Invariant chain expression in gastric cancer

Cancer Letters 168 (2001) 87±91

www.elsevier.com/locate/canlet

Invariant chain expression in gastric cancer Sumiya Ishigami*, Shoji Natsugoe, Koki Tokuda, Akihiro Nakajo, Hirohumi Iwashige, Kuniaki Aridome, Shuichi Hokita, Takashi Aikou First Department of Surgery, Kagoshima University School of Medicine,8-35-1 Sakuragaoka, Kagoshima 890 Japan Received 31 January 2001; received in revised form 15 March 2001; accepted 15 March 2001

Abstract Invariant chain (Ii) is a chaperone molecule that inhibits the binding of endogenous antigens to HLA class II. The tumor cell with overexpressed Ii chain is thought to escape attacking cytotoxic lymphocytes by suppressing the host immune. However, the relationship between Ii expression by the tumor and clinicopathological factors in gastric cancer remains unclear. We studied 126 patients with gastric cancer who had undergone curative gastrectomy at Kagoshima University Hospital between 1988 and 1997. In order to detect Ii and HLA-DR expression by tumor cells, immunohistochemical staining with anti-CD74 and anti-HLA-DR antibodies were performed by avidin-biotin peroxidase complex method. The 126 patients studied were divided into two groups based on Ii expression. Ii and HLA-DR were expressed both on the surface and in the cytoplasm of tumor cells and tumor in®ltrating lymphocytes. A total of 48 patients were identi®ed as Ii positive, while the remaining 78 patients were Ii negative. Ii expression negatively correlated with the depth of invasion of the tumor as well as the patients' clinical stage. Ii expression was negatively correlated with HLA-DR expression. Patients with Ii negative expression had signi®cantly better surgical outcomes than those with Ii positive expression (P , 0.05). Ii expression in gastric cancer affected surgical outcome and Ii expression was negatively correlated with depth of invasion and HLA-DR expression. Ii expression in gastric cancer may be a prognostic factor related to suppressive effects on host immune responses to tumor cells. q 2001 Elsevier Science Ireland Ltd. All rights reserved. Keywords: Invariant chain; HLA-DR; Gastric cancer; Surgical outcome

1. Introduction Intracytoplasmic invariant chain (Ii) protein is a chaperone molecule which regulates the function of HLA class II molecules [1,2]. Ii protein regulates antigen presentation by HLA class II molecules by inhibiting their cytoplasmic form from binding to endogenously derived antigenic peptides. Recent work on the mechanisms involved in anti-tumor * Corresponding author. Tel.: 181-99-275-5361; fax: 181-99265-7426. E-mail address: [email protected] (S. Ishigami).

immune response has indicated that humoral mediators or abnormal expression of tumor-associated antigens result in low lymphocytic response against tumor cells [3±5]. It has been elucidated that Ii in cancer cells was one of the molecules related with immunosuppression [6]. Increased Ii expression in tumor cells inhibits tumor antigen presentation by HLA class II molecules, disrupting the HLA class II mediated immune stimulation by tumor cells [7]. Thus, Ii positive tumor cells may have a signi®cant in¯uence on tumor behavior. Previous studies on HLA-DR expression in several cancers including gastric cancers revealed a positive

0304-3835/01/$ - see front matter q 2001 Elsevier Science Ireland Ltd. All rights reserved. PII: S 0304-383 5(01)00503-1

88

S. Ishigami et al. / Cancer Letters 168 (2001) 87±91

relationship between Human Leukocyte Antigen-DR (HLA-DR) expression and prognosis of the disease [8±10]. However, clinicopathological characteristics correlated with Ii expression and its prognostic implication were not fully disclosed. We herein attempt to clarify, retrospectively, the clinicopathological features of Ii positive gastric cancer.

2. Materials and methods One hundred and twenty six patients who had undergone gastrectomy for gastric cancer between 1988 and 1995 at the First Department of Surgery, Kagoshima University School of Medicine were entered into this study. The mean age of the participants was 65.0 years (range 30±87), and the male / female ratio was 2.6 (91/35). One hundred and twelve patients received curative resection, while the remaining 14 underwent non-curative gastrectomy. A total of 60 patients received total gastrectomy, 51 underwent distal gastrectomy and the remaining 15 patients had proximal gastrectomy (Table 1). None of the patients received preoperative chemotherapy. Resected specimens were longitudinally sliced into 4 mm-thick sections, ®xed in 10% formalin solution and then embedded in paraf®n. Representative sections, having the deepest lesions, were prepared and stained with hematoxylin-eosin for pathological diagnosis. Pathological data was evaluated according to the General Rules for the Japanese Gastric Cancer Study and Pathology [11]. Antibodies against HLA-DR (DAKO, Denmark) and Ii protein (CD74, Biomarker, USA) were used

as primary antibodies in the immunohistochemical study. Representative specimens were deparaf®nized, soaked with PBS and incubated overnight with either anti-HLA-DR antibody diluted at 1:100 or anti-Ii antibody diluted at 1:200. Staining was visualized using a streptavidin-biotin-peroxidase supersensitive kit. For each subject, immunohistochemically positive cells in the tumor nest including the most invasive front of the tumor were counted in 20 points under high power ®elds (£400) and the percentage of positive cells was calculated. HLA-DR and Ii expressions were evaluated according to Jiang's classi®cation [6]; patients with 10% or more positive tumor cells were de®ned as HLA-DR or Ii positive and patients with less than 10% positive tumor cells were as HLA-DR or Ii negative. A x 2 test and t-test were used to analyze correlation between Ii expression and clinicopathological factors. The cumulative survival time of the patients was calculated by the Kaplan±Meier method and analyzed by the generalized Wilcoxson test. A P level of less than 0.05 was considered statistically signi®cant. 3. Results Both HLA-DR (Fig. 1) and Ii (Fig. 2) immunoreactivity was identi®ed on the surface of the tumor cells. Some populations of tumor in®ltrating lymphocytes were also immunopositive for those markers. Patients were divided into groups according to Ii or HLA-DR

Table 1 Patients background Sex Male Female Age Operation Total Distal Proximal Curativity yes no HLA-DR positive negative

91 35 65 60 51 15 112 14 46 80

Fig. 1. Immunohistochemical staining of HLA-DR in gastric cancer. HLA-DR expression is identi®ed on the surface of tumor cells and tumor in®ltrating lymphocytes.

S. Ishigami et al. / Cancer Letters 168 (2001) 87±91

Fig. 2. Immunohistochemical staining of Ii protein in gastric cancer. Ii expression is identi®ed on the surface and in cytoplasm of tumor cells and in®ltrating lymphocytes.

expression: 48 patients were placed in the Ii positive group, and 78 patients were placed in the Ii negative group. The HLA-DR positive group consisted of 46 patients while 80 patients were placed in the HLA-DR negative group. Clinicopathologically, patients with T1 gastric cancer had lower Ii positivity than more than T2 patients (P , 0:05). Ii expression was negatively correlated with HLA-DR expression Table 2 Clinicopathological features of Ii positive gastric cancer patients

Tumor location Upper Middle Lower Whole Depth of invasion t1 t2 t3~ Lymph node metastasis Yes No Lymphatic invasion Yes No Histology Differentiated Undifferentiated HLA-DR Positive Negative

Ii positive (n ˆ 48)

Ii negative (n ˆ 78)

P value

10 9 19 10

15 18 33 12

11 22 15

42 18 18

27 21

31 47

N.S.

33 15

41 37

N.S.

27 21

38 40

N.S.

19 29

27 51

P , 0.05

89

Fig. 3. Expression of Ii and HLA-DR in each T stage. Ii expression was increased according to the depth of invasion. Adversely HLADR expression were decreased.

(P , 0:05) (Table 2). A correlation between Ii, HLA-DR expression and depth of tumor invasion was more closely investigated. Twenty-one percent of pT1, 73% of pT2, 56% of pT3, and 50% of pT4 patients were positive for Ii expression. The percentage of Ii positive patients was gradually increased in accordance with depth of invasion, which was adverse result of HLA-DR expression (Fig. 3). The postoperative prognosis of Ii positive patients was signi®cantly poorer than that of Ii negative patients (P , 0:05) (Fig. 4). HLA-DR positive patients seemed to have better survival rates than HLA-DR negative patients, although the difference was not signi®cant (Fig. 5).

N.S.

P , 0.05

Fig. 4. Survival curves of gastric cancer 112 patients received curative gastrectomy according to Ii expression. Ii negative patients (n ˆ 72) have better survival rates than Ii positive patients (P , 0:05).

90

S. Ishigami et al. / Cancer Letters 168 (2001) 87±91

Fig. 5. Survival curves of 112 gastric cancer patients with curative operation according to HLA-DR expression. There is no signi®cant difference in survival rates between the two HLA-DR groups.

4. Discussion It has been suggested that increased Ii expression in tumor cells prevents presentation of tumor antigens by HLA class II molecules [6]. Poor antigen presentation to intratumoral T-cells may then weaken the host immune defense against tumor cells. There are few clinical studies on Ii expression in cancer patients [6,12] and the results are not always similar, indicating the need for further investigation. This study describes Ii expression in gastric cancer and, for the ®rst time, shows a correlation between Ii expression and surgical outcome. For several clinicopathological factors, only depth of invasion positively correlated with Ii expression in gastric cancer patients. Jiang et al. [6] demonstrated that Ii expression correlated with tumor differentiation and depth of invasion in colon cancer, concluding that increased Ii rendered the tumour less immunogenic and less likely to stimulate a host immune response [6]. Our ®ndings in gastric cancer also support a potential role of Ii expression in tumor progression. However, we could not demonstrate a relationship between Ii expression and histological differentiation, possibly because the variety of histology of gastric cancer makes the analysis more complex than colon cancer [6]. Saito et al. [12] previously reported the expression of Ii and HLA-DR proteins in a renal cancer cell line and demonstrated that interferon-g treatment induced expression of both Ii and MHC-class II molecules by

tumor cells. Immunosuppressive cytokines which are released from tumor cells may affect the degree of Ii or HLA-class II expression. We have demonstrated that Ii expression is signi®cantly correlated with survival in gastric cancer patients. Ii expression has been reported to correlate with the clinical stage of cancer patients [6], however there is no report to date on the prognostic value of Ii expression in cancer cells. We have also found that HLA-DR expression was not a prognostic factor in patients with a curative operation, which is in agreement with previous reports [8,9]. Thus, in gastric cancer, Ii expression may play a clinically more signi®cant role in HLA class II related immunity than HLA-DR expression. Ostrand±Rosenberg demonstrated that tumor cells transfected with both MHC class II and Ii were not rejected by autologous T-cells, although tumor cells transfected with only MHC class II were easily rejected [13], suggesting that either the existence of Ii or loss of MHC class II in tumor cells could interfere with the immunologic interaction of tumor cells and in®ltrative lymphocytes. The current study suggests HLA-DR expression without coexistence of Ii may be a favorable immunological condition to eliminate autologous tumor cells. To present tumor antigens with HLA-DR, inhibition of Ii expression by the antisense method can be an effective approach in tumor immunotherapy [14]. In conclusion, Ii expression in gastric cancer is a useful prognostic marker related with HLA class II immune system.

References [1] H. Bormer, S. Viville, C. Benoist, D. Mathis, Diversity of endogenous epitopes bound to MHC class II molecules limited to invariant chain, Science 263 (1994) 1284±1286. [2] S. Ceman, A.J. Sant, The function of invariant chain in class II-restricted antigen presentation, Semin. Immunol. 7 (1995) 373±406. [3] D.R. Leach, M.F. Krummel, J.P. Allison, Enhancement of antitumor immunity by CTLA-4 blockade, Science 271 (1996) 1734±1736. [4] T.R. Mosmann, K.W. Moore, The role of IL-10 in crossregulation of TH1 and TH2 responses, Immunol. Today 12 (1991) A49±A53. [5] H. Mizoguchi, J.J. Oshea, D.L. Longo, C.M. Loef¯er, D.W. Mc Vicar, A.C. Ochoa, Alternations in signal transduction

S. Ishigami et al. / Cancer Letters 168 (2001) 87±91

[6] [7]

[8]

[9] [10]

molecules in T lymphocytes from tumor-bearing mice, Science 258 (1992) 1795±1798. Z. Jiang, M. Xu, L. Savas, P. Leclair, B.F. Banner, Invariant chain statement in colon neoplasms, Virchows Arch. 435 (1999) 32±36. V.K. Clements, S. Baskar, T.D. Armstrong, S. Ostand-Rosenberg, Invariant chain alters the malignant phenotype of MHC class II 1 tumor cells, J. Immunol. 149 (1992) 2391± 2396. S. Ishigami, T. Aikou, S. Natsugoe, S. Hokita, H. Iwashige, M. Tokushige, S. Sonoda, Prognostic value of HLA-DR statement and dendritic cell in®ltration in gastric cancer, Oncology 55 (1998) 65±69. D.A. Hilton, K.P. West, A review of the prognostic signi®cance of HLA-DR statement in gastric carcinoma, Cancer 15 (66) (1990) 1154±1157. T. Carbrera, F. Ruiz-Cabello, F. Garrido, Biological implica-

[11] [12]

[13]

[14]

91

tion of HLA-DR statement in tumors, Scand. J. Immunol. 41 (1995) 398±406. Japanese Gastric Cancer Association, Japanese Classi®cation of Gastric Carcinoma 2nd English Edition, Gastric Cancer 1 (1998) 10±24. T. Saito, M. Kimura, T. Kawasaki, S. Sato, Y. Tomita, MHC class II antigen-associated invariant chain on renal cell cancer may contribute to the anti-tumor immune response of the host, Cancer Lett. 109 (1996) 15±21. T.D. Armstrong, V.K. Clements, S. Ostrand-Rosenberg, Class II-transfected tumor cells directly present endogenous antigen to CD4 1 T cells in vitro and are APCs for tumor-encoded antigens in vivo, J. Immunother. 21 (1998) 218±224. G. Qiu, J. Goodchild, R.E. Humphreys, M. Xu, Cancer immunotherapy by antisense suppression of Ii protein in MHCclass-II-positive tumor cells, Cancer Immunol Immunother. 48 (1999) 499±506.