK expression in these tissues was stratified by 10-year age groups. A trend toward increased PLK expression in T with increasing age was observed (P = .056, ANOVA), with a more notable increase in samples from patients 60 years of age or older. Indeed, PLK expression was significantly higher in the tumors from these patients (n = 50) than from the patients <60 years old (n = 20) (median log PLK:PBGD, .1183 and -.0098, respectively; P = .0014, Mann-Whitney U-test). Because there was no significant change in PLK expression in NM with advancing age, the T:NM ratio for PLK expression was markedly higher in older patients (mean log PLK expression, .184 ± .036 in those 60 years vs. .028 ± .039 in those <60 years old; P = .005, Student’s t-test, Table 5).
Correlation Between SAK and PLK Expression
SAK expression (T:NM) was found to correlate with PLK expression (Pearson coefficient, .311; P = .01). To examine the relationship between SAK and PLK expression in individual cases of CRC, a scatter plot was constructed. For 67 CRC cases, the logs of the T:NM ratio for SAK and PLK were plotted on the x- and y-axes, respectively (Fig. 7). The Pearson correlation coefficient of the log values of SAK and PLK was .526 (P = .0001); in 52 of 67 cases (77.6%), the patterns of gene expression were congruent. Most cases (n = 41) fell in the top right quadrant, indicating that both SAK and PLK were more highly expressed in the T compared with paired NM. Twelve of the 67 cases showed increased PLK expression but reduced SAK expression in the T versus NM, whereas only 3 cases showed increased SAK and reduced PLK. In 11 cases, both SAK and PLK were reduced in T compared with paired NM.
|
A 95% confidence ellipse was constructed by using the mean and SD of the 67 log T:NM gene expression ratios for each of SAK, and PLK, and the correlation coefficient. This defined five cases as outliers. Three of these fell in the upper right quadrant, indicating that expression of each gene, SAK and PLK, was higher than predicted on the basis of the data for the group. One case fell in the bottom right quadrant and the other in the bottom left quadrant. Although there were no statistically significant findings when these cases were compared with the remaining 62 cases on the basis of clinicopathologic data, all five outliers were T3, and the proportion of node-positive cases (80.0%) was greater than that in the remaining 62 cases (46.8%). Furthermore, these five outliers seemed to have more advanced disease than the population falling within the 95% confidence ellipse (80.0% had stage III or IV disease, vs. 56.5% in the remaining 62 patients).
For 12 cases of metastatic liver disease, data on expression of both SAK and PLK were available. The Pearson correlation coefficient for the log values of SAK and PLK in this group was .515 (P = .086), demonstrating only a trend toward a correlation between the two.
|
DISCUSSION |
|---|
|
TOPABSTRACTINTRODUCTIONMETHODSRESULTSDISCUSSIONREFERENCES |
|---|
Both SAK and PLK were both found to be more highly expressed in colorectal T, when compared with NM, in the majority of cases studied. Similar findings have been observed for PLK in series of non-small-cell lung cancers, head and neck squamous cell cancers, and esophageal cancers.9–11 In these studies, overexpression of PLK was correlated with poor clinical prognosis. Because the tissue specimens examined in this study were banked relatively recently, mature clinical data on patient outcome are not yet available, and thus the prognostic significance of SAK and PLK expression in colorectal cancer will be evaluated in a future study. Disease stage is of course predictive of clinical outcome in CRC. Our data indicate no significant correlation between expression of either SAK or PLK and T stage. Indeed, neither SAK nor PLK expression was different in a group of CRCs metastatic to liver from that in the group of primary tumors we examined. Lack of correlation with stage, however, implies the potential for independent prognostic value.
Because levels of PLK and SAK mRNA correlate with proliferative activity in normal human tissues, and because overexpression of PLK has been noted in several tumor types, investigators have suggested that polo kinase expression could reflect cellular proliferation.13 SAK gene expression is observed in mitotic and meiotic cells in embryonic and adult mouse tissues.4 The gastrointestinal mucosa is a rapidly renewing tissue, and so it has a high proliferative rate, whereas normal liver contains a large proportion of quiescent hepatocytes, capable of dividing only when the appropriate signal is present. Both SAK and PLK were found to be more highly expressed in NM relative to normal liver, and these findings support earlier studies correlating expression with proliferation.
The T:NM ratio of polo gene expression was increased in samples from older patients when compared with younger individuals. This was true for both SAK and PLK. In the case of PLK, an increase in gene expression in the tumors of the individuals 60 years old was responsible for the increased ratio, whereas for SAK, the increased T:NM ratio seemed to be multifactorial. SAK expression was both reduced in NM from those 60 years old and increased in the tumors from these patients, although neither of these differences was statistically significant. Ly et al.14 have recently provided evidence to suggest that progressive alteration in the expression of genes involved in cell division occurs with age. In their study, mRNA levels were measured in actively dividing fibroblasts from young, middle-aged, and older humans. Genes whose products are involved in cell-cycle progression accounted for a significant proportion of the genes that demonstrated consistent differences among age groups. Although SAK was not among the genes assessed by these authors, other chromosomal processing and spindle assembly genes were downregulated in the older populations. PLK was among these, showing a 2.8-fold reduction in expression in middle age (37 years old) and a 3.0-fold reduction in old age (93 years old), relative to young age (8 years old). This progressive decrease in cell cycle regulators with age could potentially contribute to the associated loss of tissue regenerative capacity. One explanation for why we did not observe a significant decrease in PLK expression in normal tissue from older individuals in our study could be that we assessed intestinal epithelial tissue, which has a high baseline proliferative activity. Presumably, polo-like kinase expression in the colon is maintained with age, and thus these cells may not be subject to the degree of downregulation of mitotic genes that occurs in fibroblast cells. Also, very few of our patients were under 40 years old (n = 6), potentially minimizing our ability to detect any age-related changes in normal tissues. In this light, the reduction in SAK expression observed in NM with increased age becomes more remarkable. The age-related increase in both SAK and PLK expression in tumor tissue could be caused by the loss of a common regulatory mechanism with aging.
We observed a significantly lower level of PLK expression in rectal tumors relative to colon tumors. SAK expression showed the opposite pattern, with a trend to higher mRNA levels in cancers of the rectum. These findings raise two issues, the first being that factors regulating expression of these two polo-like kinases may differ. Indirect evidence supporting differential regulation also includes the observation of a more restricted pattern of SAK expression in human tissues (relative to PLK), and our recent evidence indicating that Sak was not essential for mouse embryo development until approximately day 7.5.5 It is interesting to note that in our study, of all 70 colorectal tumors with assessable PLK expression, the 2 with the lowest levels were rectal cancers treated with neoadjuvant chemoradiotherapy, which may have significantly decreased their proliferative rates. The second point to address is the possibility that cancers of the colon and rectum biologically differ from each other. There is evidence that different etiological factors may be important in the development of cancer at different sites along the large bowel,15,16 and this is also supported by studies on genetic alterations in colorectal tumors.17 One study has reported a higher rate of aneuploidy among rectal cancers relative to colon cancers,18 and aneuploidy has been correlated with lower survival than diploidy.19 In a recent large population-based study, both the pattern of tumor growth (expanding vs. infiltrating) and the presence or absence of lymphocytic infiltration were observed to differ significantly between colon (n = 169) and rectal (n = 106) cancers.20 Tumors with an expanding type of margin are associated with a better prognosis than those showing tumor cell infiltration into the bowel wall.21,22 It is therefore intriguing to consider that a state of relatively low PLK and high SAK, as observed in the rectal tumors in our series, might relate to these other genetic and biological properties that seem to be more prevalent in rectal cancers in general.
Possible mechanisms for the increased gene expression that was observed in the majority of CRCs in this study include the loss of a factor that negatively regulates expression of the gene or the gain of chromosomal material that results in multiple copies of the gene. Little is known about the factors that regulate expression of the polo-like kinase genes; however, the murine Sak promoter contains consensus binding sites for two transcriptional repressors, Gfi-1 and Oct-1.2 Loss of one of these negative regulators, or mutation of the binding site itself, could lead to overexpression of the gene. The latter has been observed in in vitro studies.2 Perhaps most interesting to consider is an increase in copy number of the gene, with a resulting increase in gene dosage. Galitski et al.23 have shown in yeast that expression of certain genes is ploidy induced, whereas other genes are ploidy repressed. The existence of such regulation in humans might mean that the polyploidy often seen in tumor cells, which is usually viewed as a consequence of aberrant cell cycle control, may also be a cause of disturbed cell behavior. The human SAK gene is located on chromosome 4q. Gains of this chromosomal arm were reported in 118 of 2210 solid tumors in a recent review of comparative genomic hybridization analyses of 27 human tumor types.24 Gain of 4q was notably common among pituitary tumors (5 of 23 cases, or 21.7%) and neuroendocrine tumors (7 of 20 cases, or 35%), but it was not a particularly common feature of colorectal tumors. One limitation of these findings is that variations in subregions within the same chromosomal arm were not assessed. The use of comparative genomic hybridization or fluorescent in situ hybridization studies on tumor and normal tissues may elucidate the mechanism of SAK upregulation in CRC.
What might be the functional significance of increased expression of SAK in colorectal tumors? On the basis of our recent finding that Sak-null murine embryos do not develop beyond day 7.5, with a preponderance of cells stuck in the late phases of mitosis,5 we postulate that SAK kinase activity is required for chromosomal segregation at anaphase, proper cytokinesis, or both. Thus, if overexpression of the SAK gene leads to an imbalance of kinase versus phosphatase activities and this disrupts one of these critical processes, aneuploidy could result. Furthermore, if increased SAK expression leads to inactivation of the spindle checkpoint, this could promote the acquisition of further gene defects or mutations needed for progression of the disease. Thus, SAK overexpression could represent a novel mechanism for the generation of chromosomal instability.
|
Acknowledgments |
|---|
|
Footnotes |
|---|
Received for publication March 16, 2001. Accepted for publication June 16, 2001.
|
REFERENCES |
|---|
|
TOPABSTRACTINTRODUCTIONMETHODSRESULTSDISCUSSIONREFERENCES |
|---|
This article has been cited by other articles:
 |
|
 |
|
  T. L. Schmit and N. Ahmad Regulation of mitosis via mitotic kinases: new opportunities for cancer management Mol. Cancer Ther., July 1, 2007; 6(7): 1920 - 1931. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 I. Perez de Castro, G. de Carcer, and M. Malumbres A census of mitotic cancer genes: new insights into tumor cell biology and cancer therapy Carcinogenesis, May 1, 2007; 28(5): 899 - 912. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 E. Knock, L. Deng, Q. Wu, D. Leclerc, X.-l. Wang, and R. Rozen Low Dietary Folate Initiates Intestinal Tumors in Mice, with Altered Expression of G2-M Checkpoint Regulators Polo-Like Kinase 1 and Cell Division Cycle 25c Cancer Res., November 1, 2006; 66(21): 10349 - 10356. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
P. J. Gray Jr, D. J. Bearss, H. Han, R. Nagle, M.-S. Tsao, N. Dean, and D. D. Von Hoff Identification of human polo-like kinase 1 as a potential therapeutic target in pancreatic cancer Mol. Cancer Ther., May 1, 2004; 3(5): 641 - 646. [Abstract] [Full Text] [PDF]
|
|
| ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
