Prognosis of patients with resection of stage IV gastric cancer.

We retrospectively analyzed 116 patients who underwent resection of stage IV gastric cancer to investigate the prognostic factors responsible for long-term survival. Statistically significant clinicopathological differences between patients surviving 3 years or more (n = 9) and those surviving less than 3 years (n = 107) were observed in regard to macroscopic type, curability of surgery, lymph node metastasis and blood vessel invasion. These four variables also had a statistically significant influence on survival by univariate analysis. Multivariate analysis revealed that no residual tumor (R0) surgery was a significantly beneficial factor for long-term survival. However, no patients with the diffusely infiltrative type of tumor survived for more than 2 years even if they underwent R0 surgery. Patients without the diffusely infiltrative type who underwent R0 surgery had better survival rates: a 27.3% and 13.6%, 3 and 5 year survival rate, respectively. In conclusion, R0 surgery contributes to long-term survival even in stage IV gastric cancer, and some patients with the not diffusely infiltrative type, in particular, may receive a survival benefit by R0 surgery.

[Evaluation of ileocolon interposition after total gastrectomy].

The most serious complaints after total gastrectomy are alkaline esophagitis and postalimentary syndrome. We have performed ileocolon interposition as a gastric substitution which uses the ileocecal valve to prevent bile reflux. The ileocolon segment consists of 7 cm of terminal ileum and 10 cm of coecum and ascending colon on a vascular pedicle. This isolated segment is rotated upwards between the esophagus and the duodenum. To evaluate the function of the ileocolon as a gastric substitute, Barium X rays, O-GTT and gastric emptying time (99Tc EDTA) were examined. Barium X-rays showed that the air at the coecal portion looks like a gastric fornix and there is no barium reflux into the esophagus even in the supine position. Likewise, there were no complaints from patients of symptoms of reflux esophagitis. In case of Rouxen-Y reconstructions, blood sugar level after O-GTT show oxyhyperglycemia. Those in ileocolon interposition rose rapidly, but almost to the same degree as in normal subjects. The appearance of the gastric emptying curve using Tc-DTPA in case of ileocolon interposition shows that there is good reservoir function.

Expression of MAGE genes in colorectal carcinomas.

The human genes MAGE-1 and MAGE-3 encode tumor rejection antigens recognized on melanoma cells by cytotoxic T lymphocytes (CTL). These antigens are potentially useful as targets for specific immunotherapy. Expression of MAGE genes in some malignant tumors has been reported, but MAGE gene expression in colorectal carcinomas has not been studied adequately. Therefore, we studied MAGE-1,2,3, and 4 a/4 b expression at the mRNA level, in 40 cases of surgery for colorectal carcinoma, using the reverse transcription polymerase chain reaction (RT-PCR). MAGE-1,2,3, and 4 a/4 b genes were expressed in 3 (7.5%), 6 (15.0%), 13 (32.5%), and 5 (12.5%), respectively, of these 40 cases. A total of 19 of the 40 samples (47.5%) expressed at least one of the MAGE genes. The relationships between clinicopathologic factors and MAGE gene expression were also examined. The frequency of lymph node metastasis was significantly higher in MAGE-3-positive than in MAGE-3-negative cases (p < 0.05). All these cases classified as Duke's D expressed the MAGE-3 gene. This rate of expression was significantly higher than that for all other the Duke's classifications together (p < 0.05). Our findings suggest that MAGE-specific immunotherapy against colorectal carcinomas may be feasible.

[A clinical study of recurrent breast cancer].

In a recent 14-year period from December 1978 to December 1993, 180 patients with breast cancer were treated at the hospital, and out of these 180 patients, 38 patients (21.1%) who had a recurrent breast cancer were clinically evaluated. The first recurrence occurred in the local skin in 13 patients, regional lymph node in 1, bone in 13, lung in 4, liver in 4 and in other areas in 3. Histologically, the incidence of scirrhous carcinoma recurrence was higher than that of papillotubular carcinomas. There was no recurrence in mucinous carcinomas. The frequency of recurrence became higher in accordance with TNM classification, namely, 6.7% in Stage I, 23.7% in Stage II, 37.5% in Stage IIIa, and 47.1% in Stage IIIb cancers. An increasing tendency in the recurrence rate was also noted with an increase of tumor diameter or lymph node metastasis. Otherwise, there was no difference in cumulative survival after recurrence between TNM classification. There were no correlations between the estrogen receptor and incidence of recurrence or 5-year survival rate. The disease free interval (DFI) was less than 2 years in about 60% of the recurrent cases. DFI of bone metastasis was longer than for the other sites of recurrence. The survival rate increased according to prolongation of DFI. After through evaluation of these results, one should pay close attention to follow-up after mastectomy.

[TNF-alpha gene therapy for nonimmunogenic tumor using immunogenic variant induced with mutagen (N-methyl-N'-nitro-N-nitrosoguanidine)].

We have already reported that a combination of TNF-alpha and immunogenic variant was effective for immunotherapy of nonimmunogenic tumor. Immunogenic variants (A2 and A7) were obtained from nonimmunogenic fibrosarcoma (parent cell: 1767-3) by mutagen treatment with N-methyl-N'-nitro-N-nitrosoguanidine (MNNG). In this study, we studied about the possibility of TNF-alpha gene immunotherapy for nonimmunogenic tumor. We established two types of TNF-alpha producing cell lines. One is TNF-alpha producing nonimmunogenic variant (1767 TR 2), and another is TNF-alpha producing immunogenic variant (A7 TR3). We compared them about the ability of producing specific immunity against parent tumor. TNF-alpha producing immunogenic variant (A7TR3) could induce strong specific immunity to parental cell compared to TNF-alpha producing nonimmunogenic variant (1767TR2). Immunogenicity of tumor is very important when we want to perform TNF-alpha gene immunotherapy against cancer.

Expression of HLA-class II antigen in gastric carcinomas. Its relationship to histopathological grade, lymphocyte infiltration and five-year survival rate.

To investigate whether the expression of HLA-DR antigen in gastric carcinomas is associated with the survival rate, we studied 70 cases of gastric carcinoma using a monoclonal antibody. Forty-seven cases (67%) stained positively, including 31 differentiated carcinomas, and 16 undifferentiated carcinomas. A close correlation was found between expression of HLA-DR antigen and differentiation of the tumor cells. Marked lymphocyte infiltration was seen in the HLA-DR antigen positive cancers. In advanced gastric cancers, the 5-year survival rate of patients with HLA-DR antigen positive cancers was significantly higher (67.5%) than that of patients with HLA-DR antigen negative cancers (40%). The data suggest that the expression of HLA-DR antigen in gastric carcinomas may reflect the degree of tumor cell differentiation, and influence the host immune response and prognosis.

[Prevention of postoperative recurrence after hepatic resection for metastatic colorectal cancer by adjuvant locoregional chemotherapy].

Postoperative adjuvant locoregional chemotherapy was performed on patients who underwent hepatic resection for metastatic colorectal cancer. To prevent recurrence in the residual liver after hepatic resection. 7 patients received intra-hepato-arterial infusion of MMC, 5-fluorouracil (5-FU), doxorubicin + Lipiodol emulsion. Sixteen cases did not receive such treatment. The 1-, 2- and 3-year cumulative disease free rates were 100, 80 and 40% in the patients with this adjuvant chemotherapy, against 61.4, 54.6 and 28.1% in those without this locoregional therapy, respectively. The 1-, 2- and 3-year cumulative survival rates were 100, 100 and 60% in the TAI-E group, and 73.7, 63.3 and 63.3% in the patients without TAI-E, respectively. Postoperative adjuvant locoregional chemotherapy is considered effective to prevent the recurrence of metastatic colorectal cancer in the residual liver after hepatectomy.

[Evaluation of the combination of chemotherapy with etoposide, adriamycin and cisplatin (EAP) in advanced or recurrent gastric cancer].

EAP therapy was performed on 30 cases of advanced or recurrent gastric cancer between September 1987 and July 1991. The clinical responses of 15 trial patients were evaluated. The overall response rate was 40.0% (CR, one case; PR 5 cases). The results were thus not as favorable as that reported by Preusser et al. On the contrary, with such a poor response rate, this treatment did not lead to a prolonged life span (mean survival time; 5.6 months, median survival time; 4 months). Side effects, such as myelosuppression, appetite loss, nausea, vomiting, liver dysfunction, renal dysfunction, and alopecia, were also observed. Myelosuppression was a dose-limiting factor. The rhG-CSF proved in 4 cases to be a clinically useful tool against the neutropenia induced by this treatment. It may be concluded that EAP should be given to the following selected patients: (1) those whose condition is not so far advanced: (2) those who have not received many other forms of treatment; and (3) those in excellent general physical condition.

Expression of sialyl-Tn antigen is correlated with survival time of patients with gastric carcinomas.

Expression of sialyl-Tn antigen (STN) was examined by an immunohistochemical method in 85 primary gastric carcinomas. The STN expression occurred in 53 (62.4%) cancers, and the positive staining was correlated with degree of gastric wall and lymph vessel invasion, lymph node metastasis, and stage of tumour. Five-year survival rates of patients with STN-positive cancers (47.2%) were significantly lower than those with STN-negative cancers (84.4%) (P < 0.01), and patients with STN-positive cancers at stage III and stage IV had a worse prognosis. In the cancers with serosal invasion, patients with STN-positive cancer disclosed a significantly poorer prognosis than those with STN-negative cancers (P < 0.01). Therefore, it is suggested that a careful follow-up study and intensive postoperative therapy are needed for patients with advanced gastric cancers with positive STN expression.

The induction of enhanced antitumor effect against a nonimmunogenic tumor by highly immunogenic variants obtained by mutagen treatment.

Nonimmunogenic 1767-3 fibrosarcoma was treated with the mutagen N-methyl-N'-nitro-N-nitrosoguanidine, and stable variant cell clones (M-clones) were obtained that were able to elicit an immunological rejection response in syngenic C3H mice. Mice immunized with some M-clones were protected against a challenge from the original nonimmunogenic fibrosarcoma. Furthermore, when spleen cells of immunized syngenic mice were restimulated in vitro with M-clones, cytotoxic T lymphocytes (CTL) were obtained that were able to lyse not only M-clones but also the original nonimmunogenic tumor. These in vivo and in vitro results demonstrate the immunogenicity of M-clones and the existence of a singular antigenic specificity between the original nonimmunogenic tumor and M-clones. For the purpose of application of this mutagen treatment to cancer therapy, we combined it with lymphokine-activated killer (LAK) adoptive immunotherapy (AIT). With interleukin 2 and in vitro stimulation with highly immunogenic variant clones, we tried to induce transfer cells that had not only nonspecific LAK cells but also CTL with specific immunity against the original nonimmunogenic tumor. Successful results were obtained in the LAK AIT models. These findings indicate that an immunotherapy of human cancers that are thought to be weakly or nonimmunogenic may be possible by the application of this approach to LAK AIT.

Antigenic variants isolated from a mutagen-treated guinea pig fibrosarcoma.

Antigenic variants were derived from a mutagen-treated, apparently nonimmunogenic fibrosarcoma of strain 2 guinea pigs. Fibrosarcoma line 107C3, originally induced by exposure of fetal cells to a chemical mutagen, was treated in vitro with the same mutagen. Cells that survived mutagen treatment were cloned, and the clones were tested for growth in soft agar, in conventional and immunosuppressed syngeneic guinea pigs, and in nude mice. At early passages after treatment, all clones tested in conventional syngeneic guinea pigs either failed to grow intradermally or grew temporarily and then regressed. At later passages after treatment, five of eight evaluable clones grew progressively; the characteristic of intradermal tumor growth followed by tumor regression (tum- or regressor) was a stable property of three of eight evaluable clones. The number of tumor cells required to produce progressively growing intradermal tumors in 50% of the animals (TD50) of the three tum- clones was at least 4 orders of magnitude greater than the TD50 of the parent fibrosarcoma. Tum- clones were not detected among 10 clones derived from the untreated parent tumor. Regressor clones formed colonies in soft agar and grew progressively in immunosuppressed syngeneic guinea pigs and nude mice. Regressor clones contained tumor transplantation antigens. Guinea pigs immunized with clones that grew and regressed rejected a challenge with the parent tumor when the dose of parent tumor cells was 1 to 3 times the TD50. Guinea pigs immunized by temporary growth of the parent tumor followed by excision of the local tumor and the regional lymph node did not reject a challenge with the parent tumor. These results confirm the results of experiments with murine tumors and extend the observations on tum- clones to the guinea pig. The results indicate that in guinea pigs it is possible by immunization with tumor cell variants derived from the mutagen-treated parent tumor to produce transplantation immunity to an apparently nonimmunogenic tumor.

Selection and rejection of retrovirus-expressing tumor cells from a heterogeneous murine leukemia virus-infected cell population.

We studied the basis of tumor recurrence at sites of rejection of retrovirus-infected guinea pig fibrosarcoma cells. Tumor recurrences, in contrast to the parent tumor, lacked retroviral antigens and did not release infectious virus. When reinjected into syngeneic animals, cell lines derived from tumor recurrences grew progressively. Tumor recurrences could be infected with the homologous retrovirus. Tumor rejection and recurrence were modulated by host immunity. In guinea pigs immunized to virus-infected cells, tumor recurrences occurred earlier and in a higher proportion of animals than in nonimmune guinea pigs. In some immunosuppressed guinea pigs, retrovirus-infected tumor cells grew progressively. Progressively growing tumors of immunosuppressed guinea pigs contained large amounts of infectious virus and expressed viral antigens. To identify the source of tumor recurrences, the parent virus-infected tumor was cloned. Clones were heterogeneous in virus expression; some clones released large quantities of infectious virus; others did not. Two clones formed tumors in syngeneic animals. Injection of a virus producer clone into virus-immune animals was not followed by tumor recurrence. The data suggest that the reappearance of tumors at sites of injection of retrovirus-infected fibrosarcoma cells represents immune selection and rejection of retrovirus-expressing cells. Cells with the potential to form tumor recurrences existed in the parent virus-infected tumor population.

A reliable method for reconstituting thymectomized, lethally irradiated guinea pigs with bone marrow cells.

We developed a reliable method for reconstituting thymectomized, lethally irradiated guinea pigs. Injection of 2.5-10 X 10(7) syngeneic bone marrow cells into adult thymectomized, lethally irradiated guinea pigs produced survival of 46-100% of treated animals. Gentamycin sulfate (5 mg/kg of body weight) for 10 days was required for optimal results. Acidified drinking water (pH 2.5) appeared to be required for optimal results. Thymectomized, lethally irradiated, bone marrow reconstituted ('B') guinea pigs had impaired ability to develop delayed cutaneous hypersensitivity to mycobacterial antigens and cutaneous basophil hypersensitivity to keyhole limpet hemocyanin; proliferative responses to phytohemagglutinin were impaired.

Plasma therapy of primary rat mammary carcinoma: antitumor activity of tumor-bearer plasma adsorbed against inactivated CNBr sepharose or protein A-sepharose.

We evaluated the ability of plasma adsorbed against Sepharose, inactivated CNBr Sepharose, or protein A-Sepharose to inhibit the growth of primary mammary carcinomas in Buffalo/N rats. Mammary adenocarcinomas were induced by a single intravenous injection of N-nitroso N-methylurea. When palpable mammary tumors were detected, rats were treated by infusion of plasma obtained from normal or tumor-bearing rats. Animals were killed 50 days after entry into the experiment; the index mammary tumor and any subsequently arising tumors were weighed and examined histologically. Unadsorbed plasma obtained from normal or tumor-bearing rats had no detectable antitumor effect. Normal plasma adsorbed against Sepharose or Sepharose derivatives had no significant antitumor effect. Weight (median) of index mammary tumors in groups of rats treated with normal plasma adsorbed against Sepharose alone, inactivated CNBr Sepharose alone, or protein A--Sepharose alone was not significantly different from weight of tumors of untreated, control animals. Tumor-bearer plasma adsorbed against either inactivated CNBr Sepharose or protein A-Sepharose had significant antitumor effect. Weight of index mammary tumors in groups of rats treated with tumor-bearer plasma adsorbed to inactivated CNBr Sepharose alone or against protein A-Sepharose alone was significantly less than weight of tumors in control rats. Tumor-bearer plasma adsorbed against Sepharose alone inhibited tumor growth in one experiment but not in a second experiment. Tests for endotoxin in CNBr Sepharose indicated the presence of approximately 1 ng endotoxin/g Sepharose. Administration of nanogram quantities of Salmonella enteritidis endotoxin in saline did not inhibit growth of primary mammary tumors. These experiments describe a rodent model that may be useful in the analysis of the basis of the inhibition of tumor growth that occurs following administration of adsorbed plasma.

Rejection of retrovirus-infected tumor cells in guinea pigs: effect on bystander tumor cells.

We studied the basis of rejection of retrovirus-infected tumor cells in guinea pigs by evaluating host response to injection of mixtures containing retrovirus-infected tumor cells and, antigenically and biologically distinct, uninfected tumor cells (line 10). After intradermal injection, line 10 grew progressively, metastasized to regional lymph nodes, and led to death of animals; line 107C3 4070A, a murine leukemia virus-infected fibrosarcoma cell line, grew for approximately 1 week and then regressed. Growth of the line 10 hepatoma was suppressed when the hepatoma cells were mixed with viable 107C3 4070A cells before injection into strain 2 guinea pigs. Viable virus-infected 107C3 cells were more effective than irradiated virus-infected cells in suppressing line 10 growth; mixture of line 10 with murine leukemia virus 4070A alone did not inhibit line 10 growth. Suppression of growth of line 10 cells by admixed murine leukemia virus 4070A-infected cells was less effective in animals with established viral immunity. Cyclophosphamide inhibited suppression of line 10 at sites of injection of virus-infected cells. Infection of line 10 with murine leukemia virus in vitro required cocultivation of line 10 cells with murine leukemia virus-infected fibrosarcoma cells; virus alone did not lead to acquisition of murine leukemia virus antigens by line 10 cells.

Eradication of microscopic hepatic metastases by active specific immunization.

Strain-2 guinea pigs, each with microscopic deposits of line 10 hepatocarcinoma in the liver, were treated by ID immunization with a mixture of irradiated tumor cells and an oil-in-water emulsion containing cell walls of Mycobacterium bovis strain Bacillus Calmette-Guérin (BCG CWE). Injection of line 10 hepatoma cells into the hepatic portal vein led to the development of tumor foci in the liver, metastasis in the hepatic lymph node, malignant ascites, and death. Active immunization using irradiated line 10 cells and BCG CWE was effective therapy when administered 1, 7, and 14 days after intraportal injection of line 10 cells. Effective immunization required both irradiated line 10 tumor cells and the BCG cell wall emulsion. Immunization with BCG CWE admixed with irradiated line 1 tumor cells, a hepatoma antigenically distinct from line 10, did not prevent outgrowth of line 10 deposits in the liver. Animals rendered free of disease could reject a challenge of line 10 tumor cells but not of line 1 tumor cells.

Tumor rejection mediated by an amphotropic murine leukemia virus.

We studied the effects of murine leukemia virus infection on the growth of tumors in inbred strain 2 guinea pigs. Fibrosarcomas, induced by treatment of guinea pig fetal cells with chemical carcinogens, were exposed in vitro to the amphotropic murine leukemia virus, 4070A. Tumor cells exposed to murine leukemia virus 4070A in vitro expressed virus antigens, released both reverse transcriptase and infectious virus into supernatant fluids, and grew and regressed after injection into syngeneic animals. In contrast, uninfected tumor cells did not express virus antigens and grew progressively. Rejection of virus-infected tumor cells appeared to be a host-mediated event, since murine leukemia virus 4070A infection had no detectable cytopathic effect on fibrosarcoma cells. Rejection of virus-infected tumor cells occurred in guinea pigs unable to suppress the growth of the line 10 hepatoma at sites of injection of mycobacterial cell walls and unable to develop immunity to the line 10 hepatoma. Guinea pigs immunized with virus-infected tumor cells were unable to reject a challenge of uninfected tumor cells. The results are interpreted to mean that infection of guinea pig tumors in vitro by a murine leukemia virus led to synthesis of viral antigens by the tumor cells which in turn led to recognition of the tumor as foreign with consequent tumor destruction; tumor eradication occurred without development of immunity to cryptic, intrinsic tumor antigens.