Prognostic Significance of Adoptive Immunotherapy with Tumor-associated Lymphocytes in Patients with Advanced Gastric Cancer

Patients and Clinical Evaluation.

Patients were eligible if they had histologically confirmed stage IV gastric cancer (staged according to the UICC classification), adequate bone marrow and renal function, and were <80 years old. One exclusion criterion was the presence of liver metastasis. Eligible patients were randomly assigned to the control group (chemotherapy alone) or the AIT group (AIT plus chemotherapy), on receipt of stage IV confirmation, no later than 10 days after surgery. All patients were monitored clinically using imaging analysis such as ultrasonography and computed tomography and classified according to conventional criteria: (a) complete response was defined as the disappearance of all measurable tumor for at least 1 month; (b) partial response was defined as a ≧50% decrease of all measurable tumor for at least 1 month; (c) MR was defined as a reduction of 25–49%; (d) NC was defined as a reduction of <25%; and (e) PD was defined as an increase in the tumor burden. Cytological examination of ascites and pleural effusions were performed periodically, and the cells were diagnosed as class I to class V according to the Papanicolaou classification. The present study was approved by the ethical committee of Yamanashi Medical University, and written informed consent was obtained from all individuals.

Treatment Design.

All patients in both groups underwent surgery. Then, from the 5^th week after surgery for 5 months, all patients received chemotherapy with cis-platinum (10 mg/day) plus 5-fluorouracil (500 mg/day) by venous infusion once a week and oral doxifluridine (600 mg/day) daily. In the AIT group (chemotherapy plus AIT), TALs were cultured for 6 weeks, and the cultured TALs were administered by i.p. bolus injection to autologous donors with malignant ascites or by i.v. bolus injection to autologous donors with malignant pleural effusion or lymph node metastasis.

All adverse events were evaluated according to the WHO toxicity scale, and treatment was stopped if toxic effects of grade 3 or 4 developed.

Preparation of Lymphocytes and Tumor Cells.

A single-cell suspension was prepared from malignant ascites or pleural effusions by 100% Ficoll (Pharmacia, Uppsala, Sweden) separation. The suspension was prepared from lymph node metastases obtained at surgery by mechanical mincing. Tumor- and lymphocyte-enriched populations were prepared by centrifugation of the single-cell suspensions through a discontinuous 75%/100% Ficoll gradient (5 , 11 , 12) .

Generation of TAL Lines.

To generate tumor cell lines, tumor-enriched populations from ascites, pleural effusions, or lymph node metastases in the patients were incubated for 60 min at 37°C to remove macrophages and fibroblasts by adherence to the plastic surface. Nonadherent cells were then cultured in RPMI 1640 (Life Technologies, Inc.) with 10–20% FCS, 50 units/ml penicillin, and 2 mm l-glutamine (Sigma, St. Louis, MO). Freshly separated or early-passage cultured tumor cells were frozen in aliquots or used fresh when available as feeders or stimulators in a TNF-α release assay.

TALs were generated by culturing the lymphocyte-enriched population using repetitive stimulation with an autologous tumor. The TALs (1–2 × 10⁶ cells/ml) were kept in AIM-V medium (Life Technologies, Inc.) supplemented with 50 IU/ml recombinant IL-2 (Shionogi) by stimulation with autologous tumor treated with mitomycin C (50 μg/ml; Kyowa Hakkou, Japan) every 2 weeks at a ratio of 10–20:1 (lymphocyte:tumor), as described previously (5 , 11 , 12) .

Characterization of Cultured TALs.

The phenotypes of cultured TALs were analyzed by flow cytometry using CD3, CD4, CD8, and CD16 monoclonal antibodies (Becton Dickinson, Mountain View, CA) after 6 weeks of culture. To study the reactivity of cultured TALs against autologous tumor cells or allogeneic gastric cancers, a TNF-α release assay was performed after 6 weeks of culture. Briefly, mixtures of 1 × 10⁵ effector cells (TALs) and 1 × 10⁵ stimulator cells (tumor cells) were cultured in AIM-V medium with recombinant IL-2 (50 IU/ml) at 37°C for 12 h. Autologous tumor cells and allogeneic gastric cancer cell lines (MKN28 and MKN45) were used as stimulators. The supernatants of the cultured mixtures were collected, and the TNF-α content was determined using an ELISA kit (Immunotech, Marseille, France) according to the manufacturer’s recommendations. When the amount of TNF-α for the autologous tumor cells was more than 2 times that for the allogeneic tumor cells, the reactivity was evaluated as preferential for autologous tumor cells.

Statistics.

Differences between groups were compared using the Mann-Whitney test (nonparametric statistical analysis) or χ² test and considered statistically significant when Ps were <0.05. The cumulative survival curve was estimated by the Kaplan-Meier method, and the differences in this distribution were compared with the log-rank test. Multivariate analysis of prognostic factors for overall survival was performed using Cox’s proportional hazards model. All analysis was performed with intention to treat. All time estimates were made with the date of surgery as the baseline.

Patient Groups.

Forty-four patients were randomly assigned to the control group (chemotherapy alone) or the AIT group (AIT plus chemotherapy) after surgery. There were no differences in the clinicopathological findings between the two groups, as shown in Table 1⇓ . Furthermore, there were no differences with regard to the extent of surgery performed or the extent of residual tumor (Table 2)⇓ .

Three patients in the AIT group were not able to complete the treatment protocol because TALs could not be maintained in culture for two of them (patients 6 and 11) and one patient refused any treatments (patient 13), as shown in Table 3⇓ . These three patients were included in the analysis of the AIT group because an intent to treat analysis was performed. Thus, 22 patients in the control group and 22 patients in the AIT group were evaluated in this study.

Clinical Outcome after AIT.

When the cultured autologous TALs were transferred to the patients, no complications, with the exception of low-grade fever, were observed.

Of the 22 patients in the AIT group, 2 patients showed a MR with diminished ascites or pleural effusion and negativity of cytology in the fluid, 5 showed NC, and 15 showed PD (Table 3)⇓ . In the control group (n = 22), 1 patient showed a MR, 4 revealed NC, and 17 revealed PD. The cause of death in all patients in the both groups was disease progression of gastric cancer.

The number of transferred TALs reached 9 × 10⁸ to 5 × 10⁹, and most of the transferred T cells were CD8+-dominant TAL lines (Table 3)⇓ . Preferential reactivity to the autologous tumor cells was observed in 3 (patients 1, 5, and 14) of 12 TAL lines tested. When the amount of TNF-α for autologous tumor cells was more than 2 times that for allogeneic tumor cells, the reactivity was evaluated as preferential for autologous tumor cells. There was no clear correlation between clinical response and preferential reactivity to autologous tumor cells (Table 3)⇓ . However, the amount of TNF-α for autologous tumor cells in patients with a MR or NC was significantly higher than that in patients with PD in the AIT group (P < 0.05; Table 3⇓ ).

In addition, the survival time in the AIT group did not correlate with the number of transferred cells, the amount of TNF-α released, or clinical response (Table 3)⇓ .

Survival Effect of AIT.

Cumulative survival rates were compared in two groups, as shown in Fig. 1⇓ . The 50% survival rates were 11.5 and 8.3 months in the AIT group and control group, respectively. The overall survival of the patients in the AIT group was significantly better than that of the patients in the control group, as analyzed by the log-rank test (P < 0.05).

• Download figure
• Open in new tab
• Download powerpoint

Fig. 1.

Cumulative survival rate after AIT. The cumulative survival curve was estimated by the Kaplan-Meier method, and the differences in the distributions were compared by the log-rank test.

To assess whether AIT represented a prognostic parameter, we used Cox’s proportional hazards model. The covariate parameters included age, gender, tumor size, depth of tumor invasion, peritoneal dissemination, lymph node metastasis, gastrectomy, and AIT. Multivariate analysis revealed that tumor size and AIT were independent prognostic factors (Table 4)⇓ , indicating that AIT influenced patient survival in a positive manner.