The Significance of the Ki-67 Labeling Index, the Expression of c-kit, p53, and bcl-2, and the Apoptotic Count on the Prognosis of Gastrointestinal Stromal Tumor

with the primary antibody for 30 minutes at room temperature. After rewashing, the sections were further incubated with peroxidase-labeled dextran conjugated anti-mouse and anti-rabbit immunoglobulin goat polyclonal antibody (DAKO) for 30 min at room temperature. After washing again , the sections were developed with chromogen 3,3’- diaminobenzidine tetrahydrochloride. Immunostaining was performed using the avidin-biotin complex method for Ki-67, p53, and bcl-2. Anti-human Ki-67 mouse monoclonal antibody (MM1 Novocastra, Newcastle, UK), anti-human p53 mouse monoclonal antibody (Do-7 Novocastra), or anti-human bcl-2 mouse monoclonal antibody (Novocastra) was used as the primary antibody. The deparaffinized sections were heated in an autoclave at 120 ℃ for 5 minutes in citric acid buffer (2 mmol/l citric acid and 9 mmol/l trisodium citrate dehydrate, pH6.0) and incubated with 0.03% H 2 O 2 / methanol for 30 minutes for blocking endogenous peroxidase activity. After washing in PBS, nonspecific binding was blocked by incubating the sections with normal animal serum for 20 minutes. The sections were incubated with the primary antibody overnight at 4 ℃ . After washing, the sections were further incubated with biotinylated second antibody for 30 minutes at room temperature. The primary antibody was detected using the avidin-biotin peroxidase complex (Vector Laboratories, Burlingame, CA) and 3-amino-9-ethylcarbazole as the chromogen. These thin sections were also counterstained with hematoxylin and mounted. For immunoreactivity of p53 and bcl-2 proteins, tissue sections with >10% immunopositive cells in 1000 tumor cells from three arbitrary microscopic fields were defined as being positive (+), and those with <10% immunoreactive tumor cells were defined as being negative (-).


Clinicopathological features
In six cases, the lesion was incidentally discovered during a routine screening examination. The other six cases were symptomatic and presented epigastric pain, abdominal discomfort, anemia, vomiting, an abnormal shadow on chest X-ray, or tumor palpable, respectively. The lesion size ranged from 18 to 200 mm in diameter (mean size 55.0±44.5 mm) . The GIST in the jejunum was palpable. The macroscopic aspects of GIST in the jejunum are a smooth surface, gray color, pseudocapsule, and exophitic growth. We found one case with liver metastasis, one with lung metastasis, one with bone metastasis, and no cases with lymph node metastasis at the time of surgery (Table 1).

Histology and malignancy
A Mitotic Index of > 5/50 HPF was seen in five cases and <5/50HPF in the other seven cases. A high cellularity was seen in four cases, a middle cellularity in four cases, and a low cellularity was seen in the other four cases. Concerning the degree of anaplasia, a high degree was seen in three cases, a middle degree in four cases, and a low degree was seen in the other five cases. The number of cases with ulcer formation, cases with necrosis, cases with bleeding in the tumor, and the number of cases with mucosal invasion by tumor cells was four, one, four, and one, respectively. The numbers in the malignant group, borderline group, and in the benign group were five, two, and five, respectively. Three of four cases in the metastasis/recurrence group were malignant group, but other one was benign group. All four cases in the metastasis/recurrence group had ulcer formation and high cellularity. One case with necrosis and mucosal invasion had lung metastasis. Two of four cases with bleeding in the tumor was in the metastasis/recurrence group. Two of three cases with high atypia was in the metastasis/recurrence group. In the malignant group, the tumor size was >5 cm in diameter in three cases, the cellularity in four cases was middle or high, three cases had ulcer formation, three cases had bleeding in the tumor, and one case had necrosis in the tumor with mucosal invasion by tumor cells. In the boderline group, the cellularity was middle or low, the degree of anaplasia was high or low, one case had bleeding in the tumor, and there was no ulcer formation, no necrosis in the tumor, and no mucosal invasion. In the benign group, the tumor size was 3 cm or less in diameter. The cellularity was low in two cases, middle in two cases, and high in one case. The degree of anaplasia was low in three cases, middle in two cases, and high in no case. One case with bone metastasis had ulcer formation . There was no necrosis, no bleeding in the tumor, and no mucosal invasion (Table 1).
U: upper body of the stomach, M: middle body of the stomach, L: lower body of the stomach, J: jejunum, *mucosal invasion, **upper part of the remnant stomach Table 1. Histological findings and malignancy The diameter of the tumor and the mitotic index have been reported as useful indices of biological malignancy in GISTs. We used Amin's classification, preparing evaluation criteria based on the pattern of mitosis and the diameter of the tumor. Mucosal invasion, bleeding, necrosis, and high degree of anaplasia were not recognized in the benign group. However, one case with bone metastasis in benign group had high cellularity and ulcer formation. Therefore, the diameter of the tumor and the mitotic index were not only useful indices of biological malignancy, but also cellularity, ulcer formation, necrosis, bleeding in the tumor, anaplasia, and mucosal invasion.

Treatment
Concerning the treatment, local resection was performed in eight cases, and in five of these cases, laparoscopic local resection was performed. In the other four cases, proximal gastrectomy with intrahepatic arterial infusion of adriamycin, resection of the remnant stomach with left lower lobectomy of the lung, local resection with radiation for bone metastasis and partial resection of the jejunum including the tumor were performed (Table 2).

Prognosis
Concerning survival, one patient with lung metastasis received resection of the remnant stomach with left lower lobectomy of the lung, and later died of intra-abdominal and right lung recurrence at 35 months later. One patient with multiple liver metastasis received proximal gastrectomy with intrahepatic arterial infusion of adriamycin, recurrence in the abdomen was recognized, removed and imatinib was administered from 41 months, but www.intechopen.com The Significance of the Ki-67 Labeling Index, the Expression of c-kit, p53, and bcl-2, and the Apoptotic Count on the Prognosis of Gastrointestinal Stromal Tumor 111 died of peritoneal metastasis at 58 months after the first surgery. The prognosis of the patient with bone metastasis received local resection with radiation was unknown. The patient with the large GIST in the jejunum received partial resection of the jejunum including the tumor, recurrence in the liver was recognized and imatinib was administered from 25 months, local recurrence was recognized in the small intestine and removed at 45 months, and was still alive at 75 months after the first surgery (

c-kit expression
All patients were positive for vimentin expression. Seven patients were positive for c-kit expression. Nine patients were positive for CD34 expression. In the seven c-kit-positive cases, six patients were also positive for CD34 expression. Two patients were positive for α-SMA expression. All patients were negative for both desmin and S-100 expression. All four cases in the metastasis/recurrence group were positive for c-kit expression ( Table 3). The positive rate of c-kit in the malignant group was 100% (five of five), in the borderline group, it was 50% (one of two), and in the benign group, it was 20% (one of five; Table 4)).The positive rate of c-kit in the malignant group was significantly higher than in the borderline and benign group (P < 0.05). Mean tumor size in the c-kit -positive group was 76.4 ± 62.2 mm and c-kit-negative group was 33.0 ± 21.0 mm ( Table 5).The mean tumor diameter in the c-kit-positive group was higher than that in the c-kit-negative group. Concerning any correlation between cellularity and c-kit expression, the positive rate of c-kit in the high group was 100% (four of four), in the middle group was 50% (two of four), and in the low group was 25% (one of four). Concerning any correlation between the degree of anaplasia and the c-kit expression, the positive rate of c-kit in the high group was 100% (three of three), in the middle group was 75% (three of four), and in the low group was 20% (one of five). The positive rate of c-kit in cases with ulcer formation was 100% (four of four), and in cases without ulcer formation it was 37.5% (three of eight). The positive rate of c-kit in cases with bleeding in the tumor was 100% (four of four), and in cases without bleeding in the tumor, it was 37.5% (three of eight). All malignant cases, cases with high cellularity, cases with high degree of anaplasia, cases with ulcer formation, and cases with bleeding in the tumor were positive for c-kit expression. Immunohistochemistry analyses have been reported to show a positivity reaction to CD34 in > 80% of cases, and to the c-kit in 95-100%. In our study, the positive rate to CD34 was 83.3%, and to the c-kit it was 66.7%. Therefore, the positive rate to the c-kit was lower than in other reports. However, all other reported cases lacked immunohistochemical evidence of smooth muscle or neural differentiation. Hirota reported that some GISTs lacked immunoreactivity to c-kit without immunohistochemical evidence of smooth muscle or neural differentiation. Many GISTs showed a positivity to CD34 recognized in endothelial cells and in interstitial cells of Cajal. The positive rate to α-SMA was 16.7% in our cases. Hirota reported that from 20 to 30% of GISTs were positive to α-SMA. Somatic mutation in the c-kit gene of GIST has been reported to be associated with aggressive progression and poor prognosis. Ernst reported that the 3-year survival rate of cases without mutation in ckit was >65%, but that of cases with mutation in c-kit it was < 30%. In our cases, all malignant cases were c-kit positive, and in the benign group only one case which had bone metastais was c-kit positive. The average diameter in the c-kit-positive group was higher than that in the c-kit-negative group. The expression of c-kit was considered to reflect the malignancy of the GIST. It has recently been clarified that the autoactivation of tyrosin kinase caused by upregulation of the c-kit gene was related essentially to proliferation in the GIST cells, and imatinib, which inhibits the autoactivation of tyrosine kinase, has been reported to be effective as a moleculetargeting treatment. Although adjuvant use of imatinib is currently under evaluation, it is expected to be a useful treatment for an unresectable GIST or for a patient with metastasis/recurrence.

The Ki-67 labelling index
Immunoreactivity to Ki-67 was seen in the nucleus of tumor cells (Fig. 1). The Ki-67 labeling index ranged between 4.7 to 49.8 (mean 25.4±15.9; Table 3). In the metastasis/recurrence group, the Ki-67 labeling index in patient with liver metastasis was 49.8, in patient with bone metastasis was 26.8, in patient with lung metastasis, it was 40.2, and in small intestinal GIST patient with recurrence in the liver, it was 34.6. The Ki-67 labeling index in the malignant group was 35.3±11.0, in the borderline group 37.5±9.40, and in the benign group was 10.6±9.26 (Table 4). The Ki-67 labeling index in the malignant group was significantly higher than that in the benign group (p < 0.01). The Ki-67 labeling index in the c-kit-positive cases was 35.3±10.2 and in c-kit-negative cases was 11.4±11.0. The Ki-67 labeling index in the c-kit positive cases was significantly higher than that in the c-kit-negative cases (p < 0.01). The Ki-67 labeling index in the c-kit-positive cases was 35.3±10.2 and in c-kit-negative cases was 11.4±11.0 ( Table 5). The Ki-67 labeling index in the c-kit positive cases was significantly higher than that in the c-kit-negative cases (p < 0.01). AgNOR staining, DNA ploidy pattern of the nucleus, bromodeoxyuridine labeling index, Ki-67 labeling index, proliferative cell nuclear antigen (PCNA), and p53 staining have each www.intechopen.com The Significance of the Ki-67 Labeling Index, the Expression of c-kit, p53, and bcl-2, and the Apoptotic Count on the Prognosis of Gastrointestinal Stromal Tumor 113 been reported to be a useful marker as an index of malignancy in GIST. Especially, the Ki-67 labeling index has recently been used as an excellent index of cell growth. The mitotic index reflects the M stage of mitosis only; however, because the Ki-67 labeling index can recognize most proliferating cells in stages G1, S, and G2, it is considered to be more appropriate as an objective index of the malignancy in GIST. Shimoda  , and the mean Ki-67 labeling index in the malignant group was higher than that in the benign group. Even in the benign group, one case with bone metastasis showed a high Ki-67 labeling index (26.8), and this case showed high cellularity and had ulcer formation. Such a case should be carefully followed postoperatively. The mean Ki-67 labeling index in the c-kit-positive cases was higher than that in the c-kit-negative cases. The Ki-67 labeling index was considered to be a useful marker of malignancy in GIST.

p53 expression
The immunoreactivity of p53 was also mainly identified in the nucleus (Fig. 2). Four patients were positive for p53 expression. Three cases of the metastasis/recurrence group were positive for p53 expression. The positive rate of p53 in the malignant group was 60.0% (three of five), in the borderline group was 50.0% (one of two), and in the benign group was 0%(zero of five; Table 4). The mean tumor size in the p53-positive group was 115.0±61.9 mm and in the p53-negative group was 26.3±11.9 mm. The mean tumor size in the p53-positive group was higher than that of p53-negative group. The mean Ki-67 labeling index in the p53-positive group was 38.9±8.26 and in the p53negative group was 18.6±14.6. The mean Ki-67 labeling index in the p53-positive group was significantly higher than that in the p53-negative group (p < 0.05). The p53-positive rate in the c-kit-positive group was 42.9% (three of seven), and in the ckit-negative group was 20.0% (one of five), with no significant difference between them (Table 5). Following DNA damage, p53 protein levels rise dramatically, and the entry into S is delayed until the genomic lesions are fully repaired. When the p53 function is lost, cells enter S without appropriate DNA repair, leading to fixation and propagation of genetic alterations. p53 overexpression promotes the transcription of p21, the product of which causes growth arrest through inhibition in Cdks, which are required for G1 to S transition. p21 is induced by DNA-damaging agents that trigger G1 arrest or apoptosis in cells with wild type p53 but not in tumor cells harboring a deletion or mutation in the p53 gene. Nikaido et al. reported that survival in gastric leiomyosarcoma of p53-positive cases was significantly shorter than that of p53-negative cases, that immunohistochemical p53 positivity was correlated with malignant behavior, and that p53 immunoreaction was a prognostic variable. In our cases, three cases of the metastasis/recurrence group were positive for p53 expression. The Ki-67 labeling index in all p53 -positive cases was >30%. There was no p53-positive case in the www.intechopen.com The Significance of the Ki-67 Labeling Index, the Expression of c-kit, p53, and bcl-2, and the Apoptotic Count on the Prognosis of Gastrointestinal Stromal Tumor 115 benign group. The mean tumor diameter of the p53-positive cases was higher than that of the p53-negative cases. The mean Ki-67 labeling index of the p53-positive cases was higher than that of the p53-negative cases. The expression to p53 was considered to be a useful marker as an index of malignancy and prognosis in GIST.

bcl-2 expression
The immunoreactivity of bcl-2 was mainly identified in the cytoplasm of tumor cells (Fig. 3). Seven patients were positive for bcl-2 expression (Table 3). Fig. 3. bcl-2 staining; Immunohistochemical staining with a monoclonal antibody against bcl-2. Positive staining was observed in the cytoplasm of the tumor cells (×400).
All four cases of the metastasis/recurrence group were positive for bcl-2 expression ( Table 3). The positive rate of bcl-2 in the malignant group was 60.0% (three of five), in the borderline group 50.0% (one of two) and in the benign group it was 60.0% (three of five; Table 4). The mean tumor size in the bcl-2-positive group was 77.1±64.4 mm, and in the bcl-2-negative group was 32.0±14.4 mm, with no significant difference between them. The mean Ki-67 labeling index in the bcl-2-positive group was 28.3±15.8 mm, and in the bcl-2-negative group was 21.3±17.0 mm, with no significant difference between them. The bcl-2-positive rate in the c-kit-positive group was 57.1% (four of seven), and in the ckit-negative group was 60.0% (three of five), with no significant difference between them ( Table 5).
The bcl-2-positive rate in the p53-positive group was 100% (four of four), and in the p53negative group was 37.5% (three of eight), with no significant difference between them.
Cunningham et al. reported that the patients whose tumor demonstrated staining for bcl-2 protein had a shorter survival compared with those whose tumor did not demonstrate bcl-2. Noguchi et al. reported that overexpression in bcl-2 may play an important role in increasing the malignant potential, and furthermore, that the Ki-67 labeling index and bcl-2 overexpression may be useful in predicting malignant potential. In our study, all four cases of the metastasis/recurrence group were positive for bcl-2 expression, and all p53-positive cases were positive for bcl-2 expression. However, no significant correlation was observed between the frequency of the bcl-2 overexpression and the malignancy in the GIST, tumor diameter, or Ki-67 labeling index,

TUNEL staining
Positive staining was recognized in the nuclei in the apoptotic tumor cells (Fig. 4). The apoptotic count ranged between 0.1 to 3.9 (mean 1.20±1.14; Table3). The apoptotic count in the malignant group was 1.44±1.65, in the borderline group was 0.50±0.57, and in the benign group was 0.80±0.71, with no significant difference among them. The apoptotic count in the c-kit-positive group was 1.44±1.32, and in the c-kit-negative group was 0.88±0.88, with no significant difference between them ( Table 5). The mean apoptotic count in the p53-positive group was 2.33±1.20, and in the p53-negative group was 0.65±0.62. The mean apoptotic count in the p53-positive group was higher than that in the p53-negative group. The mean apoptotic count in the bcl-2-positive group was 1.80±1.16, and in the bcl-2-negative group it was 0.38±0.36, with a significant difference (p < 0.05). All apoptotic counts of malignant group or metastasis/recurrence group were 1.0 or > 1.0. www.intechopen.com The Ki-67 Index in the malignant group was significantly higher than that in the benign group (p < 0.01). **: The positive rate of c-kit in the malignant group was significantly higher than in the borderline and benign group (P < 0.05). reported that the apoptotic cell counts of leiomyosarcoma were significantly higher than those of leiomyoma, but this seems not to be a practical index for discrimination because apoptotic cell death is a rare event in gastrointestinal myogenic tumors. In our study, there was no significant difference in apoptotic count among the benign group, borderline group, and the malignant group. The mean apoptotic count in the bcl-2-positive group was higher than that in the bcl-2-negative group, and the mean apoptotic count in the p53-positive group was higher than that in the p53-negative group. p53 induces upregulation in bax protein which accelerates apoptotic death. bcl-2 protein is able to repress a number of apoptotic death programs. bax homodimerizes and forms heterodimers with bcl-2 in vivo. The ratio of bcl-2 to bax determines survival or death following an apoptotic stimulus. Therefore, if overexpressed bax countered the death repressor activity of bcl-2, and apoptosis was considered to be induced in cases that were positive for bcl-2 expression.

Conclusion
Here, we report the Ki-67 labeling index, the expression of c-kit, p53, bcl-2, and apoptosis in eleven gastrointestinal stromal tumors (GISTs). The positive rate of c-kit in the malignant group was higher than in the borderline and benign group . The Ki-67 labeling index in the malignant GIST group was higher than that in the benign group. The Ki-67 labeling index in the c-kit-positive group was higher than that in the c-kit-negative group. The Ki-67 labeling index in the p53-positive cases was higher than that in the p53-negative cases. The bcl-2 expression was not correlated with potential malignancy. All metastasis/recurrence cases were bcl-2 positive. The apoptotic count in the bcl-2-positive cases was higher than that in the bcl-2-negative cases. All apoptotic counts of malignant group or metastasis/recurrence group were 1.0 or > 1.0. The high Ki-67 labeling index, the c-kit positive, the p53 over expression, the bcl-2 over expression and high apoptotic count were useful in predicting the potential malignancy of GIST. Almost 30 years have gone by since the postulation that GISTs derive from mesenchymal stem elements, and only 15 years have gone by since the definitive detection of origin of GISTs. Research in the last decade was more focused upon the justification of imatinib mezylate therapy in GISTs and clarification why a secondary resistance that occurred during the kinase inhibitors therapy. The era of therapy for GISTs, targeting the primary activating mutations in the KIT proto-oncogene; is being proclaimed as bringing the message of special importance to the pathologist role in multidisciplinary team that are responsible for treating patients with locally advanced or metastatic GIST. This is the first conclusive message forthcoming from this book. On the other hand, the book provides summarised and case-based knowledge on current management of gastrointestinal and extragastrointestinal stromal tumours. We hope that this book may be considered as a worthwhile timely addition to clinical science dissemination, medical education, further basic and clinical research.