Expression of survivin in primary and metastatic gastric cancer cells obtained by laser capture microdissection

Abstract

AIM: Survivin, a recently identified member of the inhibitor of apoptosis protein family, is expressed during development and in various human cancers. However, its expression in normal tissues and clinical relevance in cancers are still debated. In the present study, we analyzed the expression of the survivin gene in human primary and metastatic gastric cancer cells as well as in paired epithelial cells from normal gastric mucosa by means of a novel laser capture microdissection (LCM) technique coupled with reverse transcription - polymerase chain reaction (RT-PCR).

METHODS: Thirty patients who had undergone gastrectomy with lymph node dissection for gastric cancer without preoperative treatments were included. Neoplastic tissue, metastatic lymph nodes, and apparently uninvolved normal tissue were collected from each patient. LCM-captured “pure” cell groups were respectively subjected to RT-PCR analysis with primers specific for the survivin gene.

RESULTS: Of the paired samples from 30 gastric cancer patients studied, 24 (80%) primary gastric cancer cell groups and 7 (23%) adjacent morphologically “normal” gastric epithelial cell groups were shown to have a detectable survivin expression. There was a statistically significant difference in suvivin expression between these two groups (P < 0.01). Meanwhile, 95% (19/20) of the metastatic gastric cancer cell groups from lymph nodes had a clear expression of the survivin gene. However, no significant correlation between survivin expression and clinicopathological features of gastric cancer was observed in the present study.

CONCLUSION: Survivin expression is present in the majority of gastric cancer cell groups obtained by LCM techniques. The high expression rate in metastatic lesions suggests a possible role of survivin in cancer invasiveness and metastasis. It may contribute to the detection of gastric cancer micrometastasis as a potential molecular marker. In addition, the high expression percentage renders survivin a potential target in the therapy for gastric cancer.

Key Words: $[Keywords]

INTRODUCTION

Alteration of the balance between apoptosis and cell proliferation could result in a disturbance of tissue homeostasis, and dysregulation of apoptosis is associated with various cancers[1]. Considerable interest has focused on the identification of regulators of apoptosis, which may potentially contribute to the development of cancer. Recently, a novel member of the inhibitor of apoptosis (IAP) protein family, designated as survivin, has been identified. The surviving gene is located on chromosome 17 and spans 14.7 kb, encoding a protein of 142 amino acids. Unlike other IAP proteins, survivin contains a single baculovirus IAP repeat and lacks a C-terminal RING finger[2]. Another significant feature of survivin is its unique expression pattern as one of the most tumor-specific human gene products[3]. While embryonic and fetal organs contain abundant survivin mRNA and protein, most terminally differentiated normal tissues do not. By contrast, dramatically strong expression of survivin is detected in the vast majority of tumors[4]. Moreover, the overexpression of survivin has been shown to be correlated with aggressive and histologically unfavorable neuroblastoma[5], and associated with the poor prognostic outcome in colorectal cancer[6], non-small-cell lung cancer[7], breast cancer[8], soft tissue sarcoma[9] and B-cell lymphoma[10].

However, since its identification five years ago, there have been some contradictory reports about the clinical relevance of survivin expression in cancer. Gianani et al[11] found the expression of survivin was not a specific marker of adenocarcinoma of the colon but showed characteristic patterns of expression in normal colonic mucosa. In gastric carcinoma which is the second most frequent cause of cancer death according to a worldwide estimation, Lu et al[12] showed that survivin could promote aberrantly tumor cell viability, whereas Okada et al[13] indicated that survivin expression in nuclei of tumor cells was predictive of favorable prognosis in patients. In the present study, we used laser capture microdissection (LCM), a novel technique that allows rapid, reliable and accurate procurement of cells from specific regions of tissue sections under direct visualization, to obtain the primary and metastatic gastric cancer cells as well as the matched normal epithelial cells from the same patient and analyzed mRNA expression of the survivin gene.

MATERIALS AND METHODS

Patients

Thirty patients (25 males and 5 females) underwent gastrectomy with lymph node dissection for gastric carcinoma at China Medical University between March and September of 2 000 were included in the study. None of the patients received preoperative chemotherapy. The patients ranged in age from 33 to 78 years (median, 57 years). Routinely, the resected specimens were histologically examined by H&E staining according to the general rules of the classification of gastric carcinoma suggested by the Japanese Research Society of Gastric Cancer[14]. Details of the Borrmann gross type, the Lauren histological classification, the Ming classification of growth patterns, grade of differentiation, and levels of lymph node metastasis were obtained from the operative records and the pathology reports. The pT classification representing the depth of wall invasion and the pN classification representing the extent of regional lymph node metastasis were performed using standard criteria of the 5^th TNM staging system[15]. The total lymph node number collected from the tissue was 15 or more for a reliable pN classification.

Tissue samples

Tumor tissues, lymph nodes, and apparently uninvolved normal tissues were collected from each patient immediately after surgical removal and snap-frozen in liquid nitrogen and kept at -80 °C until use. The tissues were then embedded in OCT for subsequent LCM harvest. Sections were cut at 8 μm with a cryostat and mounted on uncoated glass slides. The slides could be immediately stored at -80 °C for a few weeks. In order to exclude the areas of necrosis, tumor tissues from all patients were selected from the most viable areas of tumors. The lymph nodes were obtained according to the standard protocol. The number of removed and positive lymph nodes for various stations was documented by the pathologists. For the selection of matched uninvolved gastric mucosa, specimens were obtained from tissues at a distance of more than 10 cm from the tumor edge and confirmed histologically.

LCM

OCT embedded blocks of frozen tissue were serially cut. For each case, the first section was stained with H&E routinely for histological analysis and digitalized. The following serial sections stained with H&E according to the standard protocol recommended by NIH[16] were subjected to LCM using a LM200 system (Olympus, Japan/Arcturus Engineering Inc, US). Areas of interest were selected under microscopic guidance, and covered with ethylene vinyl thermoplastic (EVA) film mounted on optically transparent cap. The infrared laser was activated by the push of a button, which melts the film directly above the target cells. This melt caused a binding to form between the cells and the transfer film that was stronger than the binding between the cells and the slide[17]. The parameters used for LCM included a laser diameter of 7.5 μm, laser power of 50-60 mW. Five thousand laser pulse discharges per specimen were used to “capture” approximately 10 000 morphologically normal gastric epithelial cells, malignant primary gastric cancer cells and malignant metastatic (to a perigastric lymph node) gastric cancer cells from each case. Each population was estimated to be > 95% “homogeneous” as determined by microscopic visualization of the captured cells. The caps with captured cells were then fitted onto 0.5 mL microcentrifuge tubes containing 200 μL TRIZOL Reagent. Caps briefly placed onto the section without laser activation were used as negative control.

RNA extraction from LCM-captured cells

Total RNA from each population of laser-captured cells was independently extracted using TRIZOL Reagent (Life Technologies, Inc) by means of a modification of the RNA microisolation protocol recommended by the manufacturer. Briefly, tubes containing collected cells and TRIZOL Reagent were inverted and allowed to sit at room temperature for 30 min, then 0.5 μL glycogen (20 μg/μL) carrier and 40 μL chloroform were added to each tube. The tubes were vigorously shaken by hand for 15 s and incubated at room temperature for 2 min. The samples were centrifuged at 11 000 g for 15 min at 4 °C. The aqueous layer was transferred to a fresh 0.5 mL tube, 100 μL isopropanol was added and precipitated at -70 °C for 1 h. After centrifuged at 11 000 g for 10 min at 4 °C, the pellet was then washed in 200 μL of 750 mL/L ethanol and resuspended in 7 μL diethylpyrocarbonate-treated RNase free water.

Reverse transcription-polymerase chain reaction (RT-PCR)

For RT-PCR analysis of the LCM-captured cells, first strand cDNA was prepared from total RNA by using a first-strand synthesis kit (Life Technologies, Inc). Seven microliters of total RNA isolated from approximate 10 000 cells was mixed with 1 μL of oligo-(dT)_12-18 primer, 1 μL of random hexamers primer (N₆) and 1 μL of 10 mmol/L dNTPs in a total volume of 10 μL. They were heat denaturated at 65 °C for 5 min, then chilled in ice. Four microliters of 5 × first-strand reaction buffer, 2 μL of 0.1 mol/L DTT, 2 μL of 25 mmol/L MgCl₂, 1 μL of RNasin (40 U/μL) (Promega, Madison, WI, USA), and 1 μL of SUPERSCRIPT II RTase (200 U/μL) (Life Technologies, Inc) were added and incubated at 25 °C for 10 min, followed by incubation at 42 °C for 50 min for first-strand synthesis. The reverse transcriptase was inactivated at 70 °C for 15 min. PCR amplification with specific primers was performed in a final reaction volume of 50 μL containing 1 × PCR buffer, 200 μmol/L each dNTP, 0.5 μmol/L each primer, 1.25 unit of Taq polymerase (TaKaRa Biotech, Dalian, China) and 2 μL of RT product. Programmable temperature cycling (UNO II, Biometra, Germany) was performed using the following profiles: An initial hot start at 95 °C for 2 min, followed by 38 cycles at 95 °C for 30 s, at 58 °C for 20 s and at 72 °C for 30 s. After the last cycle, an elongation step was extended at 72 °C for 5 min. For each set of PCR, parallel reactions with human genomic DNA from peripheral blood were performed to test for genomic DNA contamination, and double distilled water instead of cDNA template was also included as negative control to assure the quality of PCR. An 8 μL aliquot of the PCR product underwent electrophoresis on 15 g/L agarose gel stained with ethidium bromide and was visualized under UV trans-illuminator. The following primers were used: (A) survivin gene, forward primer 5’-CCCTGCCTGGCAGCCCTTTC-3’, and reverse primer 5’-CTGGCTCCCAGCCTTCCA-3’ (PCR product, 188 bp); and(B) glyceraldehyde-3-phosphate dehydrogenase (G3PDH), forward primer 5’-AGGGGTCTACATGGCAACTG-3’,andreverseprimer 5’-CGACCACTTTGTCAAGCTCA-3’ (PCR product, 227 bp).

Statistical analysis

Statistical comparisons for significance between nominal variables were evaluated by the χ² test and the Fisher’s exact probability test. All the statistical analyses were performed using SPSS10.0 computer software and P of 0.05 was used as significance criterion.

RESULTS

Survivin expression in microdissected gastric epithelial and cancer cells

As shown in Figure 1, LCM allowed us to precisely sample the targeted cells. In this manner, gastric normal epithelial or cancer cells were homogenously obtained and bonded to the transfer film. When the cap together with the film was lifted from the tissue section, surrounding tissues remained attached to the glass slide. As a result, the morphology of the captured cells was well preserved and could be readily visualized under a microscope. To detect survivin gene mRNA expression by RT-PCR, a pair of specific primers located at 2 flanking exons was designed and synthesized. The expected specific transcript of survivin gene represented by an 188 bp fragment was amplified (Figure 2). Of the paired samples from 30 gastric cancer patients studied, 24 (80%) primary gastric cancer cell groups and 7 (23%) adjacent morphologically “normal” gastric epithelial cell groups were shown to have a detectable survivin expression. There was a statistically significant difference in suvivin expression between these two groups (P < 0.01). Furthermore, survivin mRNA was never detected in morphologically non-tumorous gastric epithelial cell groups when the paired cancer cell group was negative for survivin mRNA.

Figure 1 Laser capture microdissection of cells from gastric carcinoma. A: Morphology map view for pathological diagnosis (× 100). B: Group of gastric carcinoma cells were selected for LCM (× 200). C: The same section revealed where carcinoma cells were lifted from the section (× 200). D: The carcinoma cells were removed on the plastic film to provide a template for PCR.

Figure 2 Representative RT-PCR results of survivin expres-sion in normal (N), tumor (T) and metastatic carcinoma (M) cells obtained by LCM from 3 patients suffering gastric cancer. Case 1 was a patient without lymphatic metastasis, cases 2 and 3 were patients with lymphatic metastasis. G: Genomic DNA from human peripheral blood. W: Water as negative control.

Strikingly, 95% (19/20) of the metastatic gastric cancer cell groups from lymph nodes had a clear expression of the survivin gene. Three cases showed positive results for the surviving-specific RT-PCR assay in metastatic cancer cells while none of their paired primary cancer cells had a detectable expression of the survivin mRNA.

Relationship between survivin expression and clinicopathological characteristics in gastric cancer patients

We examined the relationship between survivin expression in microdissected primary cancer tissues and the clinicopathological features of gastric cancer (Table 1). There was no significant correlation between survivin expression and Borrmann gross type, depth of wall invasion, Lauren histological classification, Ming’s classification of growth patterns, grade of differentiation, the presence and extent of lymph node metastasis.

Table 1 Relationship between survivin expression and different histopathological features.

	Survivin expression in			Survivin expression pattern
Variant	primary cancer			in cancer and normal tissues
	Positive	Negative	1P	N(-)T(-)	N(-)T(+)	N(+)T(+)	P
Gross type
Borrmann I + II	7	1	2NS	1	4	3	NS
Borrmann III + IV	17	5		5	13	4
Depth of invasion
pT1 + pT2	9	1	NS	1	5	4	NS
pT3 + pT4	15	5		5	2	3
Differentiation
Well and medorate	7	1	NS	1	4	3	NS
Poor and undifferentiate	17	5		5	13	4
Lauren classification
Intestinal type	21	4	NS	4	16	5	NS
Diffuse type	3	2		2	1	2
Ming growth pattern
Expanding	9	3	NS	3	6	3	NS
Infiltrative	15	3		3	11	4
Lymph node metastasis
pN0	8	2	NS	2	5	3	NS
pN1	5	1		1	3	2
pN2 + pN3	11	3		3	9	2

¹P was estimated by the fFsher’s exact test

²NS, not significant.

We also analyzed the survivin gene expression pattern in both paired normal and primary cancer cells and its relationship with clinicopathological features. Thirty patients were divided into N (-) T (-), N (-) T (+) and N (+) T (+) groups. Still, no significant correlation between survivin expression pattern and the factors mentioned above was observed.

DISCUSSION

Apoptosis plays an important role in organ homeostasis, by eliminating senescent or damaged cells. The suppression of apoptosis has been considered to contribute to carcinogenesis and cancer progression by aberrantly prolonging cell viability with accumulation of mutations[18,19]. Survivin is a newly identified bifunctional protein that suppresses apoptosis and regulates cell division. Based on the previously published studies, its mRNA is abundantly expressed in fetal tissues, but not in normal adult tissues except placenta and thymus. Meanwhile, reactivation of survivin expression has been demonstrated in tumors of lung[7], stomach[20], breast[8], colon[21], esophagus[22], liver[23], pancreas[24], bladder[25], uterus[26], ovary[27], etc.. This specific distribution pattern has aroused great interest in making the survivin gene as a potential cancer therapeutic target. However, the expression of the survivin gene in normal tissues is currently debated. Several recent studies have shown survivin expression in normal tissues including skin[28], endometrium[29], endothelial cells[30], colonic mucosa[11] and muscle[9]. In the present studies, we sought to investigate whether survivin expression could be a specific marker during the development of gastric cancer.

It has been generally accepted that molecular analysis of neoplastic tissues in vivo is challenged by the heterogeneity of samples. Genetic alternations can be masked by contaminating bystander cells. Survivin mRNA was detected in normal tissue adjacent to soft tissue sarcoma cells[9]. Therefore, it is necessary to obtain a pure population of cells to evaluate the possible role of the survivin gene in tumor progression. The recent development in LCM, a highly sophisticated technique for the transfer of isolated pure cells from the histological slide into a reaction tube, could afford the opportunity to overcome this obstacle[31]. A few published studies and our previous data (unpublished data) have shown that LCM coupled with RT-PCR technique is a reliable method for the molecular analysis of gene profiles in specific tissues. We combined RT-PCR with LCM to determine the survivin expression in primary cancer cells, metastatic cancer cells, and paired normal epithelial cells.

Our results demonstrated that 23% of morphologically normal epithelial cells had a detectable survivin mRNA expression. By contrast, Lu et al[12] reported that no survivin expression was found in normal gastric mucosa neighboring the cancer cells in an immunohistochemical analysis. Meanwhile, the rate of survivin expression was relatively lower in normal tissue we examined than in previous studies reporting 47.1% in normal esophageal epithelial tissues[22] and 29.1% or 100% in normal colorectal tissues[6,11]. One possible explanation for these results could be the different sensitivity of methods and the varied criteria for positivity determination. Further studies may be required to check out whether the expression of the survivin gene in normal tissues is caused by a high proportion of mitotically active cells.

In accordance with the recently published reports showing 88% positive survivin expression in gastric cancer tissues, we found a detectable expression of the survivin gene in 24 of 30 primary gastric cancer cell groups. The percentage was significantly higher than that in normal gastric epithelial cells. These results suggested that survivin expression in gastric cancer was a quite common event and might play an important role in the carcinogenesis of stomach. Several studies have consistently shown that survivin expression could inhibit cell death induced by various apoptotic stimuli[4]. A role for survivin in blocking apoptosis has also been found in vivo[32]. Like other proteins in the IAP family, survivin might bind specifically to caspase-3 or caspase-9, which are the effectors of apoptosis[33]. On the other hand, survivin was expressed in the G2-M phase of the cell cycle and its overexpression might contribute to overcoming the G2-M phase checkpoint to enforce progression of cells through mitosis[34]. Therefore, not surprisingly, survivin was shown to be present in the majority of gastric cancer cells.

Even though no significant correlation was observed between survivin expression in primary cancer or its distribution pattern in normal and cancer tissues and tumor invasiveness or the presence of lymph node metastasis, it is still striking to find that 95% of the metastatic cancer cells from lymph nodes had a clear expression of the survivin mRNA. In three cases, survivin expression was detected in metastatic lesions, but not in primary cancer cell groups. One possible explanation might be that survivin expression was upregulated in some metastatic cancer cells. By sensitive quantitative RT-PCR assay, significantly increased survivin expression levels were found in soft tissue sarcomas with more aggressive biological behavior[9]. Alternatively, more survivin expressions detected in metastatic lesions could result from the presence of a higher proportion of survivn-expressing cells compared with that in primary sites. Immunohistochemical analysis revealed that the percentage of survivin-positive cells in primary gastric cancer was quite variable, ranging from 20%-100%[12]. It has been shown that failures in normal apoptosis pathways could contribute to cancer progression by supporting anchorage-independent survival during metastasis. Considering the anti-apoptotic function of survivin, it seems reasonable to deduce that survivin expression might facilitate the survival of cancer cells at distant sites[35]. Also, survivin played a crucial role in angiogenesis[30,36], which is considered as one of the critical steps for cancer cell spread and metastasis. Moreover, when analyzed retrospectively, cancer patients expressing survivin exhibited a shorter survival, correlated with unfavorable prognosis and accelerated rates of recurrences[5-10]. These results together with ours suggested a possible role of survivin in cancer invasiveness and metastasis. Further analysis is necessary to assess the expression levels in metastatic lesions by accurately quantitative methods.

Furthermore, the extremely high survivin expression rate in metastatic lesions indicated a possible approach using survivin as a molecular diagnostic marker for micrometastasis in gastric cancer. Our parallel studies have shown that the detection of survivin expression in peritoneal lavaged fluid is a specific and sensitive assay for prediction of peritoneal micrometastasis in gastric cancer. Also, survivin was detected in the urine of all patients tested with new or recurrent bladder cancer, whereas normal volunteers and patients with non-neoplastic diseases were tested negative for urine survivin[4]. However, a large scale and long follow-up should be carried out to validate these results. Recently, expression of a phosphorylation-defective survivin mutant (Thr34→Ala), which acted as a dominant-negative antagonist of survivin pathway and prevented phosphorylation of endogenous protein, has been reported to trigger apoptosis in several human melanoma cell lines in vitro and to prevent tumor formation or suppress the growth of existing tumors by 60%-70% in vivo[37]. Therefore, the role of survivin in cancer progression has made it an attractive therapeutic target against cancer development.

In summary, our data indicate survivin expression is present in the majority of gastric cancer cell groups obtained by LCM techniques. The high expression rate in metastatic lesions suggests a possible role of survivin in cancer invasiveness and metastasis. It might contribute to the detection of micrometastasis in gastric cancer as a potential molecular marker. In addition, the high expression percentage renders survivin a candidate target in the therapy for gastric cancer.