The influence of autologous lymphokine-activated killer cell infusions on the toxicity and antitumor effect of repetitive cycles of interleukin-2

Lymphoma immunotherapy: vaccines, adoptive cell transfer and immunotransplant

Therapy for non-Hodgkin lymphoma has benefited greatly from basic science and clinical research such that chemotherapy and monoclonal antibody therapy have changed some lymphoma subtypes from uniformly lethal to curable, but the majority of lymphoma patients remain incurable. Novel therapies with less toxicity and more specific targeting of tumor cells are needed and immunotherapy is among the most promising of these. Recently completed randomized trials of idiotype vaccines and earlier-phase trials of other vaccine types have shown the ability to induce antitumor T cells and some clinical responses. More recently, trials of adoptive transfer of antitumor T cells have demonstrated techniques to increase the persistence and antitumor effect of these cells. Herein, we discuss lymphoma immunotherapy clinical trial results and what lessons can be taken to improve their effect, including the combination of vaccination and adoptive transfer in an approach we have dubbed 'immunotransplant'.

Antiviral NK cell responses in HIV infection: II. viral strategies for evasion and lessons for immunotherapy and vaccination

As is the case in other viral infections, humans respond to HIV infection by activating their NK cells. However, the virus uses several strategies to neutralize and evade the host's NK cell responses. Consequently, it is not surprising that NK cell functions become compromised in HIV-infected individuals in early stages of the infection. The compromised NK cell functions also adversely affect several aspects of the host's antiviral adaptive immune responses. Researchers have made significant progress in understanding how HIV counters NK cell responses of the host. This knowledge has opened new avenues for immunotherapy and vaccination against this infection. In the first part of this review article, we gave an overview of our current knowledge of NK cell biology and discussed how the genes encoding NK cell receptors and their ligands determine innate genetic resistance/susceptibilty of humans against HIV infections and AIDS. In this second part, we discuss NK cell responses, viral strategies to counter these responses, and finally, their implications for anti-HIV immunotherapy and vaccination.

Continuous infusion interleukin-2 and famotidine in metastatic kidney cancer

Infusional interleukin-2 (IL-2) is able to elicit lymphokine-activated killer cell (LAK) cytotoxicity against kidney cancer in vitro and in vivo. Famotidine may be able to augment LAK cytotoxicity against neoplastic cells. Fifteen (15) patients were treated with continuous-infusion IL-2 (9-18 MIU/m2/24 hours) for 72 hours and famotidine 20 mg intravenously twice per day. Cycles were repeated every 3 weeks. These patients had a median age of 60 years (range, 29-72), had a median performance status of 1 (range, 0-1), and had metastatic sites, including lung, bone, lymph node, and liver. The most common toxicities of this regimen were hypophosphatemia, fever, nausea/emesis, rigors, elevated creatinine, and hypomagnesemia. One (1) complete and 6 partial responses have been seen (47% response rate). The median duration of response is 9 months. The median survival for all patients is 20 months. Five (5) patients are alive at a median of 36+ months. This combination of infusional IL-2 with famotidine is active in metastatic kidney cancer.

Assessment of the Toxicities of Systemic Low-Dose Interleukin-2 Administered in Conjunction with a Melanoma Peptide Vaccine

In this phase 2 study, the authors assessed the hematologic and clinical toxicities of a melanoma peptide vaccine administered in conjunction with low-dose interleukin-2 (IL-2) therapy. Forty patients were randomized to receive a weekly vaccine paired with a regimen of subcutaneous IL-2 (3 x 10(6) IU/m2/day) administered daily for 6 weeks beginning either at week 1 or at week 4 of vaccine therapy. The differences in the time course of the IL-2 between the two groups permitted assessment of the cause of the toxicities, due either to IL-2 or to vaccine components. Both treatment regimens were well tolerated in the outpatient setting. Toxicities attributable to the vaccine components were principally limited to grade 1 injection site reactions. Systemic clinical toxicities correlated with the initiation of IL-2 therapy. These toxicities coincided temporally and in magnitude with changes in circulating eosinophil counts, suggesting that systemic clinical toxicities and eosinophilia may have common etiologic pathways. Other minor toxicities attributable to this low-dose IL-2 regimen were clinically insignificant hepatic toxicity, mild anemia, and mild thrombocytosis. The hematologic effects of this therapy were delayed in time between the two treatment groups, without dramatic differences in magnitude, which suggests minimal modulation of the IL-2 toxicity by components of the vaccine.

A systematic review of the relation between interleukin-2 schedule and outcome in patients with metastatic renal cell cancer

In Europe, interleukin 2 (IL-2) is one of the two treatment modalities officially approved for patients with metastatic renal cell cancer. Traditionally, IL-2 has been administered by three different routes: intermittent bolus injection (BIV), continuous intravenous infusion (CIV) and subcutaneous injection (SC). There have been few randomized trials designed to compare these routes of administration. This paper describes a systematic review of the literature in which an attempt has been made to determine which schedule of administration is superior. Heterogeneity of the data makes firm conclusions difficult. It appears that the number of complete remissions (CR) is similar between BIV and SC routes and that these are higher than for CIV schedules. The durability of the CRs induced by BIV appeared superior to those induced by SC IL-2 and definitely higher than with CIV protocols. This analysis highlights some of the difficulties of using evidence-based medicine to determine standard of care when the clinical-trial data are heterogeneous. These data emphasize the importance of randomized clinical trials in determining what should be regarded as optimum therapy.

Common cancers--immunotherapy and multidisciplinary therapy: Parts III and IV

The refractoriness of many solid tumors to cytotoxic chemotherapy has led to the exploration of new therapeutic modalities, including immunotherapy. Immunotherapy does not have a direct cytotoxic effect on the cancer cell but is an attempt to promote rejection of the tumor by the host, chiefly through the cellular arm of the immune system. The clinical success with immunotherapy (primarily adoptive immunotherapy) among patients with unresectable malignant melanoma and cancer of the kidney has not been marked by the large numbers of patients responding but by occasional dramatic effectiveness of therapy for these cancers, which usually are refractory to chemotherapy. Long-lasting responses and even complete disappearance of all known metastases are possible for a small percentage of patients with melanoma or renal cell carcinoma who undergo immunotherapy. A reasonable approach for patients with good performance status (no symptoms or ambulatory with symptoms but not bedridden) is entrance to clinical trials, especially trials examining adoptive or active immunotherapy for melanoma or adoptive immunotherapy for renal cancer. The overall treatment of patients with cancer has changed. Primary-care physicians detect almost all cancers. The days when "taking it out" is the best we could offer a patient are over. As we learn more about the use of adjuvant or neoadjuvant chemotherapy and radiation therapy, it is likely one or both of these modalities will be incorporated into the treatment of additional solid tumors previously managed solely with surgical resection. Increasingly, additional therapy is being given for earlier-stage disease as we define how to maximize the potential for cure with minimal toxicity. Many new therapies are on the horizon, including the use of noncytotoxic treatments as an adjunct to a surgical procedure. Such therapies include the use of angiogenesis inhibitors, tumor vaccines, and immunotherapy. Now and in the future, patients will be best served when treated in an environment that can integrate medical, surgical, and radiation oncology actively.

Comparison of three different methods for radiolabelling human activated T lymphocytes

One approach in the treatment of ovarian cancer patients involves the infusion of autologous T lymphocytes coupled with a bispecific monoclonal antibody MOv18/anti-CD3 (biMAb OC/TR), which recognizes a 38-kDa glycoprotein expressed on ovarian carcinomas and the CD3 T cell receptor. However, little is known about the in vivo biodistribution of injected activated lymphocytes, information that could be obtained by scintigraphic imaging of radiolabelled T cells in order to visualize the migratory pattern. We compared the efficiency, stability and toxicity of technetium-99m hexamethylpropylene amine oxime (HMPAO), indium-111 oxine and fluorine-18 2-fluoro-2-deoxy-d-glucose (FDG) in radiolabelling activated lymphocytes targeted with biMAb OC/TR. The mean labelling efficiencies of 111In-oxine and 18F-FDG using 2.5x10(8) lymphocytes (68% and 64%, respectively) were more than twice that of 99mTc-HMPAO (31%). Retention of the radionuclide in the cell was highest in the case of 111In-oxine labelling (less than 25% of the initial cell-bound activity released after 240 min, as compared with 44% of the 99mTc label in the same period and 45% of 18F radionuclide released after 150 min). None of the three radiolabelling reagents induced any significant alteration in cell viability or immunophenotype. However, both 111In-oxine and 18F-FDG induced a loss of cytotoxic activity of lymphocytes against the ovarian carcinoma cell line IGROV1, and all three radiolabelling reagents caused a significant reduction in the proliferative ability of labelled lymphocytes compared to controls, with cell death occurring after 8-9 days. Radiolabelling with the more stable 111In-oxine reagent using a higher number of lymphocytes (1.4x10(9)) but the same total activity (around 55.5 MBq) resulted in improved labelled T cell viability and proliferative ability, although the mean labelling efficiency decreased (35.8%). Together the data suggest that 111In-oxine at low activity per cell is the most appropriate reagent for radiolabelling activated retargeted T lymphocytes useful for in vivo biodistribution studies.

Immunological directives for biotherapy improvement in the treatment of colorectal cancer

A major cause of failure in biotherapy in cancer may be the non-existence of predictive indices to individualize the substances which might be helpful to switch the patients' immune response to tumour from an non-productive to a productive one. The defect in the patients' immune system needs to be identified and so the evaluation of their responses to biotherapeutic agents, in correlation to the disease progression is essential. We have addressed this problem in colorectal cancer at systemic level by examining the proliferative response of peripheral blood mononuclear cells (PBMC) to interleukins (IL) IL-2, IL-4, antiCD3 monoclonal antibody (antiCD3), IL-2+CD3, IL-2+IL-4, IL-4+CD3, IL-2+CD3+IL-4; the PBMC expression of phenotypic antigens CD3, CD4, CD8, CD25, DR, CD16, CD56, CD57 and CD19 in the patients and healthy subjects as control group. Analysing our data, it seems that as the disease progresses in these patients the peripheral blood cells change their ability to respond to activation agents which appears to be due to a phenotypic modification of their subsets. Our overall results might give a possible explanation of the variable responses to biotherapy in colorectal cancer patients.

Interleukin-2-induced splenic enlargement

BACKGROUND

Splenomegaly in patients with cancer raises the suspicion of tumor involvement. Splenic enlargement in the absence of splenic metastases, however, has been reported in patients treated with interleukin-2 (IL-2) immunotherapy. This study characterizes the change in spleen size that occurred in 42 patients treated with IL-2 between 1989 and 1993 for nonhematologic malignancies.

METHODS

Computed tomography (CT) scans before and during immunotherapy were available for review on all 42 patients and after immunotherapy on 16 of these patients. The splenic index was measured for each CT by a single reader blinded to the time course of IL-2 therapy.

RESULTS

Mean splenic index increased 64.1% from 646 cm3 (standard deviation [SD], 387) pre-IL-2 to 1059 cm3 (SD, 534) during therapy with IL-2 (P < 0.0001). The splenic index remained elevated at 1112 cm3 (SD, 633) after completion of IL-2 therapy.

CONCLUSIONS

Splenomegaly, temporally associated with IL-2 therapy for nonhematologic malignancies, is likely to represent a sequela of therapy and not tumor progression.

Longer local retention of adoptively transferred T-LAK cells correlates with lesser adhesion molecule expression than NK-LAK cells

The local retention of adoptively transferred lymphokine (IL-2)-activated killer (LAK) cells was examined in 11 patients with head and neck carcinoma. Unseparated lymphocytes, T and natural killer (NK) cells isolated from patients were cultured with IL-2 for 7 days, labelled with 99mTc-HMPAO, and immediately injected back into the respective donors via the superficial temporal artery or locally into the tumour tissue. The injected LAK cells were periodically traced using a gamma camera, and the LAK cell retention rate was calculated from the radioactivity. One hour after the injection, about 70% of the locally infiltrated LAK cells remained in the tumour tissue, while about half of the LAK cells transferred via the regional artery were dislodged from the tissue. LAK cells induced from T cells (T-LAK) were retained in the tissue for a longer time than LAK cells induced from NK cells (NK-LAK). T-LAK were less chemotactic and less adherent to human umbilical vein endothelial cells (EC), and showed lesser migration through EC. Flow cytometric analysis revealed higher expression of CD11a, CD11b, CD18 and CD49d on NK-LAK compared with T-LAK. MoAbs against these adhesion molecules suppressed adhesion and migration of LAK cells. These results indicate that the rapid disappearance of NK-LAK from the tissue is associated with their greater chemotactic and adhesive as well as migratory activities depending on differing expression of adhesion molecules.

Recruitment of alloreactive natural killer cells to the rat peritoneum by a transfected cell line secreting rat recombinant interleukin-2

In order to obtain large numbers of natural killer (NK) cells from single rats for functional studies, we have devised a method for the generation of IL-2-activated NK cells in vivo. Rats were implanted intraperitoneally with cell-impermeable diffusion chambers (DC) containing cultures of transfected Chinese hamster ovary (CHO) cells secreting rat recombinant interleukin-2 (rIL-2). This resulted in a dramatic increase in the peritoneal exudate cell (PEC) number with a peak (300-1000 x 10(6)) 1 week after implantation. The majority were mononuclear cells of which a large proportion were CD3-NKR-P1+ NK cells, but with substantial numbers of macrophages (M phi) and CD3+8+NKR-P1+ T cells also. The NK activity against standard tumor target cells was high among PEC from six different inbred rat strains tested. However, the NK cell-mediated reactivity against concanavalin A (ConA)-activated T cell blasts from a panel of major histocompatibility complex (MHC) congenic strains differed widely. PEC from some strains (PVG, LOU/C, and AO) efficiently lysed all the MHC-disparate lymphoblasts. In other strains (BN and LEW) more restricted allorecognition repertoires were observed, whereas PEC from one strain (DA) were unresponsive. The secretion of rat rIL-2 intraperitoneally did not lead to a significant increase in the IL-2 level in the blood or in the total number or activity of NK cells in blood and spleen. The present method represents a most potent technique for generating large numbers of functional rat NK cells and shows the high efficiency with which IL-2 can induce NK cell recruitment in vivo.

Phase II study of low dose cyclophosphamide and intravenous interleukin-2 in metastatic renal cancer

Thirteen patients with metastatic renal cancer were treated in a phase II trial with interleukin-2, 21.6 million IU/m2 intravenously daily for five days on two consecutive weeks, starting 3 days after the administration of low dose cyclophosphamide 350 mg/m2 intravenously. Treatment cycles were repeated every 21 days. No responses were seen (95% Confidence Interval: 0-22%). The most common toxicities were fever, fatigue, hypotension, nausea/emesis, and myalgia/arthralgia. There were 11 episodes of Grade III toxicity including Grade III hypotension in 7 patients. Because of the significant toxicity and the lack of observed response, the study was discontinued. Cyclophosphamide and interleukin-2 at the dose and schedule used in this study has considerable toxicity and is unlikely to improve on response rates previously seen with other IL-2 based regimens in metastatic renal cancer.

Renal, metabolic, and hemodynamic side-effects of interleukin-2 and/or interferon alpha: evidence of a risk/benefit advantage of subcutaneous therapy

Systemic immunotherapy with recombinant interleukin-2 (rIL-2) via intravenous (i.v.) and subcutaneous (s.c.) administration produces objective responses in a proportion of advanced cancer patients. While most of the previous investigations chose the i.v. route for cytokine application, there is an increasing number of trials employing s.c. rIL-2 therapy. The comparison of reported response rates for i.v. versus s.c. therapy reveals no significant differences between these modalities. In an effort to describe systemic toxicities of s.c. cytokine therapy with regard to renal, metabolic, and hemodynamic abnormalities and to compare these effects to toxicities reported upon i.v. therapy, we retrospectively evaluated 148 treatment cycles of s.c. immunotherapy given to 107 outpatients. Our study cohorts consisted of 15 patients who received s.c. rIL-2 at doses of (4.8-14.4) x 10(6) IU m-2 day-1 5 days/week for a total of 8 weeks, 20 patients who received rIFN alpha 2b at (3.0-6.0) x 10(6) m-2 day-1 thrice weekly for a total of 6 weeks, and 72 patients who were given s.c. rILFN alpha 2b at 6.0 x 10(6) U/m2, three times per week, plus s.c. rIL-2 at (14.4-18.0) x 10(6) IU/m2 on days 1 and 2, followed by 4.8 x 10(6) IU m-2 day-1 5 days/week for 6 consecutive weeks. These treatment regimens were well tolerated in the outpatient setting; no toxic death occurred, and none of the patients developed life-threatening toxicity due to a capillary leak syndrome. Upon s.c. combination therapy, dyspnea at rest occurred in 6% of patients and grade III and IV hypotension occurred in 7% and 4%, respectively; plasma protein was significantly decreased (mean nadir +/- standard deviation, 67 +/- 5 g/l). In addition, s.c. therapy led to a significant increase in serum creatinine (mean peak +/- standard deviation, 115.1 +/- 21.4 mumol/l) and urea nitrogen (mean peak +/- standard deviation, 6.5 +/- 2.5 mmol/l); electrolyte disturbances and direct nephrotoxicity never caused major clinical symptoms. This was in marked contrast to a multitude of dose-limiting and life-threatening adverse reactions reported upon i.v. rIL-2 therapy. We conclude that palliative low to intermediate-dose s.c. rIL-2/rIFN alpha combination therapy, in contrast to i.v. treatment, can be administered in the ambulatory setting with good practicability and excellent safety. This outpatient regimen is as effective against metastatic renal cell cancer as the most aggressive i.v. rIL-2 protocol reported.

Adoptive immunotherapy with peripheral blood lymphocytes cocultured in vitro with autologous tumor cells and interleukin-2

A clinical trial of adoptive immunotherapy was carried out with peripheral blood lymphocytes (PBL), cocultured in vitro with autologous tumor cells and interleukin-2 (IL-2), in 14 patients with advanced melanoma. PBL from these patients were cocultured with irradiated autologous tumor cells for 7 days, which was followed by expansion in IL-2-containing medium. These lymphocytes were returned to the patient along with intravenous IL-2 at doses up to 2 x 10(6) IU m-2 day-1. A dose of 300 mg/m2 cyclophosphamide was administered to each patient intravenously 4 days prior to each treatment. Following coculture, the lymphocytes were primarily CD3+ T cells and they expressed varied degrees of cytotoxicity against autologous melanoma cells. In 9 patients the activated cells were at least 80% CD4+ and in 2 cases they were mostly CD8+. Some of the activated cells exhibited suppressor or helper activity in a functional regulatory coculture assay. No major therapeutic response was observed in this study. Minor responses were observed in 2 patients. Toxicities were those expected from the IL-2 dose administered.

Clinical toxicity of interleukin-2

Interleukin-2 (IL-2) is increasingly used to treat patients with cancers refractory to conventional treatment. Flu-like syndromes are extremely frequent but usually mild. A variety of skin complications (mostly erythema and mucositis) have been reported. Life-threatening skin reactions have also been described. Acute reactivation of psoriasis can also occur. Immediate hypersensitivity reactions have so far not been described, but IL-2 treatment has been shown to predispose to acute hypersensitivity reactions to iodine-containing contrast media. Hypothyroidism is the major endocrine complication and antithyroid antibodies have been detected in approximately 50% of patients. Neurological and psychiatric disturbances with moderate or severe mental status changes are common and sometimes treatment-limiting. The occurrence of peritumoural oedema in patients with brain metastases can also be a major practical problem. Musculoskeletal disorders are transient and resolve spontaneously. The vascular leak syndrome is the most frequent and severe complication of IL-2 of which weight gain, generalised oedema, hypotension and impaired renal function are the main features. Even though a damaging effect on vascular endothelium cells by various cytokines released by activated lymphoid cells or mediated by non-lymphocyte-dependent factors has been proposed to be involved, the mechanism remains unclear. Other cardiovascular injuries, possibly life-threatening, including myocarditis, angina pectoris and myocardial infarction, can occur during the first days of treatment. Supraventricular arrhythmias are the most common rhythmic disorder. Decreases in myocardial contractility and haemodynamic pattern similar to those of septic shock have been encountered in most cases. Acute renal dysfunction is common but resolves with symptomatic management. Intrahepatic cholestasis with hyperbilirubinaemia is observed in most patients but permanent liver damage has not been described. Several cases of pancreatitis have been reported. Anaemia, thrombocytopenia, lymphocytopenia and eosinophilia are frequent and occur in most if not all patients. Some data suggest a high incidence of infectious complications, particularly in patients with surgically tunnelled catheters, but marked flu-like syndromes may be confounding. Finally, death directly related to IL-2 treatment has been noted in less than 1% of all patients. Investigations are under way to minimise IL-2 toxicity with varying dose regimens and combined treatments.

Induction of lymphokine-activated killer cells with low-dose interleukin 2 and interferon-gamma in oral cancer patients

Lymphokine-activated killer (LAK) cells were induced with low-dose recombinant interleukin 2 (rIL-2) and recombinant interferon-gamma (IFN-gamma) in 28 oral carcinoma patients. The patients received daily intravenous injections of rIL-2 (1.2 x 10(5) U/m2) and rIFN-gamma (7.0 x 10(4) U/m2), and both natural killer (NK) and LAK activities were periodically examined. A significant increase in CD16+CD57+ and CD16+CD57- NK subsets was observed after the induction. An increase in the T-cell population was also found, with a significant increase in CD3+HLA-DR+, CD8+Leu8-, and CD4+Leu8- cells. Significant increases in NK activity, from the original level of 32.0 +/- 13.7 to 49.9 +/- 15.2%, and LAK activity, from 4.8 +/- 3.5 to 11.0 +/- 6.1%, at Day 7 were observed. Both activities were maintained at high levels during the cytokine injections, but greater enhancement of the killing activities could not be obtained subsequently. When NK and LAK activities were investigated in each subpopulation of CD3- and CD16- cells, no remarkable cytotoxic activity could be observed before induction in any subset without NK activity in CD3- cells (31.1 +/- 14.3%). At Day 7, NK activity of CD16- cells increased up to 21.4 +/- 14.9%, accompanied by an increase in CD3(-)-cell activity (54.5 +/- 20.6%). LAK activities of both subsets were also enhanced, with activity at Day 7 of 6.5 +/- 5.6 and 9.4 +/- 6.6% in CD16- and CD3- cells, respectively. These increased activities were maintained at the same level during the induction. Phorbol myristate acetate-induced polymorphonuclear leukocyte (PMNL) O2-generation was significantly increased, from the original 81.1 +/- 28.1 to 95.6 +/- 34.9 pmol/min/10(4) cells, after 1 week of treatment. Protein kinase C activity in the cytosol decreased, and the activity in the membrane fraction conversely increased. No remarkable adverse effects except for mild fever were observed. Together with LAK induction ability and PMNL enhancement, with scarce toxicity, a combination of low-dose rIL-2 and rIFN-gamma is thought to be useful in cancer treatments.

Continuous infusion of recombinant interleukin-2 with or without autologous lymphokine activated killer cells for the treatment of advanced renal cell carcinoma

Data have been analysed for 327 patients with advanced renal cell carcinoma receiving a continuous infusion of recombinant interleukin 2 (rIL-2) alone (225 patients) or rIL-2 plus lymphokine activated killer (LAK) cells (102) on a normal oncology ward. Eligibility criteria were uniform across protocols, all patients having advanced progressive disease, but with an ambulatory performance status. The baseline characteristics of patients receiving rIL-2 alone did not differ significantly from those receiving LAK, with the exception that the LAK treated patients had a better performance status. Despite similar treatment intensity, toxicity was more severe in the patients receiving LAK. The addition of LAK did not lead to higher response rates or to prolonged response duration, progression-free survival or survival. This review confirms the activity of rIL-2 for the treatment of advanced renal cell carcinoma and demonstrates that the addition of LAK cells does not lead to increased efficacy.

Tumorigenicity of interleukin-2 (IL-2)-cDNA-transfected L1210 lymphoma and its in vivo variants is modulated by changes in IL-2 expression; potential therapeutic implications

To study parameters that affect the tumorigenicity of L1210 lymphoma we have analyzed the structure of MHC class I antigens of this tumor. In addition this tumor was transfected with interleukin-2 (IL-2) cDNA in order to determine the effects of high concentrations of IL-2 within the tumor environment. The nucleotide sequence of the class I Kd, Dd and Ld mRNAs from this tumor showed that the encoded amino acid sequence of the corresponding antigens is normal, thus suggesting that the tumorigenicity of L1210 lymphoma is not due to defective antigen presentation to tumor-specific cytotoxic T cells. In contrast, induction of IL-2 expression by cDNA transfection led to loss of tumorigenicity of the IL-2-secreting tumor cells. However, a fraction of long-term-surviving mice developed progressively growing variant tumors that showed substantial decrease or loss of IL-2 expression. These results suggest that IL-2 secretion by tumors is suicidal but, because of tumor heterogeneity, IL-2-loss-variant tumors may arise that are able to escape the immune defenses of the host. The observed consistent loss of IL-2 expression in variant tumors implies that specific targeting of large quantities of IL-2 to tumor cells may be a valuable approach to immunotherapy of cancer. In addition we find that under specific gamma ray irradiation IL-2-secreting tumor cells lose their ability to multiply yet continue to secrete IL-2 at levels equivalent to those secreted by unirradiated cells. Such IL-2-secreting irradiated tumor cells were found to be superior immunogens in comparison to the irradiated parental tumor cells, suggesting their use as tumor vaccines.