Disseminated cancer cells in the blood and expression of sialylated antigen in gastric cancer

Cancer Letters 200 (2003) 77–83 www.elsevier.com/locate/canlet

Disseminated cancer cells in the blood and expression of sialylated antigen in gastric cancer Shin-ichi Sumikura, Sumiya Ishigami*, Shoji Natsugoe, Futoshi Miyazono, Koki Tokuda, Akihiro Nakajo, Hiroshi Okumura, Masataka Matsumoto, Shuichi Hokita, Takashi Aikou First Department of Surgery, Kagoshima University School of Medicine, 8-35-1 Sakuragaoka, Kagoshima 890, Japan Received 19 March 2003; received in revised form 22 May 2003; accepted 26 May 2003

Abstract Background: In gastric cancer, disseminated cancer cells (DCC) can be detected in peripheral blood using bio-molecular techniques. It is known that patients having DCC exhibit a high occurrence of postoperative relapse in gastrointestinal cancer. However, more than half of gastric cancer patients having positive DCC do not show cancer relapse. Sialylated Lewis antigens are considered to be crucial molecules in the metastasis of disseminated cancer. The current study investigated whether combination analysis of DCC and sialylated Lewis antigen are useful in estimating the recurrence risk of gastric cancer. Patients and methods: Subjects were 106 consecutive gastric cancer patients who underwent curative gastrectomy. DCC in the peripheral blood were detected using the carcinoembryonic antigen (CEA)-mRNA by RT-PCR method. Sialylated Lewis antigen expression (sLeA and sLeX) of the primary tumor was assessed immunohistochemically. Results: Of 106 gastric cancer patients, 43 (40%) were positive for DCC. Immunohistochemically, 53 (50%) and 49 (46%) patients were positive for sLeA and sLeX, respectively. The presence of DCC did not correlate with sLeA and sLeX expression in gastric cancer. Postoperative tumors were present in 19 patients (7 hematogenous and 12 non-hematogenous), 12 of which were positive for DCC. Six sLeApositive patients (26%) with DCC and 13 sLeX-positive patients (57%) with DCC suffered from postoperative recurrence of gastric cancer. The p value of CEA-mRNA and sLeX combination analysis was more significant ðp , 0:01Þ than that of CEAmRNA alone ðp ¼ 0:02Þ: Conclusion: Analyzing both DCC and sLeX expression in gastric cancer may enable more accurate prediction of postoperative recurrence. q 2003 Elsevier Ireland Ltd. All rights reserved. Keywords: Disseminated cancer cells; Sialylated antigen; Gastric cancer

1. Introduction

* Corresponding author. Tel.: þ81-99-275-5361; fax: þ 81-99265-7426. E-mail address: [email protected] (S. Ishigami).

Disseminated cancer cells (DCC) are widely detected in the blood, peritoneal washing, bone marrow, and lymph nodes [1,2] of cancer patients. We previously reported that DCC are associated with a high occurrence of hematogenous metastasis in

0304-3835/03/$ - see front matter q 2003 Elsevier Ireland Ltd. All rights reserved. doi:10.1016/S0304-3835(03)00388-4

78

S.-i. Sumikura et al. / Cancer Letters 200 (2003) 77–83

gastric [3] and pancreatic cancer patients [4]. However, we have also observed that in gastric cancer, approximately half of patients with DCC showed no relapse. We can speculate that not all DCC in the blood are viable or possess the ability to metastasize to distant organs. Assessing the potential of DCC seeding distant organs in DCC-positive patients appears to be important, but very few studies have directly evaluated DCC. Sialylated antigen is an adhesion molecule, and expression of this antigen in gastrointestinal cancer cells is reportedly correlated with tumor progression and distant metastasis to the liver [5 –7]. DCC adhering to endothelial cells via sialylated antigen is the initial stage in distant metastasis to liver [8]. It may therefore be interesting to clarify the clinical implications of sialylated antigen expression in DCC-positive cancer patients [9]. In the current study, we evaluated the risk of postoperative distant metastasis based on the presence of DCC in the blood as well as expression of sialylated Lewis A (sLeA) and sialylated Lewis X (sLeX) antigens in gastric cancer.

Table 1 Patients’ background Age Sex Location

Depth of invasion

Nodal involvement Histology Operation

sLeA expression sLeX expression

Male Female Upper Middle Lower Whole pT1 pT2 pT3Yes No Differentiated Undifferentiated Total Distal Proximal Yes No Yes No

63 ^ 34 (median 60) 77 29 30 30 32 14 44 33 29 49 57 50 56 51 47 8 53 53 49 57

classification were assessed according to the General Rules for the Japanese Gastric Cancer Study and Pathology [10].

2. Materials and methods 2.1. Patients

2.2. RNA extraction and detection of disseminated cancer cells in the blood by PCR

A total of 106 gastric cancer patients who underwent curative gastrectomy for gastric cancer at Kagoshima University Hospital between 1997 and 2001 were enrolled in the current study. The mean age of this patient population was 63.3 years (range 30– 87; median 60 years), and the male/female ratio was 2.7 (77/29). None of the patients received preoperative chemotherapy. Of the 106 patients, 51 underwent total gastrectomy, 47 underwent distal gastrectomy, and remaining 8 underwent proximal or other gastrectomy (Table 1). During operation, venous blood samples were obtained and the nuclear cell fraction was purified using Monopoly resolving medium (Dainippon Pharmaceutical Co., Ltd, Osaka, Japan). The resected tumor was sliced into 4 mm-thick sections, fixed in buffered formalin and embedded in paraffin. Embedded sections were then used for clinicopathological examination and sialylated antigen detection. Clinical data and pathological

Blood sample treatment and detection of DCC by PCR were performed as previously reported [3,4]. Briefly, total RNA was extracted from a 5 ml sample of peripheral blood using ISOGEN-LS (Nippon Gene, Toyama, Japan). Contaminant genomic DNA was removed by treating the solution with RNase-free DNase. Total RNA was denatured for 10 min and chilled quickly. Twenty microliters of cDNA was then synthesized in a solution of 5 £ buffer, 2 mM dithiothreitol, 200 U of superscript III, 4 U of RNase inhibitor, 2.5 mM random hexamer, and 0.5 mM dNTP. The reaction mixture was incubated for 50 min at 42 8C. Using the cDNA as a template, carcinoembryonic antigen (CEA) was also amplified by nested PCR as previously reported [3,4]. To monitor the synthesis of cDNA, glyceraldehyde-3-phosphatase dehydrogenase (GAPDH) RT-PCR was performed. A total of 50 ml PCR mix contained 0.5 ml cDNA, 5 ml 10 £ PCR

S.-i. Sumikura et al. / Cancer Letters 200 (2003) 77–83

buffer, 200 mM dNTP mixture, 0.2 mM forward primer, 0.2 mM reverse primer, and 1.25 units Taq DNA polymerase (all Takara Shuzo Co., Ltd, Otsu, Japan). Denaturation for 4 min at 95 8C was performed. Amplification was performed with denaturation for 1 min at 95 8C, annealing for 1 min at 57 8C, extension for 50 s at 72 8C for 40 cycles, followed by a final extension for 9 min at 72 8C. The RT-PCR product was detected as a 390 base pair (bp) fragment. To detect the more specific CEA mRNA, we performed nested RT-PCR. We used three types of CEA-primers designed by Gerhard et al. CEA-A primer was 50 -TCTGGAACTTCTCCTGGTCTCTCAGCTGG-30 ; CEA-B primer was 50 -TGTAGCTG TTGCAAATGCTTTAAGGAAGAAGC-30 ; and CEA-C primer was 50 -GGGCCACTGTCGGCATCATGATTGG-30 . In the first RT-PCR, 20 ml reaction mix contained 0.5 ml cDNA, 2 ml 10 £ PCR buffer, 200 mM dNTP mixture, 0.2 mM CEA primer A and primer B, and 0.5 unit Taq DNA polymerase (all Takara Shuzo Co., Ltd). Denaturation for 4 min at 95 8C was performed. Amplification was performed with denaturation for 30 s at 95 8C, annealing for 1 min at 69 8C, extension for 30 s at 72 8C for 30 cycles, followed by a final extension for 6 min at 72 8C. In nested RT-PCR, 40 ml reaction mix contained 1 ml the first PCR product, 4 ml 10 £ PCR buffer, 200 mM dNTP mixture, 0.2 mM CEA primer B and primer C, and 1 unit Taq DNA polymerase (all Takara Shuzo Co., Ltd). Denaturation for 4 min at 95 8C was performed. Amplification was performed with denaturation for 30 s at 95 8C, annealing for 1 min at 64 8C, extension for 30 s at 72 8C for 32 cycles, followed by a final extension for 6 min at 72 8C. The CEA RT-PCR product was separated by 2% agarose-gel electrophoresis in Tris-acetate with EDTA (TAE) buffer and visualized by ethidium-bromide staining. The RT-PCR product was detected as a 131 bp fragment. 2.3. Evaluation of sialylated antigen expression in the cancer nest Paraffin-embedded tumor sections were sliced at 4 mm. After deparaffinization, sections were soaked with PBS. Endogenous peroxidase activity was blocked, and sections were incubated at 4 8C overnight with monoclonal antibodies against sLeA

79

(clone:N19-9) or sLeX (clone:SNH3) at dilutions of 200. Sections were washed with Tris – buffered saline, incubated with biotinylated immunoglobin M for 30 min and then reacted with biotin – peroxidase complex for 30 min. Peroxidase activity was quantified with diaminobenzidine tetrahydrochloride. A negative control assay was conducted in the absence of anti-Sialylated antigen body. Sialylated antigen positivity was classified according to staining intensity (negative: less than 25% positive cells; positive group: more than 25% positive cells), as previously reported [10]. 2.4. Clinical follow up and classification of relapse form Every one to three months, patients were followed up by physical examination, routine blood tests, serum tumor marker test and abdominal CT examination. Postoperative follow-up data were obtained in all patients, with the median follow-up period being 21 months (range 12 –60 months). Patients who died of causes unrelated to cancer were not included in the current study. Relapse form of cancer was determined based on the information present on the medical chart or pathological examination. 2.5. Statistical analysis Statistical analysis of clinical features and comparison between two unmatched groups was conducted by x2 test. A p value of less than 0.05 was considered statistically significant.

3. Results CEA-mRNA was detected in the MCF-7 cell line and indicated a concentration of one tumor cell per 106 lymphocytes. No CEA-mRNA signal was detected in patients with benign disease or in healthy volunteers (data not shown). Forty-three patients (41%) were positive for CEA-mRNA. CEA-mRNA positivity was significantly correlated with the depth of invasion, nodal involvement, lymphatic invasion, and venous invasion (Table 2). Sialyl Lewis A and sialyl Lewis X antigen were detected on the surface of cancer cells (Fig. 1).

80

S.-i. Sumikura et al. / Cancer Letters 200 (2003) 77–83

Table 2 Clinicopathological features of 106 gastric cancer with or without intravenous disseminated cancer cell CEA mRNA

Sex Depth of invasion

Nodal involvement Lymphatic invasion Venous invasion Histology sLeA expression sLeX expression

Male Female pT1 pT2 pT3Yes No Yes No Yes No Differentiated Undifferentiated Yes No Yes No

Positive ðn ¼ 43Þ

Negative (n ¼ 63)

p-value

30 13 9 18 16 27 16 33 10 22 21 18 25 23 20 23 20

47 16 35 15 13 22 41 32 31 17 46 32 31 30 33 26 37

n.s.

According to the former criteria, patients were divided into 53 (50%) sLeA-positive and 46 (43%) sLeXpositive patients. Expression of neither sialylated antigen was correlated with CEA mRNA positivity (Table 2). Expression of sLeX was positively correlated with lymph node metastasis ðp , 0:01Þ; lymphatic invasion

p , 0:01

p , 0:01 p , 0:05 p , 0:01 n.s. n.s. n.s.

ðp , 0:05Þ and venous invasion ðp , 0:01Þ; whereas sLeA expression did not reflect any clinicopathological features (Table 3). Of 106 patients, 19 suffered from a recurrence of gastric cancer. Seven of these were concomitant with hematogenous metastasis (1: lung, 6: liver) and 12 were without hematogenous metastasis. The rates of

Fig. 1. sLeA and sLeX were expressed on the surface and cytoplasm of cancer cells.

S.-i. Sumikura et al. / Cancer Letters 200 (2003) 77–83

81

Table 3 Clinicopathological features of 106 gastric cancer with or without expression of sialylated antigens SleX

Sex Depth of invasion

Nodal involvement Lymphatic invasion Venous invasion Histology

Male Female pT1 pT2 pT3Yes No Yes No Yes No Differentiated Undifferentiated

SleA

Positive ðn ¼ 49Þ

Negative ðn ¼ 57Þ

Positive ðn ¼ 53Þ

Negative ðn ¼ 53Þ

33 16 17 18 14 **29 20 *35 14 *24 25 27 22

44 13 27 15 15 **20 37 *30 27 **15 42 28 29

38 15 23 16 14 23 30 32 21 21 32 25 28

39 14 21 17 15 27 26 33 20 18 35 25 28

* p , 0:05; * * p , 0:01:

CEA-mRNA positivity in these two groups were 86 and 58%, respectively. Of 43 CEA-mRNA-positive patients, 30 did not show cancer relapse (Table 4). Twenty-three patients were classified as both CEAmRNA and sLeA positive, and 23 were classified as both CEA-mRNA and sLeX positive. Thirteen (30%) of the CEA-mRNA-positive patients, 6 (26%) of the CEA-mRNA- and sLeA-positive patients and 13 (57%) of the CEA-mRNA- and sLeX-positive patients showed a postoperative recurrence of gastric cancer. The p value of CEA-mRNA and sLeX analysis was definitely obviously significant (CEAmRNA alone; p ¼ 0:02; CEA-mRNA and sLeX; not significant, CEA-mRNA and sLeX; p , 0:001;

respectively) (Table 5). The 13 patients positive for both DCC and sLeX suffered from postoperative gastric cancer relapse. However, hematogenous metastasis occurred in 6 (45%) patients, while Table 5 Occurrence of relapse in mRNA positive patients with or without sLeA and sLeX CEA-mRNA expression Yes

No

CEA-mRNA alone (detection rate: 13/43, 30%) Recurrence 13 6 No recurrence 30 57

p ¼ 0:02

CEA-mRNA expression þ SLeA positive Table 4 Correlation between recurrence pattern and CEA-mRNA expression CEA-mRNA expression

Concomitant hematogenous metastasis Without hematogenous metastasis No recurrence

Yes SLeA (detection rate: 6/23, 26%) Recurrence 6 No recurrence 17

No

13 70

Detection rate (%)

CEA-mRNA expression þ SLeX positive

1

86

Yes

7

5

58

30

57

Yes

No

6

SLeX (detection rate: 13/23, 57%) Recurrence 13 No recurrence 10

n.s

No

6 77

p , 0:001

82

S.-i. Sumikura et al. / Cancer Letters 200 (2003) 77–83

the remaining patients died of peritoneal dissemination or lymph node metastasis.

prophylactic chemotherapy in order to prevent postoperative relapse.

4. Discussion

References

The presence of DCC in the blood, which is confirmed by PCR, is regarded as one of the critical prognostic factors in cancer patients [11]. However, some patients positive for DCC show no relapse of cancer [3]. The detection rate of DCC in the peripheral blood increases during surgery, and it is believed that surgical procedures promote the scattering of cancer cells into the blood. In order to understand the metastatic ability of DCC, the characteristics of the cells themselves must be examined. However, direct investigation of DCC is difficult. We therefore examined the expression of sialylated antigen in the primary gastric cancer site, which may indirectly reflect the characteristics of DCC. Evaluation of DCC together with sialylated antigens remarkably improves the detection rate of postoperative gastric cancer relapse. Ichikawa reported similar results in 36 colon cancer patients [9]. In the current study, sLeX was correlated with venous invasion, and sLeX expression was previously reported to be an independent prognostic factor in cancer patients [10,12 – 14]. We clarified that the presence of DCC is not correlated with sLeA or sLeX expression. The presence of DCC may indicate a different mechanism of cancer metastasis than that indicated by sialylated antigen expression. Analyzing the combination of DCC and sLeX expression is the best way to detect postoperative recurrence in cancer patients. Patients positive for DCC and sLeX were at risk for non-hematogenous metastasis [10]. Multiple organs in these patients may be already involved by cancer cells. Nakagoe observed that some sLeX-positive early gastric cancer patients died of non-hematogenous metastasis, and this observation is concurrent with our results [15]. In conclusion, molecular examination of sLeX expression in the primary tumor in addition to DCC status may enable more precise prediction of postoperative relapse in gastric cancer patients. Individuals positive for both DCC and sLeX may require

[1] M. Mori, K. Mimori, H. Ueno, N. Karimine, G.F. Barnard, K. Sugimachi, T. Akiyoshi, Molecular detection of circulating solid carcinoma cells in the peripheral blood: the concept of early systemic disease, Int. J. Cancer 68 (1996) 739–743. [2] M. Gerhard, H. Juhl, H. Kalthoff, H.W. Schreiber, C. Wagener, M. Neumaier, Specific detection of carcinoembryonic antigen expressing tumor cells in bone marrow aspirates by polymerase chain reaction, J. Clin. Oncol. 12 (1994) 725–729. [3] F. Miyazono, S. Natsugoe, S. Takao, K. Tokuda, F. Kijima, K. Aridome, S. Hokita, M. Baba, Y. Eizuru, T. Aikou, Surgical maneuvers enhance molecular detection of circulating tumor cells during gastric cancer surgery, Ann. Surg. 233 (2001) 189 –194. [4] F. Miyazono, S. Takao, S. Natsugoe, K. Uchikura, F. Kijima, K. Aridome, H. Shinchi, T. Aikou, Molecular detection of circulating cancer cells during surgery in patients with biliarypancreatic cancer, Am. J. Surg. 177 (1999) 475–479. [5] M.L. Phillips, E. Nudelman, F.C. Gaeta, M. Perez, A.K. Singhal, S. Hokomori, J.C. Paulson, ELAM-1 mediates cell adhesion by recognition of carbohydrate ligand, sialyl-Lex, Science 250 (1990) 1130–1132. [6] A. Takada, T. Ohmori, K. Yoneda, A. Tsuyuoka, A. Hasegawa, M. Kiso, R. Kannagi, Contribution of carbohydrate antigen sialyl Lewis A and sialyl Lewis X to adhesion of human cancer cells to vascular endothelium, Cancer Res. 53 (1993) 354 –356. [7] M. Tatsumi, A. Watanabe, H. Sawada, Y. Yamada, Y. Shino, H. Nakao, Immunohistochemical expression of the sialyl Lewis x antigen on gastric cancer cells correlates with the presence of liver metastasis, Clin. Exp. Metastasis 16 (1998) 743 –750. [8] S. Hakomori, Tumor malignancy defined by aberrant glycosylation and shingolipid metabolism, Cancer Res. 56 (1996) 5309–5318. [9] D. Ichikawa, K. Kitamura, N. Tani, S. Nishida, H. Tsurutome, S. Hakomori, E. Ikeda, F. Mutoh, H. Kurioka, H. Yamagishi, Molecular detection of disseminated cancer cells in the peripheral blood and expression of sialylated antigens in colon cancers, J. Surg. Oncol. 75 (2000) 98–102. [10] Japanese Gastric Cancer Association, Japanese Classification of Gastric Carcinoma, Gastric Cancer, 2nd English Edition., Vol. 1, 1998, pp. 10–24. [11] S.B. Hosch, J. Kraus, P. Scheunemann, J.R. Izbicki, C. Schneider, U. Schumacher, K. Witter, M.R. Speicher, K. Pantel, Malignant potential and cytogenetic characteristics of occult disseminated tumor cells in esophageal cancer, Cancer Res. 60 (2000) 6836– 6840. [12] T. Nakagoe, K. Fukushima, M. Hirota, H. Kusano, H. Ayabe, M. Tomita, S. Kamihara, Immunohistochemical expression of

S.-i. Sumikura et al. / Cancer Letters 200 (2003) 77–83 sialyl Lex antigen in relation to survival of patients with colorectal carcinoma, Cancer 72 (1993) 2323–2330. [13] N. Futamura, S. Nakamura, M. Tatematsu, Y. Yamamura, R. Kannagi, H. Hirose, Clinicopathologic significance of sialyl Le(x) expression in advanced gastric carcinoma, Br. J. Cancer 83 (2000) 1681– 1687. [14] M. Amado, F. Carneiro, M. Seixas, H. Clausen, Dimeric sialyl-Lex expression in gastric carcinoma correlated with

83

venous invasion and poor outcome, Gastroenterology 114 (1998) 462 –470. [15] T. Nakagoe, K. Fukushima, T. Sawai, T. Tsuji, M. Jibiki, A. Nanashima, K. Tanaka, H. Yamaguchi, T. Yasutake, H. Ayabe, K. Arisawa, H. Ishikawa, Increased expression of sialyl Lewis(x) antigen as a prognostic factor in patients with stage 0, I, and II gastric cancer, Cancer Lett. 175 (2002) 213 –221.