Histological and Immunohistochemical Studies on Spontaneous Rat Astrocytomas and Malignant Reticulosis

Abstract

Among spontaneous neoplasms of the rat central nervous system, the discrimination between astrocytoma and malignant reticulosis (MR) is sometimes difficult because of their similar cell morphology and infiltration patterns. In the present study, we carried out histological and immunohistochemical analyses on a total of sixty-four cases in Sprague-Dawley and F344 rats. These cases were diagnosed as benign/malignant astrocytoma containing no neoplastic oligodendroglial elements or MR according to the diagnostic criteria of the World Health Organization International Classification of Rodent Tumors (Mohr et al. 1994). Astrocytomas were divided into three types and MR into two types based on the number of lesions, cellularity and infiltration patterns, and so on. Although the neoplastic cells from all types showed various immunoreactivities for RM-4 (anti-rat macrophages and dendritic cells), ED-1, and/or vimentin, there were no distinctive differences among these types, and most cells that were positive for RM-4 were also positive for ED-1. None of the tumor types showed any reactivity for GFAP or S-100 protein. From the results of morphological and immunohistochemical examinations, it was indicated that there are no distinctive differences between spontaneous astrocytomas and MR in rats, and they are probably derived from the same cell lineage, that is, microglia, macrophage, or radial glia.

Introduction

In neoplasms of the rat central nervous system, spontaneous astrocytoma occurs most commonly, whereas malignant reticulosis rarely occurs (Zwicker et al. 1992). The diagnostic criterion of rat astrocytoma was first proposed by Garner, Innes, and Nelson in 1967. On the other hand, that of malignant reticulosis has been used from the proposal of Garman in Monographs on Pathology of Laboratory Animals sponsored by the International Life Sciences Institute in 1988. Similar tumors have previously been reported as neoplastic reticulosis, reticulo-sarcoma, polymorphic cell sarcoma, and so on (Gopinath 1986; Krinke et al. 1985; Mawdesley-Thomas and Newman 1974; Newman and Mawdesley-Thomas 1974).

In rats, although astrocytomas are sometimes found with oligodendrogliomas, many spontaneous astrocytomas are composed of relatively uniform neoplastic cells showing no oligodendroglial differentiation. The cellular morphology of astrocytoma containing no neoplastic oligodendroglia and malignant reticulosis is quite similar, and the only differentiation between these two types of tumors is the main site of cellular proliferation. Namely, it is generally accepted that neoplastic astrocytes proliferate mainly in the parenchyma, whereas neoplastic cells of malignant reticulosis infiltrate into the leptomeningeal and perivascular areas. However, neoplastic astrocytes frequently infiltrate into perivascular spaces and leptomeninges, and neoplastic cells of malignant reticulosis also infiltrate into parenchyma. Therefore, in some cases, it is hard to accurately diagnose these two types of tumors. Malignant astrocytoma or malignant reticulosis is reported to be induced by exposure to acrylonitrile or acrylamide (Bigner et al. 1986; Johannsen and Levinskas 2002a; Johannsen and Levinskas 2002b; Rice 2005). Accurate diagnosis of these tumors is important in the oncogenicity assessment of chemicals. In addition, the origin of malignant reticulosis is suggested to be lymphocyte, microglia, or histiocyte, but it has yet to be proved (Garman 1988; Krinke et al. 2000; Mohr et al. 1994; Solleveld and Boorman 1990). In rat astrocytoma, neoplastic astrocytes show negative reactivity for glial fibrillary acidic protein (GFAP) (Krinke et al. 2000; Solleveld and Boorman 1990), although neoplastic astrocytes in other species are positive for GFAP (Lopes and VandenBerg 2000; Okazaki and Scheithauer 1988; Summers et al. 1994). Thus, it is difficult to reach a definitive diagnosis of rat astrocytoma.

In the present study, morphological and immunohistochemical examinations were carried out on naturally occurring astrocytoma containing no neoplastic oligodendroglial elements and malignant reticulosis in rats. We propose that these two types of tumors are derived from the same lineage, namely, microglia, macrophage, or radial glia.

Materials and Methods

The materials available were collected from control and test groups without treatment-related effects on the central nervous system in one chronic and fifteen carcinogenicity studies using F344/DuCrlCrlj rats (Charles River Laboratories Japan, Inc., Atsugi, Kanagawa, Japan) and Crl:CD(SD) rats (Charles River Laboratories Japan). The studies were conducted at Bozo Research Center Inc. with permission from the Animal Care and Use Committee of Bozo Research Center Inc.

All animals were housed individually in wire mesh cages in an animal room under controlled conditions (23°C ± 3°C room temperature, 50% ± 20%relative humidity, and a twelve-hour light/dark cycle), fed a commercial diet (CRF-1, Oriental Yeast Co., Ltd., Tokyo), and checked daily. A complete necropsy was performed, and the brain and spinal cord were fixed with 10% phosphate buffered formalin, trimmed, and embedded in paraffin.

All central nervous tissue sections with any kind of neoplasm were re-examined. Then, benign/malignant astrocytoma, benign/malignant oligodendroglioma, benign/malignant mixed glioma, malignant reticulosis, medulloblastoma, or similar findings were diagnosed according to the diagnostic criteria of the World Health Organization (WHO) International Classification of Rodent Tumors (IARC No.122). From these tumors, a total of sixty-four neoplasms, which were astrocytomas containing no neoplastic oligodendroglia or malignant reticulosis, were selected for this study. Namely, all these sections of the brain and spinal cord were stained with hematoxylin and eosin, and Watanabe’s silver impregnation method for reticulin. In addition, immunohistochemical staining for ED-1, RM-4, GFAP, S-100 protein, vimentin, and nestin was done on the sections by the labeled polymer method using EnVision kits (DAKO Japan, Kyoto). The first antibodies used were as follows: (1) anti-rat CD68 mouse monoclonal antibody (clone: ED-1, 1:100, SEROTEC. Co., Ltd. Oxford, UK); (2) anti-rat macrophages/dendritic cells mouse monoclonal antibody (clone: RM-4, 1:25, Trans Genic, Inc. Kumamoto, Japan); (3) anti-GFAP rabbit polyclonal antibody (1:500, DAKO Japan); (4) anti-S-100 protein rabbit polyclonal antibody (1:500, DAKO Japan); (5) anti-vimentin mouse monoclonal antibody (clone: V9, 1:100, DAKO Japan); and (6) anti-nestin mouse monoclonal antibody (clone: Rat-401, 1:100, Santa Cruz Biotech, Inc., Santa Cruz, CA, U.S.A.). Positive control sections for GFAP and S-100 protein were obtained from the rat brain, and those for rat RM-4 and ED-1 from the rat spleen and lymph nodes.

For the purposes of this evaluation, astrocytomas were divided into three types (A, B, and C) based on the number of lesions, cellularity, or existence of necrosis with peripheral palisading, and MR were divided into two types (D and E) based on infiltration patterns. The number of cases in each type is shown in Table 1. The diagnostic criteria for each type are as follows. Type A (astrocytoma) was composed of one lesion with moderate cellularity in the parenchyma. Type B (astrocytoma) included not only a single lesion with dense cellularity, but also multiple lesions extending into the parenchyma and infiltrating into the surrounding leptomeninges or perivascular space. Type C (astrocytoma) showed characteristics of type B and had necrosis with peripheral palisading. Type D (malignant reticulosis) showed obvious parenchymal involvement, as well as more diffuse infiltration into the leptomeninges and perivascular spaces. Type E (malignant reticulosis) showed diffuse infiltration of neoplastic cells into the leptomeninges and perivascular spaces without obvious parenchymal involvement.

Results

Histological Characteristics of the Tumors

Histological features of each type of astrocytoma and malignant reticulosis are summarized in Table 2. The result of reticulin staining is shown in Table 3.

Histologically, there were no distinct differences in cell morphology between each type of astrocytoma and malignant reticulosis. Neoplastic cells were generally round, oval, polygonal, or slightly elongated with indistinct cell borders, and their nuclei were round or oval (Figures 1A–5A). In addition, the border between the neoplasm and the normal tissue was generally not clear (Figures 1A, 2A, and 4A). In types B, C, and D, cellular pleomorphism was more conspicuous compared with other types. In type C, mitotic figures were more frequently observed, and some tumors of this type were clearly separated from surrounding normal tissues. The morphological feature of the cells palisading around the necrosis was similar to that of the remaining cells in the neoplasm (Figure 3A). Reactive astrocytes, satellitosis of neurons by neoplastic cells (Figure 6A), and invasion into perivascular spaces and leptomeninges were observed in all types. Eosinophilic granular cells (Figure 6B), which had several or many eosinophilic granules in their cytoplasm, were seen in types B, C, and D, but not in the types A or E. Reticulins were produced around the neoplastic cells in types B, C, D, and E, but not in type A, and perivascular reticulins were observed in all types (Figures 1G–5G). Among all types of astrocytoma and malignant reticulosis, the only essential histological difference was the absence of a significant parenchymal proliferation of neoplastic cells in type E.

Immunohistochemical Characteristics of the Tumors

Immunohistochemical features of each type are shown in Table 4. Although neoplastic cells from all types of astrocytoma and malignant reticulosis showed various immunoreactivities for RM-4 (marker of macrophages and dendritic cells) (Figures 1C–5C) and vimentin (Figures 1F–5F), there were no distinctive differences in such immunoreactivities among these tumors. Astrocytoma generally showed positive reactivity for ED-1 (Figures 1B–3B), but some cases in types A and B were negative for ED-1. As to malignant reticulosis, two cases in type D were positive, whereas two cases in type E were negative for ED-1 (Figures 4B and 5B). Most of the RM-4 positive cells were simultaneously positive for ED-1. The reactivity for vimentin was more intense than that for RM-4 and ED-1, and a few neoplastic cells of types B and C showed positive reaction for nestin (Figure 6C). There was no reactivity for GFAP and S-100 protein observed in any types (Figures 1D–5D and 1E–5E). However, non-neoplastic and reactive astrocytes were positive for vimentin, GFAP, and S-100 protein (Figures 1D–3D, 1E–3E, 1F and 2F). Microglias and macrophages were positive for vimentin, RM-4, and ED-1. In general, the results of immunohistochemical analyses on malignant reticulosis were similar to those on astrocytoma.

Discussion

In the present study, to differentiate astrocytomas containing no neoplastic oligodendroglia from malignant reticulosis and to discuss their cell origin, a total of sixty-four cases obtained from F344 and SD rats were examined histologically and immunohistochemically. Histologically, the morphological characteristics of neoplastic cells were almost the same between astrocytomas and malignant reticulosis. Eosinophilic granular cells and perineuronal satellitosis which are characterized in rat astrocytoma (Krinke et al. 2000; Pruimboom-Brees et al. 2004) were observed in parenchymal lesions of type D of malignant reticulosis. Type E malignant reticulosis without obvious parenchymal involvement may have a parenchymal lesion similar to that in the case of type D in unexamined tissues. Consequently, it was difficult to divide these tumors morphologically.

There were no differences in the results of immunohistochemical staining between astrocytomas and malignant reticulosis. They both showed positive reactivity for ED-1, RM-4 (marker of macrophages) (Iyonaga et al. 1997), and vimentin, but not for GFAP or S-100 protein. It is consistent with the historical data in the literature that astrocytomas are uniformly GFAP negative (Krinke et al. 2000). In the present study, the two cases of type E malignant reticulosis were negative for ED-1, differing from many astrocytomas and another type of malignant reticulosis (type D). However, these two cases of malignant reticulosis (type E) were positive for RM-4, and the reactivity for RM-4 is generally said to correspond to that for ED-1. In addition, in WHO International Classification of Rodent Tumors (IARC No. 122), malignant reticulosis is described to be positive for ED-3, which is also a marker of macrophages. The reason the present two cases of malignant reticulosis (type E) were negative for ED-1 is unclear. However, in the present study, some of the astrocytomas were also negative for ED-1, and we were able to examine only two cases of type E malignant reticulosis. Therefore, it is difficult to consider negative reactivity for ED-1 to be specific to this type of malignant reticulosis. If we have the opportunity to examine more cases of type E malignant reticulosis in the future, we may be able to identify an ED-1–positive case.

In nonrodent astrocytomas, several cytological patterns such as fibrillary, protoplasmic, or gemistocytic astrocytoma are recognized (Lopes and VandenBerg 2000; Okazaki and Scheithauer 1988; Summers et al. 1994). However, the astrocytomas examined in this study could not be differentiated further based on the cytological pictures. Pseudopalisading around necrotic foci that characterizes glioblastoma in nonrodents (Lopes and VandenBerg 2000; Okazaki and Scheithauer 1988; Summers et al. 1994) was also observed in rat astrocytoma in this study, but glomeruloid vascular proliferation, which is considered to be a feature of malignancy in nonrodent brain tumors (Lopes and VandenBerg 2000; Okazaki and Scheithauer 1988; Summers et al. 1994), was not observed in any malignant astrocytomas in this study. In addition, perineuronal satellitosis, which is a common finding in rat astrocytomas, is generally described in oligodendrogliomas in humans (Okazaki and Scheithauer 1988). Moreover, differing from that in other species, rat astrocytomas in this study were immunonegative for GFAP. Thus, the nature of rat astrocytomas is different from those in other species.

It is well known that ethylnitrosourea (ENU) induces glial tumors, which are composed of mixed glioma, oligodendroglioma, and astrocytoma. Ethylnitrosourea-induced astrocytoma is accepted to be derived from astroglia. However, ENU-induced astrocytomas (Zook et al. 2000) seem to be morphologically different from the astrocytomas examined in this study, because the former contain oligodendroglial cells. On the other hand, acrylonitrile-induced brain tumors (Bigner et al. 1986) are morphologically similar to the astrocytomas examined in this study. To our knowledge, the precise origin of an acrylonitrile-induced brain tumor is unknown.

The immunohistochemical staining pattern of the tumors diagnosed as astrocytomas or malignant reticulosis is most consistent with the tumors which are derived from macrophage or microglia lineage but not from a neuroepithelial lineage. However, Wu et al. (2005) also reported that ED-1–positive macrophage/microglia were also stained by radial glial marker in rats, and this result suggested that the origin of the tumors diagnosed as astrocytoma in this study could be of radial glia lineage. Radial glia is generally recognized to be neuronal progenitor cells, which differentiate into astrocytes, oligodendrocytes, and neurons (Barry and McDermott 2005), and it is reported to show positive reaction for nestin (Shibuya et al. 2003). In the present study, although neoplastic cells showed positive reaction for nestin, it was limited to only a few neoplastic cells of a few neoplasms, and it is not definitive whether the neoplastic cell is a radial glia. Further studies are needed to clearly define the exact histogenesis of these tumors.

In conclusion, our data indicated that there were no distinctive histological and immunohistochemical differences between spontaneous astrocytoma containing no neoplastic oligodendrogia and malignant reticulosis in rats. These two tumors most likely originate from the same cell lineage, namely, microglia, macrophage, or radial glia.