Intramedullary spinal cord tumors (IMSCTs) are relatively rare neoplasms, accounting for only 2% - 4% of all central nervous system (CNS) tumors and 15% of all spinal cord tumors. Of these, approximately 70% are lowmalignancy lesions, such as low-grade astrocytomas and ependymomas [1-3]. High-grade astrocytomas occur less frequently. The ratio of highto low-grade astrocytomas has been reported to be 1:3 [2]. Glioblastoma multiforme (GBM) of the spine is rare, accounting for only 1.5% of all spinal cord tumors and 1% to 5% of all GBM cases [4]. These highly malignant lesions occur mainly in the cervicothoracic segments, have a slight tendency to occur in the first decades of life, and have a short clinical history before diagnosis [5]. From the surgical point of view, the optimum treatment of these tumors is controversial. With recent advances in neuroimaging equipment, e.g., magnetic resonance imaging, microsurgical techniques, e.g., intraoperative ultrasound, laser therapy, and intraoperative electrophysiological monitoring, and total resection of most tumors is possible [6].

The development of a novel model of IMSCTs using glial cell lines would facilitate biological and histopathological studies, as well as lowering costs and increasing the utility of the model. In previous studies aimed at generating a model of IMSCTs, tumor lines of intracranial glial tumors, such as 9L gliosarcoma and F98 glioma, were commonly used [7].

In this study, we present for the first time a novel rat model of IMSCTs using glioblastoma cell line C6 and discuss the methodology and histopathological features.

Glioma cell lines and culture condition: The rat glioblastoma cell line C6 was obtained from ATCC (RockvilleUSA), and serial passages were made in modified Eagle medium of Dulbecco, which contained 15% heat-inactivated fetal calf serum, 0.2 mM glutamine, 50 mg/ml neomycin, and 100 mg/ml streptomycin. Culture flasks were kept in an electronic incubator under a humidified atmosphere with 5% CO₂ at 37˚C.

Animals and implantation procedure. The cells were harvested via 2 ml trypsin-EDTA solution C (Biological Industries, Israel) and centrifuged at 1 × 10³ rpm for 3 minutes after the addition of 1.5 cc of F12 medium and the fetal calf serum mixture. After removing the supernatant, the pellet was resuspended with 2 cc of medium. The cell suspension was concentrated; so that 5 microliters of any injection volume contained 5 × 10⁵ cells. It was then placed in a microcentrifuge tube and kept in a water-ice mixture environment during the whole implantation procedure, which always lasted less than 2 hours.

Fourteen male Wistar rats were randomized into two groups. The rats in group 1 (control group, n = 7) received a 5 μl intramedullary injection of serum physiologic (SF). The rats in group 2 (experimental group, n = 7) received a 5 μl intramedullary implantation of media containing 5 × 10⁵ C6 glioma cells. The rats were sacrificed for histopathological examination at 21 days.

The rats were anesthetized with an intraperitoneal (i.p.) injection (0.4 - 0.6 ml) of a stock solution containing ketamine hydrochloride (50 mg/ml) (Ketalar^®, Pfizer İla- çları Ltd. Sti., Istanbul), xylazine hydrochloride (10 mg/ ml, Rompun^®, Bayer), and 14.25% ethanol in normal saline. The animals were placed on a sterile field, and their backs were shaved and prepared with a betadine solution. The spinous process of thoracic vertebrae 5 (T5) was identified, and a 2-cm longitudinal incision was made over the dorsal midthoracic region. The underlying fascia and the paravertebral muscles were retracted laterally, the spinous process of T5 was removed with rongeurs, and the ligamentum flavum was removed, exposing the intervertebral space. The cell suspension was injected through the dorsal intervertebral space with a 26-gauge Hamilton syringe (Hamilton Company, Reno, NV). The needle was advanced until the dorsal aspect of the vertebral body was felt and then retracted slightly (1 - 2 mm). The incisions were sutured closely in layers after the injection. After recovery, the rats were returned to their cages where the room temperature was 28˚C.

After sacrifice, the spinal column of each animal was exposed, and a segment encompassing all macroscopically visible tumor was excised en bloc, placed in 4% formalin, and embedded in paraffin. The spinal cord was sectioned in 5-μm slices for hematoxylin-eosin staining.

In the histopathological examination of the control rats, there were no significant findings relating to tumors, and normal cytoarchitectures were observed in the spinal cord (Figures 1(A) and (B)).

The histopathological examination of the group 2 rats implanted with the C6 glioblastoma cells demonstrated intramedullary spinal cord tumors. Tumor tissue protruding from the skin was visible (Figure 2(A)). The tumors had highly cellular, well-defined lesions, with compression of the surrounding structures (Figure 2(B)). In the tumor tissue, cellularity changes were less frequently demonstrated, and tumor-containing myxoid areas were present (Figure 2(C)). The bone structure was found to be infiltrated by the tumor (Figure 2(D)). Within the tumors, glial tumor cells, ovoid and round nuclei, pink fibrillary cytoplasm, moderate pleomorphism, and high mitotic activity were observed (Figure 2(E)). In addition, the area of necrosis and the tumor cells surrounded palizadik were found (Figure 2(F)).

As mentioned above, astrocytomas and ependymomas represent the most common intramedullary neoplasms. It is known that the intracranial to spinal ratios for astrocytomas and ependymomas are 10:1 and 3:1 to 20:1, respectively [8]. Treatment of IMSCTs includes resection, radiotherapy, or a combination of these approaches [9- 12]. Although the mechanisms by which IMSCTs produce neurological dysfunction are not precisely known, the effects of pressure and edema on axonal conduction probably play a key role. The clinical presentation of IMSCTs is determined in part by the location of the tumor [13]. Midline back pain is the most common initial symptom of spinal cord tumors. Leg weakness and sensory changes are the other symptoms that bring patients to diagnostic evaluation [14]. The improvement in function after surgical resection of the tumor is finding a

good for prognosis [15-17].

There have been limited advances in the diagnosis and management of spinal cord tumors because of the rarity of the disease. The best treatment for IMSCTs has so far to be determined. Hence, a suitable animal model of IMSCTs needs to be developed. In contrast to the many existing intracranial animal tumor models [17-19], the literature contains few animal models of IMSCTs [20-22]. Caplan et al. [22] found that animals injected with 9L or F98 consistently develop hind-limb paresis in a reliable and reproducible manner. The progression of neurological deficits was similar to that seen in patients with IMSCTs. They suggested that this model mimics the behavior of IMSCTs in humans and may be used to examine the efficacy of new treatment options for both lowand high-grade intramedullary tumors. Ren et al. [23] described a murine model of IMSCG using 9L gliosarcoma cells, which are syngeneic to Fischer 344 rats. They reported on the functional progression, neuroimaging, and histopathological characterization of this tumor model. The rats implanted with the 9L cells regularly developed paraplegia in a reliable and reproducible manner. In the present study, we found that the animals injected with the C6 tumor had a median onset of hindlimb paresis. The histopathological examination of the rats implanted with the C6 glioblastoma cells revealed intramedullary spinal cord tumors. The tumors displayed highly cellular, well-defined lesions with compression of the surrounding structures. Within the tumors, polymorphic malformed nuclei with clearly mitotic figures and multinucleated cells were observed. Endothelial proliferation was evident with necrosis also.

In conclusion, our results indicate that a rat C6 IMSCT model was successfully established. This current working rat model may resemble the behavior of IMSCTs in humans. However, we believe that further studies of the existing IMSCT model are needed.