Does cyclin E and p57^kıp2 expression have prognostic and survival value in colorectal adenocarcinoma?

Abstract

Introduction: Colorectal cancer is still one of the main causes of cancer death in the world. There is a continous need for novel biomarkers for diagnose, treatment modalities and follow-up. Cyclin E and p57^KIP2 as the positive and negative regulators of cell cycle seem to be an important target for investigations. Materials and methods: In a retrospective setting, primary colorectal adenocarcinoma cases examined in Mustafa Kemal University, School of Medicine, Pathology Department between 2008-2015 were reviewed. Immunohistochemical expressions of cyclin E and p57^KIP2 in 80 pairs of colorectal carcinoma and adjacent normal mucosal tissues were evaluated and the findings were compared with clinicopathological parameters and survival time. Results: There were no statistically significant difference between two groups both in cyclin E and p57^KIP2 stained tissues (P>0.05). There were 40 (50%) patients in high-expression group and 40 (50%) patients in low-expression group for cyclin E. p57^KIP2 was negative in 55 (68.75%) patients and positive in 25 (31.75%) patients. There were no statistically significant relation between p57^KIP2 and cyclin E expressions with clinicopathologic parameters defined as age, gender, lymphovascular invasion, perineural invasion, depth of invasion, nodal involvement, emergency in operation, perforation before operation and overall survival except that there was significant relation between p57^KIP2 expression and histological grade (P=0.012). Conclusions: Immunohistochemical studies of cyclin E and p57^KIP2 should be performed with larger series of patients supported by more detailed technical research methods to be candidates as predictive markers for treatment modalities and prognostic factors.

Introduction

Colorectal cancer is one of the leading causes of cancer death in the world, and according to World Health Organization GLOBOCAN database, incidence and mortality rates of colorectal cancer are 9.7% and 8.5%, respectively (http://globocan.iarc.fr/) [1]. The basic risk assesment and therapeutic approach is accomplished by pathologic stage and histologic subtype with the aid of different biomarkers in newly diagnosed colorectal cancer [2]. But there is a need for new biomarkers for risk assesment, early diagnose, treatment modalities and predicting response to therapy protocols with this lethal disease.

Proteins involved in the control of the cell cycle like cyclins and cyclin-dependent kinases (CDKs) and their inhibitors are among the novel therapeutic targets and are of great interest for new investigations. Among them CDK4-6 inhibitors are new agents targeting the cell cycle and have shown promising effect in breast cancer and gynecologic tumors [3,4]. The fundemantal ability of cancer cells is to sustain chronic proliferation by supporting the cell cycle independently [5].

Abnormalities in the cell cycle are a leading factor in tumor progression and there are several regulators for governing the cell cycle. Cyclins activate cyclin-dependent kinase and control the progression of cells through the cell cycle [6,7]. Cyclin E, which includes full-length (FL) cyclin E, low-molecular weight (LMW) cyclin E and total molecular cyclin E (full-length plus low-molecular weight), is one of the most important regulators of cell cycle by playing a role in G1 phase, G1-S transition and causing chromosomal instability. Abnormalities in cell cycle may lead to malignant transformation and tumor progression. Notably, cyclin E has been defined as a poor prognostic marker in various solid organ carcinomas including breast. Interestingly overexpression of cyclin E has been related with lower clinical benefit rate and progression-free survival from some recombinant monoclonal antibody therapy agents in advanced breast cancer patients [8,9]. But there are conflicting data about the role of cyclin E in colorectal adenocarcinoma (CRA) [2,10,11].

The gene encoding p57^KIP2 is on chromosome 11p15. 5 [12], that often undergoes maternal allele loss of heterozygosity in neoplasia. p57^KIP2 is known to be the negative regulators of cell cycle belonging to KIP family. The CIP/KIP family is composed of p21^WAF1/CIP1, p27^KIP1 and p57^KIP2. These proteins share a conversed N-terminal domain and inhibit a broad range of CDKs by binding to several cyclin/CDK complexes, including cyclin D/CDK4 (or CDK6), cyclin E/CDK2 and cyclin A/CDK2 contribute to passage of cells through the G1, S, G2, and M phases of the cell cycle [13]. p57^KIP2 also suppresses cellular transformation by binding to proliferating-cell nuclear antigen [14] inhibits the conversion of conditionally immortal cells to the fully immortal phenotype [15], overexpression of p57^KIP2 induces cellular senescence, and arrests cells in G, phase completely [12], Its role in carcinogenesis and metastasis to lymph nodes of tumor remains unclear [16].

There are limited data searching the role of p57^KIP2 in colorectal carcinogenesis with its interaction by another cell cycle regulator; cyclin E [16,17]. The aim of our study is to search the expression patterns of cell cycle regulators cyclin E and p57^KIP2 in our CRA series and their relations with clinopathologic parameters and survival rates.

Materials and methods

This was a retrospective archival study including 80 patients diagnosed with colonicadenocarcinoma in the Pathology Department of Mustafa Kemal University, School of Medicine between 2008-2015. Only patients that were diagnosed with resection specimens and had no neo-adjuvant chemotherapy were included in this study.

The ethical committee on human research at our institution approved the protocol for all human research with number 4298783/05001 date 17/01/2015. The haematoxylin eosin (HE) stained cross sections of the cases have been re-evaluated for histopathology on the basis of prognostic factors. The preparate representing the tumor in the best manner was selected.

Sections of 3-4 mm thickness were cut from the paraffin blocks of these preparations and then were de-paraffinized and rehydrated through a graded series of alcohol, microwave antigen retrieval method was used, followed by incubation with cyclin E (HE12): sc-247 (monoclonal mouse antibody, Santa Cruz Biotechnology, CA, USA) and P57^KIP2 (C-20): sc-1040 (poyclonal rabbit antibody, Santa Cruz Biotechnology, CA, USA).

Immunohistologic staining (IHS) was applied. Cytoplasmic and nuclear staining in normal testis tissue for cyclin E and nuclear staining in fetal kidney for p57^KIP2 accepted as positive controls. Immunoreactivity scored by nuclear immunostaining in tumor for cyclin E and p57^KIP2 was calculated as the percentage of positive epithelial cells in relation to the total number of cells encountered in at least 5-10 representative high power fields (500-1000 neoplastic epithelial cells).

The immunoreactivity was interpreted by light microscopy. Staining was evaluated only in areas with well-preserved tissue morphology and away from necrosis or artifact. Staining was found in the nucleus, and tumor cells showed a range of intensities of staining.

Every stained nucleus was considered positive, irrespective of intensity. In agreement with previous studies [11,18], cases were regarded as negative when <2% of the tumour nuclei showed staining or no staining at all (for the detection of cyclin E and p57^KIP2).

Statistical analysis

The collected data were analyzed by using the SPSS version 21.0 (SPSS Inc, Chicago, IL, USA). Tumor and non-tumor tissues of colorectal cancer patients were compared by paired t-test and Spearman’s correlation test. Clinicopathologic parameters, p57^KIP2, and cyclin E comparisons of tumor cases were performed with Pearson Chi-Square non-parametric test. The postoperative survival rate was analyzed by Kaplan-Meier method, and differences in survival rates were assessed with log-rank test. Cox regression test was used for univariate and multivariate analysis. P values <0.05 were deemed significant.

Results

Cyclin E and p57^KIP2 expression were investigated immunohistochemically in CRA patients’ cancerous and normal tissues (Figures 1, 2, 3 and 4). While the relative expression levels of cyclin E in CRA patients’ cancerous and non-cancerous tissues were found to be a mean of 14.9167 and 4.8194, p57^KIP2 expression levels were 1.1667 and 1.5556, respectively. There were no statistically significant difference between two groups both in cyclin E and p57^KIP2 stained tissues (P>0.05). But there was apositive correlation between cyclin E and p57^KIP2 expressions in cancerous tissues (r=0.409) (Figure 5A, 5B).

Figure 1.

Strong nucleer cyclin E expression in carcinoma tissue (Cyclin E ×200).

Figure 2.

Weak and focal expression of cyclin E in adjacent normal mucosa (Cyclin E ×200).

Figure 3.

P57^KIP2 expression of adenocarcinoma (p57^KIP2 ×200).

Figure 4.

P57^KIP2 expression in surface epithelium of normal colonic mucosa (p57^KIP2 ×200).

Figure 5.

A, B. Cyclin E and p57^KIP2 expressions in carcinoma and normal mucosa.

To study the relationships of clinicopathological characteristics and cyclin E expression with p57^KIP2 in CRA patients, patients were divided into two groups. Cyclin E expression levels were divided into high and low-expression groups according to mean value, and p57^KIP2 expression was grouped as negative or positive.

There were 40 (50%) patients in high-expression group and 40 (50%) patients in the low-expression group for cyclin E. The relation between cyclin E expression level and age, gender, lymphovascular invasion, perineural invasion, histological grade, depth of invasion, nodal involvement, emergency in operation, and perforation before operation was investigated and we did not find any statistically significant relation between these parameters and cyclin E relative expression levels. p57^KIP2 was negative in 55 (68.75%) patients and positive in 25 (31.75%) patients.

There were no statistically significant correrelations between p57^KIP2 expression and clinicopathologic parameters defined as age, gender, lymphovascular invasion, perineural invasion, depth of invasion, nodal involvement, emergency in operation, and perforation before operation but there was a significant correlation between p57^KIP2 expression and histological grade (P=0.012) (Table 1).

Table 1.

Association of Cyclin E and p57 expressions with clinicopathologic factors

Variable	Cyclin E expression			p57 expression
	Low (n)	High (n)	p value	Negative (n)	Positive (n)	p value
Age	61.37±14.3235	60.20±15.1322	0.722	60.58±14.6675	61.24±14.9088	0.854
Gender
Male	20	14	0.129	13	12	0.334
Female	20	26		33	22
Histologic type
Well, moderate	30	24	0.152	42	12	0.012
Poor, mucinous	10	16		13	13
Depth of invasion
T1, T2	4	7	0.339	7	48	0.235
T3, T4	36	33		4	21
Location
Colon	31	24	0.091	37	18	0.672
Rectum	9	16		18	7
Lymph node metastasis
Absent	20	23	0.453	27	16	0.364
Present	20	17		28	9
Lymphovascular invasion
Absent	24	31	0.091	40	15	0.255
Present	16	9		15	10
Perineural invasion
Absent	29	31	0.606	44	16	0.126
Present	11	9		119
Emergency in operation
Absent	28	32	0.302	40	20	0.486
Present	12	8		15	5
Perforation at initial evaluation
Absent	36	35	0.723	50	21	0.365
Present	4	5		5	4
Liver metastasis
Absent	35	34	0.745	48	21	0.694
Present	5	6		7	4

For the relationship of clinicopathological characteristics with cyclin E and p57^KIP2 expression levels with patients’ survival, univariate analysis of clinicopathological factors and cyclin E and p57^KIP2 was done. Of the 80 patients, 35 (43.48%) died within the follow-up period. The cumulative overall survival rate was 56.3%. There were 7 (43.8%) deaths in the low-cyclin E expression group and 28 (56.9%) in high-expression group. There were 22 (62.9%) deaths in the p57^KIP2 negative expression group and 12 (37.1%) in the p57^KIP2 positive expression group (Figure 6A, 6B). Survival time was associated with the depth of invasion (P=0.05, r=-0.313) in the p57^KIP2 group and there was a negative correlation between p57^KIP2 positive expression group and survival time (r=-0.324). No correlationswere found with other clinicopathologic characteristics both for cyclin E and p57^KIP2. (Table 2).

Figure 6.

A, B. Kaplan-Meier curves for overall survival in colorectal cancer patients according to Cyclin E and p57^KIP2 expression.

Table 2.

Univariate analysis of clinicopathological factors for overall survival

Variable	Number	Hazard ratio	95% CI	** p value
Age
≤65	53	1	0.764-3.179	0.223
>65	27	1.558
Gender
Male	34	1	0.524-2.790	0.655
Female	46	1.210
Histologic grade
Well, moderate	54	1	0.403-2.025	0.804
Poor, mucinous	26	0.903
Depth of invasion
T1, T2	11	1	0.359-3.521	0.841
T3, T4	69	1.124
Location
Colon	55	1	0.445-3.266	0.714
Rectum	25
Lymph node metastasis
Absent	43	1	0.707-2.988	0.309
Present	37	1.454
Lymphovascular invasion
Absent	55	1	0.137-3.179	0.603
Present	25	0.659
Perineural invasion
Absent	60	1	0.988-19.884	0.052
Present	20	4.443
Emergency in operation
Absent	60	1	0.299-2.888	0.898
Present	20	0.929
Perforation at initial evaluation
Absent	71	1	0.410-7.776	0.439
Present	9	1.786
Liver metastasis
Absent	69	1	0.330-2.931	0.977
Present	11	0.984

^**p<0.05.

Cyclin E and p57^KIP2 expression levels, clinicopathologic characteristics, and overall survival were searched by Kaplan-Meier analysis, log-rank test, and Cox-regression test. There was no statistically significant correlation (P>0.05) between cyclin E and p57^KIP2 expression, clinicopathologic characteristics, and overall survival (Table 2).

Also, the relation of cyclin E and p57^KIP2 expressions with carcinomas localised left and right colon was interrogated. There was no statistically significant correlation (P>0.05).

Discussion

There is a need to find new prognostic biomarkers which may be applied in clinical practice for the identification of patients with CRA who have higher survival potential. The detection of these markers may aid in risk stratification and cyclin E with p57^KIP2 may be included in the list of possible candidates.

Abnormalities in the expressions of cell cycle regulatory genes might lead to neoplastic transformation, dysregulation in positive or negative regulators by aberrant or lost expression, resulting in malignant transformation of cells.

Cyclin E, a positive regulator of the cell cycle, has a leading role in malignant transformation of cancer cells and tumorigenesis due to chromosomal instability with resulting metastasis [19-22].

There are several studies supporting overexpression of cyclin E in CRA carcinogenesis [11,23-25]. Yasui et al. defined cyclin E expression in 56% of adenocarcinomas; compared to this study we have found higher cyclin E expression rates of 80% in our series [24]. There are some studies demonstrating high expression levels of cyclin E in their CRA patient groups. Some studies indicate that higher levels of cyclin E expression might be related to gene amplification, chromosomal instability, and play a role in colorectal carcinogenesis [26,27]. We also have low and high cyclin E expressing CRA groups which might also show the role of these carcinogenesis pathways in our patients.

Bioabnormalities in the mucosa adjacent to tumor are shown in a few reports that included mucosa adjacent to tumor in their studies [22,25,28]. Sutter et al. showed overexpression of cyclin E with corresponding normal mucosal samples in their 20 patient series by immunohistochemistry. They defined scattered staining of the nuclei by cyclin E with lower averages compared to tumor tissue [25].

On the other hand Qi et al. explored cyclin E1 expression in 50 rectal cancer and 16 pericarcinoma tissues and found higher protein expression levels in carcinoma compared to pericarcinoma tissues suggesting a synergistic effect of cyclin E in development and progression of rectal cancer [29].

We have also included normal mucosal samples of the same cancer patients and found expression in both of the groups; there was difference in expression levels in tumor and normal mucosal samples but it was not statistically significant. Therefore it is possible to show that there are different alterations in macroscopically normal mucosa of CRA patients. There are conflicting results about cyclin E expression and its correlation with clinicopathological parameters including prognosis. The clinicopathologic characteristics of our patients correlated with low and high expression groups of cyclin E; unfortunately there was not any statistically significant difference. There are also studies that were not successful in showing a correlation [11,25,30]. Iochim et al. also detected cyclin E overexpression in 18 of 60 cases but expression did not correlate with conventional clinicopathological features [11].

However, in a study conducted by Li et al., a decrease in cyclin E was associated with some of the parameters like tumor size, mucinous type, venous invasion, and poor prognosis [17]. Moreover, Perea et al. also proved that lack of cyclin E expression might be an indicator of poor prognosis and a marker of advanced stage disease [31]. But it should be cautioned that number in their study group [28] was less than ours (80) which could explain the differences in results among various studies. p57^KIP2 has an important role in neoplastic transformation of various human cancers but its expression has been shown limited in numbers of studies in CRA [16,32-34]. Li et al. investigated 189 primary CRA and 22 normal mucosa specimens for p57^KIP2 immunohistochemically and found a significant decrease in primary carcinomas. Beside this, Noura et al. examined the expression of p57^KIP2 in 110 pairs of colorectal non-tumor and cancer tissues. Their immunohistochemical analysis showed weak expression in normal mucosa compared to cancer tissues [16,17]. We have also studied cancer and adjacent normal mucosa to compare expression paterns and show the development of neoplasia. We determined p57^KIP2 in both of the sample groups but unfortunately could not determine a significant difference in expression levels or patern. The limitation of our study was the smaller sample size and a different clone of the immunohistochemical reagent. Some reports have also investigated the relationship of p57^KIP2 with clinicopathologic parameters and survival time. Li et al. could not show a correlation of p57^KIP2 with clinicopathologic indices but the patients having loss of this protein tended to show poorer prognosis. Noura et al. found correlation of p57^KIP2 and large tumor size or female gender. They stated that p57^KIP2 did not influence the prognosis [16,17]. We have also searched with the same analysis and determined the relation of p57^KIP2 with histopathologic differentiation. Interestingly, positive p57^KIP2 expression was related with poorer survival time.

One of the main goals of our study was to explore the interaction of these two proteins (cyclin E and p57^KIP2) in the same CRA cases due to their roles in the cell cycle. The positive correlation between cyclin E and p57^KIP2 expressions in cancerous tissues (r=0.409) proved the imbalance of positive and negative regulators in the neoplastic process.

To date, studies are investigating right and left colon cancers that are suggested to be oncologically different [35,36]. We also evaluated cyclin E and p57^KIP2 expression, grouping our cases according to their localisation as left and right side,searching the role of cancer pathways. There was no statistically significant correlation.

Some aspects of our study did not support our hypothesis. The immunohistochemical methods might be affected by many factors: laboratory conditions, antibody clone, and most importantly number of patients in a single center.

In conclusion; immunohistochemical studies of cyclin E and p57^KIP2 might have additional roles in colorectal carcinogenesis but before accepting as potential prognostic and predictive biomarkers for CRA, we need larger series of patients supported by more detailed technical research methods.

Disclosure of conflict of interest

None.