A Case Report of a Spontaneous Gastrointestinal Stromal Tumor (GIST) Occurring in a F344 Rat

Abstract

We report a case of gastrointestinal stromal tumor (GIST) that developed in a male F344 rat at week 101 of an experiment in a carcinogenicity study. Macroscopically, the primary tumor, which measured 1 cm in diameter, involved the submucosal tissue of the forestomach at the lesser curvature extending to the glandular stomach and esophagus. Histopathologically, the tumor was composed of neoplastic cells with small- to medium-sized spindle-shaped single nuclei and fibrillary cytoplasm lacking distinct cell borders. It invaded extensively into the tunica muscularis and subserosa, further extending to the lamina propria mucosa and serosal surface. A few densely proliferating portions showed a tendency to storiform pattern. Metastatic tumor nodules were found in the liver, spleen, and femur bone marrow, with multiple nodules, up to 1 cm in diameter, apparent in the liver. Immunohistochemically, diffuse, but weak cytoplasmic immunoreactivity for KIT was evident, and most neoplastic cells also exhibited strong immunoreactivity for α-smooth muscle actin and vimentin. Sparse nuclear S-100-immunoreactive cells were further observed, but none of neoplastic cells were immunoreactive for CD34, caldesmon, desmin, cytokeratin, or synaptophysin. Collectively, these features meet the criteria for a GIST, with limited potential for differentiation to smooth muscle and neural cells.

Introduction

In humans, most nonepithelial tumors arising in the gastrointestinal tract are composed mainly of spindle-shaped cells and have been collectively referred to as gastrointestinal stromal tumors (GISTs). This category was originally applied to tumors not presenting authentic morphological characteristics of leiomyomas or neurinomas (Mazur and Clark, 1983). In 1996, Rosai proposed GIST as an entity for a broad range of submucosal mesenchymal tumors and introduced criteria for their classification into smooth muscle, neural, combined smooth muscle-neural, and uncommitted types. Recently, most tumors of uncommitted type in Rosai’s classification, i.e., those not derived from smooth muscle cells or Schwann cells, have demonstrated characteristic immunoreactivity for KIT (product of c-kit oncogene) in more than 90% of the cases (Sircar et al., 1999; Robinson et al., 2000; Miettinen et al., 2005) and an undifferentiated mesenchymal antigen CD34 in 70–80% of the cases (Moniham et al., 1994; Miettinen et al., 1995).

Based on these features, GISTs are recognized as a group of nonepithelial gastrointestinal tumors expressing KIT and CD34 in the recent WHO classification (Miettinen et al., 2000). However, lack of immunoreactivity for either KIT or CD34 (Hirota et al., 1998), or positive immunoreactivity for both smooth muscle and neural markers (Rosai, 2003), has sometimes been encountered in small populations within tumors in this category. In addition, GIST cases immunoreactive for caldesmon, an actin and tropomyosin binding protein used as a smooth muscle cell marker, are not infrequent (Miettinen et al., 1999; Tazawa et al., 1999; Orosz et al., 2005). We are not aware of any reports of spontaneous GISTs in experimental animals in which immunohistochemical studies were performed. In the present paper, we therefore document immunohistochemical analysis of a gastric submucosal tumor developing in a male rat at week 101 of an experiment in a 2-year carcinogenicity study.

Case Report

Four-week-old male and female F344 rats were purchased from Charles River Japan Inc. (Kanagawa, Japan) for a 2-year carcinogenicity study of Madder color, a natural ingredient of madder root derived from Rubia tinctorum LINNE and that has been used as an existing food additive in Japan. The case was found in 1 of the 50 males of the middle-dose group (2.5% in diet). Animals were housed in polycarbonate cages on wood-chip bedding and maintained in an air-conditioned animal room (temperature 24 ± 1°C, relative humidity: 55 ± 5%) with a 12-hour light/dark cycle. They were allowed ad libitum access to feed and tap water. CRF-1, a regular rodent diet, obtained from Oriental Yeast Co. Ltd. (Tokyo, Japan) was employed as the basal diet for the feeding study of Madder color (a mixture of products of 2 companies: San-Ei Gen F.F.I., Osaka, Japan; OCI Co., Ltd., Kobe, Japan). The animal protocol was reviewed and approved by the Animal Care and Use Committee of the National Institute of Health Sciences, Japan.

The animal in question was euthanized and necropsied at 101 weeks because of cachexia with adoption of a stooping position and apparent hypoactivity. The necropsy revealed a 1 cm tumor mass in the lesser curvature of the forestomach with extension into glandular stomach and esophagus (Figure 1a). The mass was firm and the cut surface whitish. Grossly, metastatic deposits were found in the liver and spleen (Figure 1b). The primary tumor, metastatic deposits and all other organs were fixed in 10% buffered neutral formalin (pH 7.4), processed routinely for embedding in paraffin, sectioned at 4 μm, and then stained with hematoxylin and eosin for histological examination.

Sections of the gastric submucosal tumor and metastatic deposits in the liver and spleen were immunohistochemically stained with antibodies to KIT (affinity purified rabbit antibody, x50 dilution; DakoCytomation, Glostrup, Denmark), S-100 (rabbit immunoglobulin, without dilution; DakoCytomation), α-smooth muscle actin (α-SMA; mouse IgG_2aκ, clone 1A4, × 200 dilution; DakoCytomation), vimentin (mouse IgG₁κ, clone V9, × 100 dilution; DakoCytomation), CD34 (mouse IgG_2a, clone MEC 14.7, ×10 dilution; Abcam Ltd, Cambridge, UK), caldesmon (mouse IgG₁, clone hHCD, ×500 dilution; SIGMA, St. Louis, MO), desmin (mouse IgG₁κ , clone D33, without dilution; DakoCytomation), cytokeratin (mouse IgG₁κ, clones AE1/AE3, ×20 dilution; DakoCytomation), and synaptophysin (rabbit IgG, ×200 dilution; Lab Vision Corporation, Fremont, CA).

Incubation with each primary antibody was performed at 4°C overnight and the sections were then processed according to the manufacturer’s instructions for the StreptABComplex/HRP kit (DakoCytomation), with 3,3’-diaminobenzidine/H₂O₂ as the chromogen. Prior to application of each primary antibody, all deparaffinized sections were subjected to antigen retrieval by autoclaving at 120°C for 10 minutes in 0.01M citrate buffer (pH 6.0), blocking of nonspecific endogenous peroxidase activity by treatment with 0.3% hydrogen peroxide in absolute methanol for 10 minutes, and then masking with 10% normal serum from a goat (for KIT, S-100, and synaptophysin) or rabbit (for α-SMA, vimentin, CD34, caldesmon, desmin, and cytokeratin) for 1 hour at room temperature. After immunoreactions, nuclei were lightly counterstained with hematoxylin.

Histological examination of the primary tumor revealed the tumor cells to have small- to medium-sized spindle or irregular-shaped single nuclei, with fibrillary cytoplasm and indistinct cell borders (Figure 1c and 1d). A distinct infil-trative proliferation pattern was evident within tunica muscularis and subserosa with a few densely proliferated foci showing a tendency to storiform pattern (Figure 1a, 1c and 1d). The tumor cells had directly infiltrated into the lamina propria mucosa, muscularis mucosa and tunica submucosa in both the forestomach and glandular stomach, as well as the serosal surface of the forestomach (Figure 1a). Metastatic tumor nodules were found in the liver, spleen, and femur bone marrow, with multiple tumor nodules, up to 1 cm in diameter, apparent in the liver (Figure 1b).

Immunohistochemical analysis of the primary tumor and metastatic deposits revealed weak diffused KIT reactivity (Figure 2a and b). Rare tumor cells were immunoreactive for S-100 (Figure 2c). In contrast, most tumor cells exhibited strong diffused staining with antibodies against α-SMA and vimentin (Figure 2d and e), while staining against CD34, caldesmon, desmin, cytokeratin, and synaptophysin was negative. Metastatic deposits exhibited similar immunoexpression patterns on these molecules. The animal also had C-cell adenoma and C-cell carcinoma of the thyroid, adrenal pheochromocytoma, and bilateral Leydig cell tumors in the testis. No gastrointestinal submucosal tumors that could be diagnosed as GISTs were observed in any of the other males treated with Madder color for up to 2 years.

KIT is expressed in the interstitial cell of Cajal, and therefore, GISTs have also been called gastrointestinal pacemaker cell tumors (Kindblom et al., 1998) or interstitial Cajal cell tumors (Sircar et al., 1999). CD34 immunoreactivity, which is present in 70–80% of human GISTs was not present in the present rat tumor. Demonstration of KIT immunoreactivity along with the location of the primary tumor in the submucosa supports an origin from the interstitial cell of Cajal. Only 5% of human GISTs are positive for desmin and 18% are positive for αSMA (Miettinen et al., 2005). Our present case was also negative for desmin as well as caldesmon, while strong immunoreactivity against αSMA was evident with this tumor, suggesting a limited potential for differentiation to muscle cells. Also, sparse S-100-immunoreactivity suggested a Schwann cell differentiation potential in part. These cellular features are compatible with the combined smooth muscle-neural subcategory of GISTs according to Rosai’s classification (1996).

We are not aware of any spontaneous GISTs developing in experimental rodents. However, in a rat gastric carcinogenesis model with reflux of duodenal contents, frequent KIT and CD34-immunoreactive GISTs developed along with adenocarcinomas (Mukaisho et al., 2005). In other animal species, spontaneous KIT-immunoreactive GISTs have been reported in dogs and horses (Hafner et al., 2001; Frost et al., 2003). Interestingly, these lesions frequently occur in the jejunum, cecum, and colon (LaRock and Ginn, 1997; Del Piero et al., 2001; Hafner et al., 2001; Frost et al., 2003). In contrast, a majority of human tumors occur in the stomach (Miettinen and Lasota, 2001). The present rat case showed a malignant phenotype with direct invasion of the serosa and metastasis to the liver, spleen, and bone marrow. In humans, malignant GISTs are also associated with liver metastasis and abdominal cavity involvement (Orosz et al., 2005). Similar pattern of spread was also reported in dogs (Frost et al., 2003). Considering the common immunoreactivity across species, KIT might be important cell marker for diagnosis of GISTs. In humans, a large number of tumors diagnosed previously as gastrointestinal smooth muscle tumors most likely represent GISTs (Rosai, 2003). Therefore, it should be kept in mind that many gastrointestinal submucosal tumors previously encountered in long-term rodent studies and considered to have a mesenchymal origin should have been diagnosed as GISTs. Application of immunohistochemical analysis of KIT and CD34 is necessary for diagnosis of gastrointestinal submucosal tumors.