Monoclonal antibodies against the 4-1BB T-cell activation molecule eradicate established tumors.

The 4-1BB glycoprotein is a member of the tumor necrosis factor receptor superfamily and binds to a high-affinity ligand (4-1BBL) expressed on several antigen-presenting cells such as macrophages and activated B cells. Expression of 4-1BB is restricted to primed CD4+ and CD8+ T cells, and 4-1BB signaling either by binding to 4-1BBL or by antibody ligation delivers a dual mitogenic signal for T-cell activation and growth. These observations suggest an important role for 4-1BB in the amplification of T cell-mediated immune responses. We now show that administration of anti-4-1BB monoclonal antibodies can eradicate established large tumors in mice, including the poorly immunogenic Ag104A sarcoma and the highly tumorigenic P815 masto cytoma. The immune response induced by anti-4- 1BB monoclonal antibodies is mediated by both CD8+ and CD4+ T cells and is accompanied by a marked augmentation of tumor-selective cytolytic T-cell activity. Our data suggest that a similar approach may be efficacious for immunotherapy of human cancer.

Costimulation by CD48 and B7-1 induces immunity against poorly immunogenic tumors.

Genetic modification of many types of mouse tumors to express the B7-1 or B7-2 molecules, natural ligands for the T cell-costimulatory molecule CD28, increases their immunogenicity. However, even after transfection with the B7-1 and/or B7-2 genes, poorly immunogenic tumors fail to elicit and efficient immune response. We report here that two such tumors, the Ag104A sarcoma and the K1735-M2 melanoma, become immunogenic after transfection of the genes encoding murine B7-1 together with CD48, which is the natural ligand for CD2. Tumor-specific CD8+ cytotoxic T lymphocytes were readily generated and were effective for adoptive immunotherapy of metastasis induced by wild-type Ag104A sarcoma cells. A similar approach may be useful for developing therapy for other poorly immunogenic tumors, including those in humans.

Regression and inhibition of sarcoma growth by interference with a radiosensitive T-cell population.

BALB/c mice were inoculated subcutaneously with 10(6) cells from either of two syngeneic sarcomas 1315 and 1425. 6--8 days later, the mice were randomized into groups which were left untreated or given 400 rads of whole body irradiation. Irradiation significantly retarded the growth of both sarcomas, and complete regressions were seen of approximately equal to 30% of the small, established 1315 tumors. The anti-tumor effect of irradiation was abolished if the irradiated mice were inoculated with a T-cell-enriched (but not with a T-cell deprived) suspension of syngeneic spleen cells, suggesting that the irradiation inhibited tumor growth by affecting a radiosensitive population of host suppressor T cells.

Cell-mediated immunity and blocking serum activity to tolerated allografts in rats.

W/Fu rats were neonatally inoculated with bone marrow cells from B/N rats and vice versa. Of the inoculated rats, some were capable of accepting a foreign (B/N or W/Fu) skin graft over the period of observation (i.e. for more than 100 days), while other rats rejected their skin grafts as early as control animals (within 8-12 days) or after a prolonged period of acceptance (20-96 days). Using a microcytotoxicity test, it could be shown that both those rats that rapidly rejected skin grafts and those that kept their grafts during the observation period had lymphocytes capable of destroying cultivated allogeneic cells from the respective strains with whose cells the rats had been inoculated as newborns. The degree of lymphocyte reactivity decreased upon time, so that 4 of 13 rats that had carried "tolerated" skin grafts over more than 84 days had lymphocytes which were nonreactive in the highest dose tested, and the degree of reactivity in the other 9 rats was less than seen early after tolerance induction. Rats that were capable of accepting skin grafts over prolonged periods of time had sera that could specifically block lymphocyte-mediated cytotoxicity, while sera from rats that had rejected their grafts did not block. Sera from rats that rejected their skin grafts after 20-96 days lost the blocking activity 3-10 days before rejection.

Tumor immunogenicity determines the effect of B7 costimulation on T cell-mediated tumor immunity.

A costimulatory signal through B7 to its counter-receptor CD28 on T cells enhances T cell activation. We have generated recombinant retroviruses containing cDNA for murine B7 and transduced a panel of murine tumor lines with varying immunogenicity to study the effect of B7 costimulation on antitumor immunity. In contrast to the progressive outgrowth of all wild-type (B7-) tumors in unimmunized syngeneic mice, four immunogenic tumors, lymphoma RMA, EL4, mastocytoma P815, and melanoma E6B2, regressed completely when transduced with the B7 gene. In contrast, four nonimmunogenic tumors, sarcomas MCA101, MCA102, and Ag104, and melanoma B16, remained tumorigenic after transduction of the B7 gene. Immunization with B7-transduced immunogenic tumors enhanced protective immunity and increased specific cytotoxic T lymphocyte (CTL) activity against the respective wild-type tumors as compared to immunization with nontransduced or mock-transduced tumors. Moreover, cocultivation of CTL with B7-transduced EL4 cells augmented the specificity of tumor-reactive CTL in long-term cultures. Treatment by injection of B7-transduced tumor cells cured 60% of mice with established wild-type EL4 lymphoma. In contrast, immunization with nonimmunogenic tumors transduced with B7 did not provide protective immunity and did not increase specific CTL activity. Our results show that tumor immunogenicity is critical to the outcome of costimulation of T cell-mediated tumor immunity by B7.

Use of combinatorial peptide libraries to construct functional mimics of tumor epitopes recognized by MHC class I-restricted cytolytic T lymphocytes.

Identification of cytolytic T lymphocyte (CTL) epitopes presented by major histocompatibility complex (MHC) class I molecules on tumor cells is critical for the design of active immunotherapy. We describe the use of combinatorial peptide libraries with defined amino acids in two MHC anchor positions to search for epitopes that are recognized by H-2Db- and Kb-restricted CTL specific for the mouse lymphoma EL4. An iterative strategy was used for screening libraries in which 16 amino acids were divided into 3 groups and 3 subgroups: alpha (AL, VT, FY); beta (GS, P, DE); gamma (KR, H, NQ). The proportions of each group and subgroup at individual peptide positions were changed in the library synthesis, and the effect of these changes on CTL activity was measured in a sensitive RMA-S cell assay. A single H-2Db epitope mimic was deduced from the original library that contained > 2 x 10(8) potential peptides and was at least 9 logs more potent than the original library. Immunization of syngeneic mice with this peptide elicited CTL that lysed EL4 cells as well as RMA-S cells pulsed with peptides isolated from Db molecules of EL4 cells, indicating functional similarity between the mimicking peptide and the naturally processed CTL epitope. Furthermore, adoptive transfer of such a CTL line had a therapeutic effect in mice with EL4 established as an ascites tumor. Two H-2Kb-restricted epitope mimics of the same tumor were also identified. Our method represents a novel approach for the construction of MHC class I-restricted targets that can serve as immunogens for active immunotherapy of cancer.

B7-CD28 costimulation unveils the hierarchy of tumor epitopes recognized by major histocompatibility complex class I-restricted CD8+ cytolytic T lymphocytes.

Immunization of mice with tumors genetically engineered to express the B7 costimulatory molecules amplifies the antitumor immune response mediated by CD8+ cytolytic T lymphocytes (CTL). In this report, we examined the effect of B7-CD28 costimulation on the hierarchy of tumor epitopes. Using a combination of affinity chromatography/reversed-phase high performance liquid chromatography and CTL cloning, we show that major histocompatibility complex (MHC) class I molecules from EL4 lymphoma cells can present at least six distinct CTL epitopes presented by MHC class I molecules. Nevertheless, mice immunized with wild-type B7-negative EL4 cells develop CTL only to one immunodominant epitope. In contrast, immunization with B7-transduced EL4 cells led to not only the amplification of the CTL response to this immunodominant epitope, but also to the recognition of five otherwise silent subdominant epitopes. The adoptive transfer of a CTL clone against such a subdominant epitope cured mice bearing EL4 lymphoma growing as an ascites tumor. The fact that CTL response can be spread to normally silent epitopes as a result of B7-CD28 costimulation suggests a novel approach to manipulate the hierarchy of CTL epitopes and offers an opportunity to explore novel targets for T cell-mediated cancer therapy.

Anti-tumor antibody BR96 blocks cell migration and binds to a lysosomal membrane glycoprotein on cell surface microspikes and ruffled membranes.

BR 96 is an internalizing antibody that binds to Lewis Y (Le(y)), a carbohydrate determinant expressed at high levels on many human carcinomas (Hellström, I., H. J. Garrigues, U. Garrigues, and K. E. Hellström. 1990. Cancer Res. 50:2183-2190). Breast carcinoma cell lines grown to confluence bind less BR96 than subconfluent cultures (Garrigues, J., U. Garrigues, I. Hellström, and K. E. Hellström. 1993. Am. J. Path. 142:607-622). However, when the confluent cells are induced to migrate by scratch wounding, they again bind BR96 suggesting that antigens bearing the Le(y) determinant may promote cell migration. In the present study, BR96 was found to be highly enriched on microspikes and ruffled membranes, cell surface structures involved in cell migration. In addition, BR96 was a potent inhibitor of cell migration in vitro. When stationary BR96 treated cells were exposed to fresh culture media, membrane ruffles and microspikes developed at the cell margin and migration resumed. Immunogold microscopy showed that BR96 antigens were enriched on these membrane protrusions. BR96 cell surface immunoprecipitation analysis of 3H-glucosamine labeled breast carcinoma cells identified antigens with approximate molecular weights of 135 kd (upper antigen) and 85 kd (lower antigen). A short amino terminal sequence (8 residues) of the upper antigen matched that of human lysosomal membrane glycoprotein 1 (LAMP-1). In addition, the upper antigen was detected on immunoblots probed with anti-LAMP-1, and within the intracellular compartment BR96 was found predominantly in endosomes and lysosomes. A soluble LAMP-1/immunoglobulin fusion protein (LAMP-1/Ig) was transiently expressed in both BR96 binding and nonbinding cell lines. Immunoblot analysis of LAMP-1/Ig's from the various cell lines showed that (a) acquisition of the BR96 epitope is probably controlled at the level of polylactosamine modification (e.g., fucosylation) rather than LAMP-1 gene expression; (b) alternate forms of LAMP-1/Ig comigrate with the lower BR96 antigen raising the possibility that it may be a degradation product of the upper antigen; and (c) LAMP-1/Ig expressed in 3396 breast carcinoma cells has approximately 30-fold more BR96 epitopes than LAMP-1/Ig from non-tumorigenic mammary epithelial cells. Together these data indicate that a major BR96 antigen, LAMP-1, is present on unique cell surface domains involved in cell locomotion as well as membranes of the endocytic compartment. Altered glycosylation of LAMP-1 expressed in transformed cells may contribute to their ability to disseminate.

The melanoma proteoglycan: restricted expression on microspikes, a specific microdomain of the cell surface.

A cell surface chondroitin sulfate proteoglycan associated with human melanomas and defined by mAb's F24.47 and 48.7 has been characterized biochemically and localized by indirect immunogold electron microscopy. These antibodies recognize distinct epitopes on the intact proteoglycan. In addition, mAb 48.7 also recognizes an epitope on a 250,000-D glycoprotein and is therefore similar to antibody 9.2.27 (described by Bumol, T.F., and R.A. Reisfeld, 1982, Proc. Natl. Acad. Sci. USA., 79:1245-1249). Furthermore, it was shown that the glycosaminoglycan chains released by alkaline borohydride treatment of the proteoglycan recognized by mAb 48.7 had a size of approximately 60,000 D. Since the intact proteoglycan was estimated to be 420,000 D, there are probably three chondroitin sulfate chains attached to the 250,000-D core glycoprotein. Furthermore, an oligosaccharide fraction containing 42% of the 3H activity (glucosamine as precursor) was isolated. Immunolocalization studies using whole-mount electron microscopy revealed that the chondroitin sulfate proteoglycan was present almost exclusively on microspikes, a microdomain of the melanoma cell surface. These processes were present as 1-2-micron structures on the upper cell surface and as longer (up to 20 micron) structures at the cell periphery. Peripheral microspikes were involved in the initial interactions between adjacent cells and formed complex footpads that made contact with the substratum. Immunogold-labeled cells were also thin sectioned and the specific localization of the chondroitin sulfate proteoglycan antigen was quantitated. The data confirmed the results of whole-mount microscopy and demonstrated a statistically significant association of the antigen with the microspike processes as compared with other areas of the cell surface. By using two different mAb's (48.7 and F24.47) that recognize epitopes on either the core glycoprotein or the intact proteoglycan, respectively, we have demonstrated that both molecules have the same restricted distribution at the cell surface. The specific localization of the antigen to microspikes at the cell surface suggests it may play a role in cell-cell contact and cell-substratum adhesion, which could be important in the metastatic process.

Cell-bound immunity to autologous and syngeneic mouse tumors induced by methylcholanthrene and plastic discs.

A colony-inhibition assay, was lused to demonstrate specific immunity in vitro against syngeneic mouse sarcomas induced by mlethlylcholanthrene or implanted Dictabelt plastic discs. An immunity to methylcholanthrene-induced tumors could also be demonstrated with lymphocytes derived from the autologous primary tumor-bearing host 3 to 8 days after removal of the primary tumor.

Immunological studies on urinary bladder tumors of rats and mice.

Human neoplasms derived from the same tissue have been previously shown to have tumor associated antigens characterizing that tissue type. Evidence is now presented for the existence of analogous antigens common to both rat bladder papillomas and carcinomas, and for antigens common to mouse bladder carcinomas. Rats immunized with syngeneic urinary bladder papillomas, then challenged with a methylcholanthrene pellet inserted into the bladder, develop (4 to 6 months later) fewer primary bladder tumors than rats immunized with normal bladder tissue.

Cure of xenografted human carcinomas by BR96-doxorubicin immunoconjugates.

Immunoconjugates (BR96-DOX) were prepared between chimeric monoclonal antibody BR96 and the anticancer drug doxorubicin. The monoclonal antibody binds an antigen related to Lewis Y that is abundantly expressed at the surface of cells from many human carcinomas; it has a high degree of tumor selectivity and is internalized after binding. BR96-DOX induced complete regressions and cures of xenografted human lung, breast, and colon carcinomas growing subcutaneously in athymic mice and cured 70 percent of mice bearing extensive metastases of a human lung carcinoma. Also, BR96-DOX cured 94 percent of athymic rats with subcutaneous human lung carcinoma, even though the rats, like humans and in contrast to mice, expressed the BR96 target antigen in normal tissues.

Species restriction of a monoclonal antibody reacting with residues 130 to 137 in encephalitogenic myelin basic protein.

A monoclonal antibody (immunoglobulin G1) has been produced that reacts against myelin basic protein present in or extracted from the brains of many mammals-with certain important exceptions. Because of known species differences in amino acid sequences of basic protein and of certain peptide fragments, the binding site for this particular antibody appeared likely to include residues 130 to 137. Confirmation of this hypothesis was obtained by amino acid composition of the major immunoreactive peptides produced by thermolysin digestion of human basic protein and isolated by high-performance liquid chromatography.

Localization of 131I-labeled p97-specific Fab fragments in human melanoma as a basis for radiotherapy.

33 patients with advanced malignant melanoma were studied after intravenous administration of 131I-labeled Fab fragments specific for p97, an oncofetal glycoprotein of human melanoma. In all, 47 gamma camera imaging studies were performed for the purpose of localization of metastatic deposits. In addition to tumor, 131I-Fab uptake was also seen in liver and kidney. 20 of these studies included simultaneous administration of both an 131I-labeled Fab specific for p97, and an 125I-labeled Fab not specific for p97. Blood clearance of p97-specific Fab was significantly more rapid than for nonspecific Fab. Eight of these patients had biopsies of subcutaneous nodules at 48 and 72 h postinjection in order to assess whether localization of radioactivity was antigen specific. Antigen-specific localization was observed with average ratios of specific/nonspecific uptake of 3.7 (48 h) and 3.4 (72 h); uptake was strongly correlated with tumor p97 concentration (r = 0.81, P less than 0.01). Also, imaging studies of the bio-distribution of 131I-labeled anti-p97 Fab in patients selected for high p97 tumor concentration showed avid tumor uptake and more prolonged retention of labeled Fab in tumor than in normal tissues. Based on these studies, we estimated that total 131I doses of 500 mCi could be safely given to patients before dose-limiting toxicity would be observed. Accordingly, in seven selected patients, phase I radiotherapeutic trials were begun. For improved radiation safety, we developed automated methods to label Fab fragments with up to 200 mCi of 131I. So far, a total of 12 individual therapeutic doses, ranging from 34 to 197 mCi of 131I-labeled to 5 to 10 mg of Fab, have been administered with excellent tumor localization and without major target organ toxicity. Cumulative doses ranged from 132 to 529 mCi 131I. Side effects attributable to the radiation were mild, with a transient drop slightly greater than 50% in platelet and absolute neutrophil counts being observed in the two patients who received cumulative doses greater than 500 mCi. In the combined series of 47 diagnostic and 12 therapeutic studies, four acute reactions were observed: one episode each of transient chills and fever; flushing and hypotension; and two skin rashes. All of these reactions responded promptly to symptomatic therapy. After multiple administrations of 131I-(anti-p97) Fab (IgG1), isotype-specific immunity was observed in three patients. In two of these patients it was possible to successfully reinfuse after immunity had developed with 131I-(anti-p97) Fab of a different isotype (IgG2a). Dosimetry estimates were performed based on the biodistribution of (131)I-Fab in these patients,and for every 100 mCi of (131)I-Fab given, tumor receives 1,040 rads; liver. 325 rads; and bone marrow, 30 rads. Marrow would be expected to be the critical organ, if doses >500 mCi (131)I-Fab are given. These studies demonstrated that, with proper precautions, large doses (of an (131)I-labeled murine Fab fragments immunologically specific for a human melanoma-associated antigen) could be safely given to humans by using repetitive intravenous injections.

Immunity to tumor antigens: potential implications in human neuroblastoma.

Immune reactions to tumor-specific and tumor-associated antigens have been demonstrated in animals with neoplasms with in vitro and in vovo techniques. Some of the antigens detected in vitro induce transplantation resistance in vivo, while others do not. Human neuroblastoma cells cultivated in vitro have been shown to possess common antigens to which lymphocytes from neuroblastoma patients react. Whether it is possible to augment the immune reactivity of patients with neuroblastoma to these common antigens and, if so, whether this heightened immune reactivity would have clinically beneficial effects are as yet unknown. These reactions are complex, involving both cellular and humoral mechanisms. The fact that one type of immune response can be detected to one type of antigen present in a tumor in vitro does not necessarily mean that the immune response is effective in vivo. Responses to other tumor antigens may be deficient, or the immune response may be depressed. This may be due to active suppression of and/or selective deficiencies in critical cell populations required for an augmented immune response; this possibility may be evaluated with techniques allowing for in vitro sensitization to tumor antigens.

Transfected mouse melanoma lines that express various levels of human melanoma-associated antigen p97.

The gene for the human melanoma-associated antigen p97 has been introduced by cDNA transfection into cells from clone M2 of the K1735 mouse melanoma, which metastasizes to the lung when injected iv into syngeneic C3H/HeN mice. Tumor clones were established from the transfected cells and found to differ in the level of p97 expression. Their outgrowth in immunocompetent syngeneic mice was shown to inversely correlate with p97 antigen expression, and lines that express higher p97 levels elicited a stronger delayed-type hypersensitivity response when injected into the footpads of mice immune to p97. Five clones which expressed very high levels of p97 failed to grow in immunocompetent C3H/HeN mice while they formed tumors in nude (nu/nu) mice. The highest expressing clone, 2A, grew slightly faster than any of the other clones when cultured in vitro. Since several of the transfected clones were found to express a stable level of p97 and have consistent in vivo growth behavior, they provide a useful model for various forms of antigen-specific active and passive immunotherapy with the same agents as those intended for human application.

The need to establish whether the leukocyte adherence inhibition test is a reliable assay of tumor immunity in humans.

Lymphocytes from human cancer patients can react to tissue-type-specific antigens shared by many tumors of the same histological type and different for tumors of different types. Such reactions have been detected, for example, by using microcytotoxicity assays and leukocyte migration inhibition techniques. Work presented at this workshop indicates that the leukocyte adherence inhibition test is an excellent assay for detecting reactivity to antigens shared by neoplasms of the same histological type. However, very little is known of the nature of these antigens; for example, we do not know whether they are tumor specific or just normal tissue antigens. The usefulness of the leukocyte adherence inhibition test for patient monitoring and for diagnostic purposes also needs to be studied more.

Antibody-mediated killing of human tumor cells by attached effector cells.

Cultured human melanoma, lung carcinoma, and colon carcinoma cells were isotope labeled and incubated with a combination of effector cells and mouse monoclonal antibodies to tumor-associated cell surface antigens. The former were derived from the peritoneal cavity of mice or from peripheral blood of healthy human subjects. Monoclonal antibodies MG-21, 96.5, and L6, which are IgG3, IgG2a, and IgG2a, respectively, were all cytolytic when added in the presence of mouse effector cells to target cells expressing the relevant antigens. MG-21 and L6 were cytolytic also with human effector cells, while monoclonal antibody 96.5 was not. The effector cells attached to plastic surfaces, stained with neutral red, were peroxidase positive and mediated their effect over a 24- to 72-h time period as compared to the 4 h generally sufficient for antibody-dependent cellular cytotoxicity by natural killer cells. In tests on human effector cells with a fluorescence-activated cell sorter, they stained with antibody LCM-3C10 to the CD14 antigen, as well as with antimonocyte antibody 61D3. The cytolytic effect of human effector cells and antitumor antibody was not abolished by incubation with antibodies FC2 or 60.3 to CD16 and CD18, respectively, known to interfere with the antibody-dependent cellular cytotoxicity activity and natural killing of natural killer cells. This suggests, together with the other findings, that the effector cells were macrophages.

An unmodified anticarcinoma antibody, BR96, localizes to and inhibits the outgrowth of human tumors in nude mice.

The antitumor effects of an unmodified murine monoclonal antibody, BR96, were examined in nude mice bearing human lung adenocarcinoma xenografts. BR96, a murine IgG3 that internalizes and is cytotoxic to cells expressing the antigen in vitro, also elicits strong antibody-dependent cellular cytotoxicity and complement-dependent cytotoxicity effector functions. Its in vivo antitumor effects were compared with those of its F(ab')2 fragments, a mouse-human chimeric form, and an IgG1 class switched variant of the original (IgG3) BR96. Antitumor effects were observed with antigen-positive tumor lines (but not with tumors which did not bind with BR96) and correlated with the levels of antigen expression as detected in vitro. The chimeric form of BR96 gave the strongest antitumor effects, followed by the murine IgG3, while limited effects were seen with the IgG1 and with F(ab')2 fragments of BR96, indicating that Fc-dependent host effector functions are primarily responsible for its in vivo activity. The antitumor effects observed were modest unless the antibody treatment was started on the day following tumor grafting.

Long-term transplantability and morphological stability of three experimentally induced urinary bladder carcinomas in rats.

Three transitional cell carcinomas induced in Fischer 344 rats by a methylcholanthrene pellet or a foreign body inserted locally into the bladder have been serially transplanted in the syngeneic strain for up to 6.5 years. There have been no changes in the individual morphological characteristics of the tumors during this time. Cells cultured in vitro for varying numbers of passages reproduce regularly the morphology of each tumor when they are injected back into the animals and results from a microcytotoxicity assay for cellular immunity indicate that they retain a common, bladder tumor-specific antigen. These tumors are useful for research in turmo biology and are offered to other scientists seeking transplantable carcinomas for experimentation.