This article is mentioned in:

Introduction

Colorectal cancer (CRC), one of the most common malignant neoplasms in Western countries, has become increasingly frequent in China, mostly due to the improvement of living standards and alteration of eating habits. Although overall survival has climbed to approximately 65% after 5 years, partly due to the development of systematic therapeutic modalities in past decades, one-third of advanced patients succumb to progressive disease (1). Mounting evidence has demonstrated that several genetic alterations are involved in the carcinogenetic process of CRC (2). Identification of multiple cancer-associated molecules may further our knowledge of the gene regulation throughout the tumorigenetic process in human CRC.

Peroxisome proliferator-activated receptor γ (PPARγ) is a member of the steroid receptor superfamily and belongs to the ligand-activated transcription factors. PPARγ forms a heterodimer with retinoid X-receptor α in the presence of its ligands, such as polyunsaturated fatty acids and arachidonic acid derivatives, following which it regulates the expression of target genes. In addition to functioning as a significant regulator of adipocyte differentiation and fatty acid metabolism, PPARγ was found to be correlated with several types of cancer, including CRC (3,4). Although aberrations of PPARγ expression have been observed in several other cancer types, previous studies revealed that PPARγ was abundant in normal and malignant tissue in the colon (5,6), where its function is largely unknown. Moreover, the roles of PPARγ in various stages of CRC are also poorly understood. One recent study demonstrated that activation of PPARγ inhibited cell growth and induced cell differentiation in human colorectal cancer through several pathways involving the induction of PTEN over-expression (7). PTEN undergoes genetic or epigenetic inactivation in many malignancies including colorectal cancer (8,9). The role of the lipid phosphatase activity of PTEN as a negative regulator of the cytoplasmic phosphatidylinositol-3-kinase/Akt pathway is well known. However, certain studies have also revealed that PTEN may relocate to the nucleus and indicate an additional modulation mechanism (10,11). Additionally, the question of whether aberrant PTEN is involved in the initial stages or is a later event in CRC remains controversial.

To date relatively few studies have focused on the correlation of expression of PPARγ and PTEN in various tissues. In this study, we examined the protein expression of PPARγ and PTEN using immunohistochemical (IHC) staining on tissue microarray (TMA) in CRC, adenoma and corresponding normal para-cancerous colorectal tissues and explored their potential clinical significance.

Materials and methods

Materials and patients

Formalin-fixed and paraffin-embedded colorectal cancer tissues, adenomatous polyps and paraneoplastic normal tissue specimens were obtained from the archives of the Department of Gastroenterology, the First Affiliated Hospital of Soochow University and the National Engineering Center for Biochip in Shanghai, China. Demographic and clinical data were collected retrospectively (Table I). A total of 139 patients comprising 84 males and 55 females (age range 54±28), with CRC (colon cancer, 85; rectal cancer, 54) were included in this study. None of the patients had received radiotherapy or chemotherapy prior to surgery. Specimens were viewed by one pathologist (Dr Jing Fang). CRC tissues were diagnosed, staged and graded according to the guidelines on the national diagnosis and treatment of CRC. In these cohort CRC patients, the median age was 54. The incidence of CRC in China demonstrates a younger tendency. Therefore, in our study, we chose 54 years as a cut-off value to separate patients into two groups. In regard to the degree of cancer cell differentiation, 27 cases were well-differentiated, 81 were moderately differentiated and 31 were poorly differentiated. Lymph node metastases were observed in 50 patients and distant metastases were observed in 23 patients, 13 patients had missing parameters. With the exception of 10 patients with unclear stage parameters, the ramaining 129 patients were classified into Dukes’ A (25 cases), B (44 cases), C (37 cases) and D (23 cases), according to the clinical stages. The 18 adenomatous polyp specimens were removed by colonoscopy and the 50 para-cancerous normal tissues were resected at least 5 cm away from the corresponding cancer tissues. Written informed consent was obtained from each patient according to the institutional regulations.

Table I Characteristics of study population.

Table I

Characteristics of study population.

	Age at enrolment (mean ± SD)	Gender		Location
		Male	Female	Colon	Rectum
Cancer (n=139)	54±28	84	55	85	54
Adenoma (n=18)	48±12	11	7	12	6
Normal (n=50)	55±27	35	15	32	18

Reagents

Primary rabbit polyclonal antibodies against human PPARγ and PTEN were obtained from Santa Cruz Biotechnology (Santa Cruz, CA, USA). The EnVision+™ goat-anti rabbit kit was a product of GenTech company.

Construction and sectioning of tissue microarray

The TMAs were assembled using a manual tissue arrayer. In brief, the CRCs, adenomas and normal tissues were embedded in paraffin blocks and 5-μm sections stained with hematoxylin-eosin (HE) were obtained to select representative areas for biopsies. The selected donor cores (diameter 1.0 or 1.5 mm) were extracted from areas of individual donor paraffin blocks and precisely arrayed into a new recipient block with a custom-built instrument (Beecher Instruments, Silver Spring, MD, USA). Tissue specimens from 139 CRCs, 18 adenomatous polyps and 50 para-cancerous benign mucosas were arranged in 4 recipient paraffin blocks. A total of 3 core tissue biopsies were obtained from each specimen. After construction, 4-μm slides were cut from these recipient blocks and used in series for IHC staining.

IHC staining to TMA sections

Microarray sections (4-μm) were soaked in xylene overnight to remove any adhesive from the tape transfer system. Slides were deparaffinized and rehydrated and antigens were then retrieved by autoclave. Following incubation in 10% normal goat serum for 20 min at room temperature, the sections were stained with the primary antibodies at 1:100 dilution overnight at 4°C. The antibodies used were monoclonal (rabbit antihuman) with PPARγ and PTEN purchased from Santa Cruz Biotechnology Inc. The primary antibodies were further examined using a biotin-free, horseradish peroxidase enzyme-labeled polymer conjugated to anti-rabbit secondary antibodies (GeneTech).

Each of the IHC-stained sections was scored according to the sum of the extension area and intensity (12). Nuclear staining with yellow and/or brown was scored for PPARγ, whereas complete cytoplasm staining was assessed for PTEN. The intensities were scored as 0 (negative), + (faint/), ++ (moderate), and +++ (strong). The immunoreactive areas were categorized as 0 (<5%), + (5–25%), ++ (26–50%), +++ (51–75%) and ++++ (>75%). Combined scores ≥ ++ were considered positive and staining scores ≥ +++ were regarded as strong positive.

Statistical analysis

Statistical analysis of PPARγ and PTEN stainings and their correlations with clinicopathological variables were carried out using the SPSS program, version 13.0 (SPSS, Inc.). Pearson’s χ2 test or Fisher’s exact test were applied to compare the staining index and clinicopathological parameters of the two molecules. The significance level was defined as P<0.05.

This study was approved by the Ethics Committee of Soochow University.

Results

Expression of PPARγ and PTEN in different tissues

The expression of PPARγ protein was mainly located in the cell nucleus, with a minority in the cytoplasm. Positive expression was detected in 125 specimens in 139 CRC tissues, 15 specimens in 18 adenomas and 41 in 50 corresponding normal colorectal tissues, whereas strong positive rates were 25% (35/139), 17% (3/18) and 34% (17/50), respectively (Fig. 1). No significant differences were found among the CRC, adenoma and normal colorectal mucosa regarding IHC staining in our study (P=0.055). (Table II). Furthermore, there was also no signifcantly different expression observed between the cancer tissues and their normal paracancerous counterparts following the exclusion of the adenoma data (P=0.112).

Figure 1 Immunophenotype of the investigated antigens (PPARγ and PTEN) in three colorectal tissues (original magnification, ×100). Positive stainings of PPARγ were located in the cell nuclei, while those of PTEN were in the cytoplasm. (A) HE staining of the TMA section of colorectal cancer tissue. (B, C and D) PPARγ expression in benign mucosa, adenomas and cancer tissues. (E, F and G) PTEN expression in benign mucosa, adenomas and cancer tissues, respectively.

Table II Expression of PPARγ and PTEN in different colorectal tissues.

Table II

Expression of PPARγ and PTEN in different colorectal tissues.

Groups	Numbers	PPARγ (%)				P-valuea	PTEN (%)				P-valuea
		−	1+	2+	3+		−	1+	2+	3+
Cancer	139	13 (9.4)	1 (0.7)	90 (64.7)	35 (25.2)	0.055	6 (4.3)	4 (2.9)	74 (53.2)	55 (39.6)	0.100
Adenoma	18	3 (16.7)	0 (0.0)	12 (66.6)	3 (16.7)		1 (5.6)	2 (11.1)	12 (66.7)	3 (16.6)
Normal	50	7 (14.0)	2 (4.0)	24 (48.0)	17 (34.0)		1 (2.0)	1 (2.0)	29 (58.0)	19 (38.0)

a Tested by χ² or Fisher’s exact test.

The expression of PTEN protein was commonly observed in the cytoplasm in CRC, adenoma and normal colorectal mucosa. Positive immunostainings were observed in 129 tissues of 139 CRC specimens, 15 of 18 adenomas and 48 in normal colorectal mucosa tissues (Fig. 1). No statistically significant difference was found between these tissue types. (P=0.100) (Table II). Similarly, when the cancerous and para-cancerous tissues were isolated, a significantly different expression of PTEN was not observed (P=0.846).

We further investigated the relationship between the PPARγ and PTEN, according to our originally conservative plan. Notably, the expression of PPARγ was found to be positively correlated with that of PTEN in this cohort tissue in our study (r=0.375, P=0.000).

Correlations between the expression of PPARγ/PTEN and the clinicopathological parameters

The association between the IHC pattern of PPARγ and clinicopathological features including age, gender, histological grade, lymph node and distant metastasis were not statistically significant (P>0.05). Similarly, there were no significant associations between the expression of PTEN and age, gender, clinical stage and lymph node metastasis. However, PTEN protein expression was found to be correlated with CRC histological grade (P=0.006) and distant metastasis (P=0.015). PPARγ expression was also shown to correlate with clinical stage in this cohort of CRC patients (P=0.004). (Table III).

Table III Correlations between the expression of PPARγ or PTEN and the clinicopathological data in the CRC patients.

Table III

Correlations between the expression of PPARγ or PTEN and the clinicopathological data in the CRC patients.

Clinicopathological data	Number	PPARγ				P-valuea	PTEN				P-valuea
		−	1+	2+	3+		−	1+	2+	3+
Age
<54	45	7	0	26	12	0.274	3	2	24	16	0.576
≥54	94	6	1	64	23		3	2	50	39
Gender
Male	84	9	1	51	23	0.647	5	4	45	30	0.234
Female	55	4	0	39	12		1	0	29	25
Histological grade
Low	31	6	0	17	8	0.165	5	2	18	6	0.006
Moderate	81	6	0	53	22		1	2	43	35
High	27	1	1	20	5		0	0	13	14
Lymph-node metastasis
Negative	89	6	0	59	24	0.259	3	2	42	42	0.069
Positive	50	7	1	31	11		3	2	32	13
Distant metastasis
Negative	103	7	1	65	30	0.175	2	3	58	40	0.015
Positive	23	4	0	16	3		4	1	8	10
Dukes’ stage
A	25	2	0	21	2	0.004	0	1	12	12	0.095
B	44	1	0	23	20		1	1	23	19
C	37	5	1	22	9		1	1	25	10
D	23	4	0	16	3		4	1	8	10

a Tested by χ² or Fisher’s exact test. CRC, colorectal cancer.

Discussion

Although it has been identified in other cancer cells, including prostate, gastric, pancreatic and lung cancer, but found to be absent from their corresponding benign tissues (13–16), PPARγ was found to be expressed ubiquitously in colorectal epithelial cells and cancer cells by certain authors (17–19). The first problem to be addressed in the study is therefore the exact expression status of PPARγ in various colorectal specimens.

In their study, Mansen et al (17) found that expression levels of PPARγ and PPARβ/δ in colorectal tissues of mice were higher than those in the small intestine, and that the expression of PPARγ was gradually increased from crypt to lumen. Similarly, Fujisawa et al (18) found that PPARγ was expressed highly in human normal colon epithelium and cancer cells. In the present study, our results revealed that PPARγ was highly expressed in three tissues and that no statistically significant difference was observed, consistent with the study by Eun et al (19). Therefore, in combination with the results of previous studies and our investigations, we propose that the colorectal basal cells decreased in fissionability and developed differentiated potential followed by the increased expression of PPARγ during the differentiation process. These results demonstrate that, although a variety of gene aberrations such as deletion of APC and P53 and activation of RAS occur during the progression of CRC, the protein expression of PPARγ may exhibit no marked change, indicating that the variation of PPARγ protein expression may not be involved in the progression of CRC. However, whether other patterns of PPARγ aberrations such as mutation, rearrangement or post-translational modification are involved in the pathogenesis of CRC is poorly understood. Therefore, together with other studies, we propose that PPARγ expression may not be considered as an early diagnostic and progression index in CRC. Nevertheless, the decreasing expression tendency of PPARγ found in advanced colorectal cancers in this study may indicate the prognostic potential of PPARγ. Furthermore, activation of PPARγ demonstrating an anti-oncogene role in CRC may indicate that although PPARγ was expressed equally in benign tissues and cancer cells of the colorectum, PPARγ activity may decrease in malignant tissues through several pathways such as mutation, post-translational modification, competitive inhibition of PPARβ/δ and decreasing of the endogenic ligands (20,21). Therefore, the specific role of PPARγ in the development and progression of CRC requires additional studies.

As a tumor suppressor gene, PTEN on chromosome 10q23.3 encodes a dual-specificity phosphatase that negatively regulates the phosphoinositol-3-kinase/Akt pathway and mediates cell-cycle arrest and apoptosis. Several studies have revealed that PTEN aberration is involved in the development of a number of types of cancer (22,23). One study also found that PTEN is involved in CRC (24). However, Zhou et al (25) revealed that distinct PTEN aberrations may occur in variant genetic backgrounds. Furthermore, it was found that the mutation in the ATP motif of PTEN may affect its subcelluar localization and tumor suppressive function (11,26). In our study, PTEN protein was mostly expressed in the cytoplasm, but not in the nucleus, and no significant difference was observed between CRC, adenoma and normal colorectal mucosa, which was consistent with previous studies (27,28). However, whether the unmatched paired design and different genetic backgrounds were responsible for the results requires additional studies. Notably, the correlations between PTEN expression and histological grade or distant metastasis were also noted in CRC tissues, which demonstrated that deletion or low expression of PTEN may not play a role in the primary stage of malignant transformation but may be involved elsewhere in the progression of cancer, such as aggravation of malignant differentiation and metastasis. Based on our present results, PTEN expression may not be a marker for early diagnosis of CRC but may have some potential to predict malignant aggravation and distant metastasis and may also be a potential prognostic marker in CRC, which requires verification by future predictive or prognostic statistical analyses. Moreover, the positive correlation between the expression of PPARγ and PTEN observed in our study may indicate the modulation relationship of these two proteins during colorectal carcinogenesis, which provides additional evidence to support the hypothesis that PTEN may be a downstream regulatory factor of PPARγ.

In this study, we elevated efficiency and reduced cost and labor in studies of target gene expression through the use of TMA, a new molecular biotechnology, when compared with immunostaining of conventional specimens. The cell morphology and protein expressions were studied in parallel and the variance in results under various experimental conditions currently experienced with conventional technology may be avoided. Therefore, it is feasible that CRC may be studied using TMA in combination with other molecular biotechnology in the near future. TMA technology, which is fast, convenient and economical, may have a potentially dominant position in macro-scale examination of tissue specimens.

Acknowledgements

This work is supported by the 135 Medical Important Talent Foundation of Jiangsu Province, China (No.37RC2002037) and the National 863 Project about Functional Genomic and Biochip (No. 2002AA2Z2021).

References