High-level mRNA quantification of proliferation marker pKi-67 is correlated with favorable prognosis in colorectal carcinoma

Abstract

Purpose

The present study retrospectively examines the expression of pKi-67 mRNA and protein in colorectal carcinoma and their correlation to the outcome of patients.

Methods

Immunohistochemistry and quantitative RT-PCR were used to analyze the expression of pKi-67 in 43 archival specimens of patients with curatively resected primary colorectal carcinoma, who were not treated with neo-adjuvant therapy.

Results

We determined a median pKi-67 (MIB-1) labeling index of 31.3% (range 10.3–66.4%), and a mean mRNA level of 0.1769 (ΔC_T: range 0.01–0.69); indices and levels did not correlate. High pKi-67 mRNA ΔC_T values were associated with a significantly favorable prognosis, while pKi-67 labeling indices were not correlated to prognostic outcome. A multivariate analysis of clinical and biological factors indicated that tumor stage (UICC) and pKi-67 mRNA expression level were independent prognostic factors.

Conclusion

Quantitatively determined pKi-67 mRNA can be a good and new prognostic indicator for primary resected colorectal carcinoma.

Introduction

The Ki-67 antigen, pKi-67, one of the most commonly used markers of proliferating cells is a large nuclear protein, which was initially described by Gerdes et al. 1983. The original Ki-67 antibody was derived from mice immunized with cell nuclei from the cell line L428 and is a prototype for other antibodies which also identify epitopes of pKi-67, e.g., MIB-1 and MIB-3 (Cattoretti et al. 1992). Further investigations about the distribution of the protein during the cell cycle revealed that it was exclusively expressed in proliferating cells (G₁, S-, G₂-, and M-phase; Gerdes et al. 1984; Verheijen et al. 1989). Cloning of the pKi-67 cDNA and of the gene locus revealed a very large gene (about 12.5 kb) consisting of 15 exons (Schluter et al. 1993; Duchrow et al. 1996). Exon 13 encodes the largest part of pKi-67 (6845 bp) and contains sixteen 16 tandemly repeated elements, each encoding for about 122 amino acids. These “Ki-67 repeats” are between 43 and 62% identical and contain the 66 bp-”Ki-67 motifs,” which are from 72 up to 100% identical. The epitopes FKEL and FKELF, which are recognized by Ki-67 and MIB-1, respectively, are encoded in this region (Kubbutat et al. 1994). As shown by yeast and mammalian two-hybrid screenings it is known that the “Ki-67 motif” is involved in a large number of protein interactions (Schmidt et al. 2003) including self-association (Schmidt et al. 2002a).

The protein can be isolated from proliferating cells in two isoforms with approximately 395 and 345 kD encoded by the same gene. Alternative splicing of the mRNA of pKi-67 results in a short version without exon 7 (Duchrow et al. 1994). Analyzing 93 tissues mainly consisting of brain tumor specimens we recently found evidence that long and short isoform can be expressed independently of each other (Schmidt et al. 2004). Moreover, we presented evidence that the pKi-67 N-terminus is multiply spliced resulting in at least five different isoforms which fulfill different functions in cell cycle control. It is known that expression of pKi-67 antisense oligonucleotides (Schluter et al. 1993; Kausch et al. 2003), microinjection of anti-pKi-67 antibodies (Starborg et al. 1995), or overexpression of recombinant pKi-67 tandem repeats (Duchrow et al. 2001b) hampers cell division and attenuates progression of the cell cycle. Recently, we demonstrated that pKi-67 is involved in the regulation of cell division by organizing the life cycle of nucleoli dependent on cyclin B and ran (Schmidt et al. 2003), the effects of which were self-regulated (Schmidt et al. 2002a).

Because of its absence in quiescent cells (G0 phase) this protein developed into a widely used tumor marker in research and pathology. The standard antibody to detect pKi-67 is MIB-1. The fraction of MIB-1 positive tumor cells (the MIB-1/Ki-67 labeling index) is often correlated with the clinical course of cancer; pKi-67 is of prognostic value for many types of malignant tumor (Brown and Gatter 2002). The best studied examples in this context are carcinomas of the prostate and the breast. For these types of tumors the prognostic value for survival and tumor recurrence has repeatedly been proven in univariate and multivariate analyses.

For colorectal cancer the prognostic relevance is inconsistent. Recently, we compared the MIB-1 labeling index (immunohistochemical staining with pKi-67 antibody MIB-1; Fig. 1) with the pKi-67 in situ hybridization index (hybridization of a complementary strand of a pKi-67 probe on paraffin slides) and found an improved outcome of colorectal cancer patients with high MIB-1 labeling index but low pKi-67 in situ hybridization index (Duchrow et al. 2003). The present study retrospectively examines the expression of pKi-67 protein and mRNA by immunohistochemical assessment of MIB-1/Ki-67 labeling index and quantitative RT-PCR in 43 archival specimens of patients with colorectal carcinoma and the correlation between the outcome of patients and the pKi-67 expression.

Fig. 1.

MIB-1 staining. Examples for immunohistochemical staining of sections of colorectal carcinomas with MIB-1 (×200). A MIB-1 staining of a tumor with low labeling index (18.8%; UICC stage I, tumor grade 2). B Medium labeling index (37.6%; UICC stage III, tumor grade 3). C High labeling index (56.3%; UICC stage III, tumor grade 2)

Materials and methods

Patients

In this study 43 patients were included, who underwent curative surgery for colorectal cancer between January 1995 and October 1998. There were 24 men and 19 women, with a mean age of 72.1 years (age range 31.5–92.8 years). All patients were followed up at Lübeck University in a standardized follow-up program. Follow-up information included overall survival and cancer-related death, and data collected were entered prospectively into the colorectal cancer registry database. Mean follow-up was 41.2 months (range 2–92 months).

The tumors were classified according to the International Union Against Cancer (UICC; Sobin and Wittekind 1997) to stages I, II, or III. All patients were free of distant metastases and did not receive chemotherapy or radiation prior to the operation. Patients who died within 30 days after surgery were excluded from analysis. Clinical and histopathological data were obtained from the prospective colorectal cancer registry database, and included: age; gender; tumor site; UICC stage; histological differentiation (grades 1–3: well; moderately; and poorly differentiated); histology; tumor size and depth of invasion (pT); lymph node status (pN); preoperative serum carcinoembryonic antigen (CEA) level (normal, <5 ng/ml; increased, 5–20 ng/ml; excessive, >20 ng/ml); and follow-up information regarding overall survival. Patients’ clinical primary characteristics are presented in Table 1.

Table 1.

Relation between clinicopathological features and pKi-67 index or pKi-67 mRNA expression

Clinicopathological features	No. of patients	pKi-67	pKi-67
		mRNA expression	labeling index (%)
Age (years)		n.s.	n.s.
<45	1	0.687	21.5
45–65	10	0.183±0.096	26.9±11.0
65–85	27	0.177±0.122	33.7±14.0
>85	5	0.060± 0.032	40.4±10.0
Tumor stage		n.s.	n.s.
UICC I	13	0.165±0.109	34.7±2.9
UICC II	25	0.175±0.160	33.1±3.1
UICC III	5	0.215±0.093	25.3±3.7
UICC IV	0
Depth of invasion		n.s.	n.s.
T1	3	0.114±0.127	28.2±9.4
T2	10	0.180±0.106	36.8±2.7
T3	27	0.189±0.153	31.7±2.8
T4	3	0.110±0.112	33.3±11.1
Lymph node status		n.s.	n.s.
N0	38	0.171±0.143	33.3±2.2
N1	3	0.236±0.112	29.8±3.0
N2	2	0.183±0.080	18.6±6.0
Histopathological grading		n.s.	n.s.
Grade 1	1	0.257	10.3
Grade 2	34	0.183±0.144	32.0±2.0
Grade 3	8	0.138±0.119	38.4±5.9
Gender		n.s.	n.s.
Male	24	0.199±0.151	31.8±3.1
Female	19	0.148±0.117	34.0±2.5

Immunohistochemistry

Tissue samples

Tumor tissue had been fixed in 4.5% formalin and embedded into paraffin blocks as surgical pathology specimens in a routine manner. The paraffin blocks were stored at ambient temperature before analysis. Formalin-fixed, paraffin-embedded tissue samples of colorectal cancer were cut in sections of 4 µm, mounted on Superfrost slides (Menzel, Braunschweig, Germany), and dried overnight at 56°C. Representative sections were routinely stained with hematoxylin–eosin prior to immunostaining to ensure that the slides contain predominantly tumor tissue as regions of interest. For immunostaining the sections of the paraffin blocks were dewaxed in xylene, rehydrated in graded alcohol, and pretreated for antigen retrieval in citrate buffer (pH 6) in a microwave oven (800 W) for four cycles of 10 min. Sections were immunostained with MIB-1 (a mouse immunoglobulin of the IgG1 subclass) using the APAAP method (Alcaline-Phosphatase-Anti-Alcaline-Phosphatase) according to the manufacturer’s recommendations (DAKO, Glostrup, Denmark) and counterstained with hematoxylin before mounting.

Assessment of pKi-67 staining index

A minimum of two MIB-1 immunostained sections per patient were light-microscopically evaluated using a total magnification of 400x and a 10×10 square grid placed in the ocular. Avoiding margins of sections and areas of poorly presented morphology, areas with the highest number of positive nuclei were chosen to evaluate the MIB-1 labelling index (Ki-67 index) ensuring that the whole section was scanned. Any nuclear staining of tumor cells was considered to be positive. Cytoplasmic staining was considered as artifact. The indices were determined in a minimum of five high-power fields per specimen by counting more than 1000 of the total number of tumor cells per section. The MIB-1 labelling index was calculated as the percentage of positive labeled nuclei in relation to all tumor cell nuclei. Intratumoral variations of the labeling indices remained unconsidered. All slides were examined in blinded fashion without knowledge of clinical or histopathological data at the time of assessment.

PCR

RNA isolation

Sample and RNA preparation

Colon tumors were obtained intraoperatively from 43 colorectal cancer patients at Lübeck University Clinic. There were 24 men and 19 women, with a median age of 72.1 years (age range of 31.5–92.8 years). After routine histopathological examinations, a portion of each specimen was used for RNA isolation. All samples were immediately frozen in liquid nitrogen and stored at −180°C until further use. Total RNA was isolated using QuickPrep Micro mRNA Purification Kit (Amersham Pharmacia Biotech, Nr.035-00-19) following the manufacturer’s instructions.

cDNA preparation

The RNA was precipitated, washed twice in 70% ethanol (Roth, Karlsruhe, Germany), dried, resuspended in 40 μl of distilled water, and reversely transcribed using random hexamers (Amersham Pharmacia, Freiburg, Germany) into cDNA with MMLV (Life Technologies, Eggenstein, Germany) according to the manufacturer’s instructions.

Quantitative PCR

TaqMan PCR assay

The TaqMan 5’ nuclease fluorigenic quantitative PCR assay that we used is a well-established method of analyzing gene expression in a wide range of samples. Five microliters of cDNA were used in a total of 50-μl PCR reaction mix. Fifty microliters of reaction mixture for quantitative PCR were prepared in a single tube: 30 ng of the extracted cDNA; 10 µl TaqMan buffer A [50 mM Bicine, 115 mM potassium acetate, 0.01 mM EDTA, 60 nM Passive Reference, 8% glycerol (pH 8.2)]; 3 mM MgCl₂; 200 mM dATP, dGTP, and dCTP, dUTP; 0.3 mM forward primer; 0.3 mM reverse primer; 0.2 mM TaqMan probe; 5 units of rTth DNA Polymerase; and 0.5 unit of AmpErase UNG (the enzymes and the buffer containing the passive reference were from Applera Deutschland GmbH Applied Biosystems, Darmstadt,Germany). The conditions of the quantitative PCR were: 2 min at 50°C; 10 min at 95°C; and then 40 cycles of amplification for 15 s at 95°C and 1 min at 60°C, finally 25°C.

Primers and TaqMan probes

Primers and the TaqMan probe for pKi-67 were designed in accordance with Applied Biosystems guidelines. Primers and the TaqMan probe for ß-actin (TaqMan ß-actin control reagent kit) were also purchased from Applied Biosystems. The probes were labeled with a reporter dye [FAM or JOE (2,7-dimethoxy-4,5-dichloro-6-carboxy-fluorescein)], situated at the 5’ end of the oligonucleotide, and a quencher dye (TAMRA), located at the 3’ end. The sequences of primers and probes used were: pKi-67-forward: 5’ >AATTCAGACTCCATGTGCCTGAG <3’ (bases 9776–9798 of the pKi-67 coding sequence); pKi-67-reverse: 5’ >CTTGACACACACATTGTCCTCAGC <3’ (bases 9902–9925 of the pKi-67 coding sequence); pKi-67 probe: 5’ >(FAM) TCAAGAAAGACAAAAAGCCAGCCTGCAG(TAMRA) <3’ (bases 9800–9828 of the pKi-67 coding sequence); ß-actin-forward: 5’ >TCACCCACACTGTGCCCATCTACGA <3’ (bases 2141–2165 of the ß-actin coding sequence); ß-actin-reverse: 5’ >CAGCGGAACCGCTCATTGCCAATGG <3’ (bases 2411–2435 of the ß-actin coding sequence); ß-actin-probe: 5’ >(FAM)ATGCCC-X(TAMRA)-CCCCCATGCCATCCTGCGTp <3’ (bases 2171–2196 of the ß-actin coding sequence).

AmpliTaq DNA polymerase extended the primer and displaced the TaqMan probe through its 5’-3’ exonuclease activity. If the probe is intact the emission spectrum of the reporter is suppressed by the quencher. The nuclease degradation of the hybridization probe releases the reporter, resulting in increased fluorescence emission. The use of a sequence detector (ABI Prism 7700, Applied Biosystems) allows measurement of the amplified product in direct proportion to the increase in fluorescence emission, continuously, during the PCR amplification. The amplification plot is examined early in the reaction, at a point that represents the logarithmic phase of product accumulation. The point representing the detection threshold of the increase in the fluorescent signal associated with the exponential growth of the PCR product for the sequence detector is defined as the threshold cycle (C_T) which is predictive of the quantity of input target (i.e., when the PCR conditions are the same, the larger the initial template concentration, the lower the C_T); ΔC_T is the difference between C_T of pKi-67 and C_T of ß-actin used as reference gene.

Statistical analysis

Statistical evaluation was carried out with the SPSS program package (release 11.0.1, SPSS Software, Munich, Germany). Univariate analysis (χ2 and Student’s t-tests) examined whether immunohistochemical variables or mRNA levels are correlated with clinico-pathological data and can be used to predict recurrence or survival. For bivariate correlation the Pearson coefficient (R) and the level of significance (P) were calculated. For multivariate analysis we used logistic regression analysis to determine independent factors predictive of recurrence. Survival rates were calculated by the Kaplan-Meier method for analysis of censored data. The significance of differences in survival was analyzed by means of the log-rank test: P values ≤0.05 were considered significant. Deaths due to unknown causes or related to causes other than colorectal cancer were treated as censored observations at the time of death. In multivariate analysis, independent prognostic factors were determined by the Cox proportional hazards model. The level of significance was set at P≤0.05. Survival data were obtained from the tumor surveillance program of our clinic.

Results

A total of 43 samples of colorectal cancer were processed for analysis of pKi-67 expression by immunohistochemistry and quantitative TaqMan PCR. MIB-1 positive tumor cells showed a clear red staining of their nuclei. In contrast to the normal or adjacent mucosa, tumor specimens revealed a heterogenic staining pattern. A clear morphological distinction between tumor cells and other cells was always possible. The Ki-67 indices of the investigated tumors ranged from 10.3 to 66.4% with a mean pKi-67 index of 32.8% (median=31.3%). No significant correlation between pKi-67 index and any clinico-histopathological parameter could be shown (Table 1).

For all investigated tumors the mean mRNA level (ΔC_T values) was 0.1769 (median=0.141). The content varied in a wide range from 0.01 to 0.69. Similar to the pKi-67 index no significant correlation between pKi-67 mRNA level and clinico-histopathological parameters were detected (Table 1). Likewise, there was no significant correlation between pKi-67 index and mRNA expression (Fig. 2; R=−0.119; P= 0.457).

Fig. 2.

Correlation between pKi-67 mRNA expression and MIB-1 labeling index X/Y plot of pKi-67 mRNA ΔC_T values and MIB-1 labeling indices of 43 patients with a colorectal carcinoma. Between pKi-67 mRNA ΔC_T values and MIB-1 labeling indices no significant correlation could be found (Pearson coefficient for bivariate correlation: R=−0.119; level of significance: P=0.457).

The overall survival of all 43 patients had been followed up. At the end of the follow-up period (2–92 months, mean 41.2 months), 26 of the 43 patients (60.4%) were still alive. Analysis of the immunohistochemical expression and quantitative mRNA expression of pKi-67 with respect to survival is shown in Figs. 3 and 4. To find a possible correlation between pKi-67 index or pKi-67 mRNA level and patients’ prognosis we patterned three groups with 13 or 14 patients representing high, medium, and low MIB-1 labeling index / pKi-67 mRNA ΔC_T values, respectively. Analysis based on the log rank test revealed that patients with a high pKi-67 mRNA level had a significantly more favorable overall survival than patients with low or medium pKi-67 mRNA levels (P=0.0304; Fig. 4); however, there was no significant difference for the pKi-67 staining index (Fig. 3). A multivariate analysis of the clinical, immunohistochemical, and mRNA data indicated that only T-category and pKi-67 mRNA expression were independent prognostic factors (Table 2).

Fig. 3.

Survival curves (MIB-1 labeling index). No significant difference could be observed in the survival rates of patients with a colorectal carcinoma with high (≥31.6%), medium (31.6% >x<43.1%), and low (≤43.1%) MIB-1 labeling indices.

Fig. 4.

Survival curves (pKi-67 mRNA ΔC_T values). A significant (P=0.0304) difference could be observed in the survival rates of patients with a colorectal carcinoma with high (≥0.093), medium (0.093>x<0.236), and low (≤0.69) pKi-67 mRNA level (ΔC_T values).

Table 2.

Multivariate survival analysis (Cox regression)

Variable	P	χ2	Exp(B)
pKi-67 mRNA level	0.017	4.822	0.458
T category	0.028	6.209	2.937
N category	n.s.
Histopathological grading	n.s.
Tumor localization	n.s.
pKi-67 labeling index	n.s.
UICC stage	n.s.
Preoperative CEA level	n.s.
Histology (WHO category)	n.s.
Gender	n.s.

Discussion

In our study we immunohistochemically determined a median pKi-67 labeling index of 32.8 % (10.3–66.4%) with MIB-1, which is in accordance with the majority of previously published data (Table 3; Duchrow et al. 2003; Porschen et al. 1989; Hemming et al. 1992; Palmqvist et al. 1999; Kimura et al. 2000; Duchrow et al. 2001a; Brown and Gatter 2002; Allegra et al. 2003; Dziegel et al. 2003). Until today, results of pKi-67 mRNA quantifications only have been published once: in 2001 we quantitatively determined pKi-67 mRNA expression of twenty five resected colorectal adenocarcinomas of different stages (UICC; I–IV) and grades by competitive reverse transcriptase–polymerase chain reaction followed by product quantification using an ELISA (Duchrow et al. 2001a) resulting in 0.2 –4.4 amol Ki-67 protein specific mRNA per µg total RNA (median=0.88 amol). In contrast to this method, for the present study we used a quantitative Real-Time Taq-Man PCR and choose (a) curative resection, and (b) lack of neo-adjuvant treatment as criteria for inclusion of tumors. Consequently, the results of both studies cannot be well compared, because the methods for mRNA quantification and the selected patients are quite different.

Table 3.

Summary of published investigations of pKi-67 with prognostic relevance in colorectal carcinoma. IHC immunohistochemical method, ISH in situ hybridization, mRNA determination of pKi-67 mRNA

No. of patients	pKi-67 index (%; mean)	Method	Reference
61	38.7	IHC	Porschen et al. (1989)
51	59/42	IHC/ISH	Duchrow et al. (2003)
56	43.7/36.8	IHC	Palmqvist et al. (1999)
110	50.6	IHC	Kimura et al. (2000)
25	60/0.88 amol	IHC/mRNA	Duchrow et al. (2001a, 2001b)
41	<50–80.4	IHC	Petrowsky et al. (2001)
	>50–19.6
706	<20–11.4	IHC	Allegra et al. (2003)
	21–40–37.5
	41–60–41.6
	61–80–9
	>81–0.5
81	Negative - 2.4	IHC	Dziegel et al. (2003)
	<10–20.9
	10–30–43.2
	>30–33.3
43	<5–58.1	IHC	Ishida et al. (2004)
	5–24–16.2
	25–49–4.6
	>50–20.9

In the present study, no significant correlation between pKi-67 index and pKi-67 mRNA level could be demonstrated. There are several possible reasons to explain this finding:

1. The use of immunohistochemical markers often depends on a subjective assessment of specimens. Depending on the quality of fixation and labeling, even experienced pathologists will find different results on the same sections. A MIB-1 negative cell could still contain some amount of Ki-67 mRNA.

2. The Ki-67 antigen is expressed during all active phases of the cell cycle (G₁, S, G₂, and M) and consequently may be expressed in cells arrested in G₁/S or G₂/M as demonstrated with cell cultures by using synchronizing inhibitors. This could be shown in cell-cycle-arrested osteosarkoma cells (van Oijen et al. 1998). In conclusion, not all cells containing the antigen are actively proliferating and cell cycle arrested tumors cells might grow slower than indicated by the MIB-1 labeling index. Recently, we could demonstrate this phenomenon also for colorectal cancer (Duchrow et al. 2003). From these results we postulated that the better prognostic outcome of these patients may be explained because tumor cells being able to get into a cell-cycle arrest are thought to be less degenerated than tumor cells, which completely lost this ability and therefore may be more susceptible to adjuvant therapies and also to a response of the patient’s immune system. This postulation may also be valid for this study. Although no correlation between pKi-67 labeling index and mRNA level could be found, high pKi-67 indices are related to low pKi-67 mRNA levels (negative correlation coefficient; see Fig. 2); therefore, pKi-67 is detectable in tumor cells, although no de novo synthesis has taken place. With a half-life of the Ki-67 protein of approximately 1 h, this is impossible, but half time rises up to 60 h in cell-cycle-arrested cells (van Oijen et al. 1998). As a result, a great number of tumor cells could be cell-cycle arrested in these tumors.

3. The function of pKi-67 is modulated by its phosphorylation status, which continuously increases during the cell cycle. At the G₂ to M transition, pKi-67 undergoes massive phosphorylation, the so called phosphorylation shift. This changes the chemical characteristics and the function of the usually positively charged protein (MacCallum and Hall 2000; Schmidt et al. 2003). The detection of pKi-67 by MIB-1 also depends on the phosphorylation status. MIB-1 staining decreases rapidly when cells leave mitosis due to massive phosphorylation of pKi-67 (Endl and Gerdes 2000). The MIB-1 labeling index could be imprecise in such phases.

4. Overexpression of the “Ki-67 motif,” which is present in the pKi-67 repeats, can lead to inability of MIB-1 to detect its antigen. Thereafter, in order to prevent the underestimation of Ki-67 proliferation indices in MIB-1 labeled preparations, additional antibodies should be used. Overexpression of the “Ki-67 motif” leads to further growth inhibition of tumor cells (Schmidt et al. 2002b).

5. Further regulation of pKi-67 at the level of transcription or translation is possible. Presently, however, no study is known to prove such further regulation.

6. Detection of different protein isoforms with quantification methods could lead to missing correlation. In this study, we used quantitative real-time PCR and Immunohistochemistry for pKi-67 detection. Both methods recognize all pKi-67 isoforms.

Therapeutic decisions for cancer patients are often made on the basis of prognostic factors. In addition to the traditional pTNM classification, other factors have been validated for routine clinical use, e.g., expression of tumor suppressor genes and oncogenes. The influence of the proliferation rate of tumor cells on patients’ survival has been discussed for a long time and this rate is considered to reflect the course of tumor disease. Proliferative activity of tumors can be determined by various methods, e.g., by measuring doubling time, by calculating labeling index after ³H-thymidine or bromodeoxyuridine incorporation, by calculating mitotic index, or by using cytometry. Flow cytometric determination of S-phase cells is expensive and technically challenging. The other techniques are time-consuming and generally impractical for most clinicians, while real-time PCR will become a routine method.

pKi-67 is of prognostic value for many types of malignant tumors (Brown and Gatter 2002). In colorectal carcinomas the proliferative activity of tumor cells was determined in several investigations with different methods, almost exclusively by immunohistochemical means, but the prognostic relevance is inconsistent (a summary is shown in Table 3).

Our finding that high levels of pKi-67 mRNA are associated with improved outcome seems counterintuitive, given that several investigations in other malignant tumors found that a high pKi-67 labeling index was associated with a more aggressive tumor behavior and worse clinical outcome.

In colorectal tumors the relation between cell proliferation and clinical outcome is uncertain. It could be shown that high Ki-67 antigen expression in liver and lymph node metastasis correlated with a poor prognosis (Petrowsky et al. 2001; Ishida et al. 2004). Considering colorectal tissue itself, some authors concluded that colorectal carcinomas with low pKi-67 expression had a poor prognosis (Palmqvist et al. 1999; Allegra et al. 2003; Dziegel et al. 2003), whereas others found that those cases showing a high pKi-67 expression had a worse prognosis (Kimura et al. 2000; Duchrow et al. 2001a). Some studies showed no correlation at all (Hemming et al. 1992; review by Brown and Gatter 2002). The biological reason why high levels of Ki-67 should be associated with improved outcome will require additional research and clinical investigations.

Malignant tumors create an immuno-modulatory answer of the body. The answer consists mainly of an invasion of lymphocytes into the tumor (tumor-infiltrating lymphocytes, TILs; Ropponen et al. 1997). A drawback of the quantitative RT-PCR method is that it is impossible to localize the cellular source of the signal. As the pKi-67 gene is also expressed in nontarget cells, like TILs, results are indistinguishably influenced by a background signal. High pKi-67 mRNA levels could indicate a high activity of TILs and therefore result in a prognostic advantage.

This disadvantage can be reduced by careful microscopic selection of tumor probes. Proliferating nontumor cells express the Ki-67 protein and its mRNA on a much lower level than tumor cells. We were able to prove this in two previous independent studies with both quantitative RT-PCR and in situ hybridization (Duchrow et al. 2003; Duchrow et al. 2001a); therefore, unavoidable contamination of tumor samples with adjacent tissue would have only insignificant influence on the quantification of the Ki-67 protein mRNA expression of tumors. A new approach is microdissection that helps selecting only target cells for pKi-67 mRNA detection.

Another reason for missing relation between cell proliferation and clinical outcome could be the marked heterogeneity of pKi-67 expression in colorectal carcinomas; however, our results have to be interpreted carefully because of the small number of patients included in this study. Quantitative RT-PCR is an objective alternative method to determine the proliferative activity of tumor cells in contrast to immunohistochemistry. The performance of RT-PCR for the quantification of pKi-67 mRNA levels provides several advantages (e.g., non radioactive method, greater linear dynamic range, no post-amplification steps, high number of probes per plate, method suitable for automatization, sensitive and exact method, same protocol and universal Mastermix for amplification of other sequences) over previous methods used for quantification in clinical tumor research.

The application of real-time RT-PCR decreases the time and costs of mRNA quantification. The time needed from surgical tumor resection to the quantification was less than 48 h; however, future investigations have to show whether the Ki-67 index is an independent prognostic marker for colorectal adenocarcinoma and whether the quantitative RT-PCR is useful to clinical investigations.