Search

The neural cell adhesion molecule L1 has been implicated in a variety of neurologic processes, including neuritogenesis and cerebellar cell migration. Here we propose an expanded role for this cell adhesion molecule in the human immune system based on its expression on cells of myelomonocytic and lymphoid origin. Freshly isolated circulating monocytes had minimal L1 expression; however, activation of these effector cells by IFN-gamma, or as a result of density gradient isolation, resulted in a significant induction of L1 expression. Constitutive L1 expression was further evident on myelomonocytic cell lines but was absent on mature tissue macrophages. In further studies, positive selection with a L1-specific Ab enriched both B cells and dendritic precursor cells from peripheral blood. Significantly, L1 expression was not detected on mature dendritic cells but could be induced by treatment with LPS, PHA, and TNF-alpha. Immunohistochemical analysis further demonstrated significant L1 expression on follicular dendritic cells and on endothelial cells associated with the arterioles and red pulp of normal spleen. Based on these observations and known functions of L1, we propose that this cell adhesion molecule may contribute to cell-cell adhesion events associated with the effector function or extravasation of these immune effector cells. In support of the latter, we present evidence that L1 can be recognized by endothelial cells via the integrin alpha(v)beta3.

mAb against GD3 ganglioside demonstrate homophilic binding in which soluble anti-GD3 mAb bind, through the GD3 binding site, to a VH idiotope (designated IdHOM) on solid phase anti-GD3 mAb. In this way, homophilic binding provides a mechanism for amplifying the binding of mAb to cell surface GD3. We show that serine 52a, within CDR2, is required for IdHOM expression, homophilic binding, and high avidity binding to cell surface GD3. Computer modeling based on the crystal structure of anti-GD3 mAb R24 showed serine 52a situated at the mouth of the GD3 binding pocket, but not directly involved with GD3 binding. Substitutions at position 52a predicted to maintain the GD3 binding pocket (e.g., threonine) resulted in the loss of IdHOM expression and homophilic binding and markedly decreased binding to cell surface GD3, but maintained low avidity GD3 binding as measured by ELISA. All other substitutions at position 52a were predicted to significantly distort the GD3 binding pocket and resulted in the loss of both homophilic binding and any detectable avidity for GD3. We have structurally defined IdHOM and conclude that this idiotope is not required for the GD3 binding pocket, but that the idiotope is necessary for homophilic binding, which is required for high avidity binding to cell surface GD3. We speculate that selection of certain VH genes may result in the expression of idiotopes that allow homophilic binding, and this may represent a general mechanism for increasing the avidity of Abs against T cell-independent Ags.

A mouse/human chimeric antibody (ch14.18) was developed that reacts with the disialoganglioside GD2 on the surface of tumor cells of neuroectodermal origin. ch14.18 has the constant regions of a human IgG1 antibody and was expressed in a murine hybridoma. This antibody was produced in tissue culture at concentrations up to 180 mg/liter of spent culture fluid. ch14.18 was characterized and compared to 14.G2a, a murine mAb against GD2 of IgG2a isotype derived from the same parental hybridoma as ch14.18. Scatchard plot analysis of data from saturation binding studies on M21 melanoma cells showed identical binding for ch14.18 and 14.G2a. Indirect immunofluorescence revealed the same staining pattern for ch14.18 and 14.G2a on different melanoma cell lines. Both antibodies were equally capable of targeting M21 xenografts in athymic nude mice. ch14.18- and 14.G2a-activated human C-mediated cytolysis of melanoma cell; however, ch14.18-mediated antibody-dependent cytotoxicity of human effector cells against melanoma cells 50- to 100-fold more efficiently than 14.G2a.

The induction of a CTL response capable of eradicating disseminated tumor metastases and the establishment of a persistent tumor-protective immunity remain major goals of cancer immunotherapy. Here, we demonstrate for the first time that the combination of interleukin 2 (IL-2) targeted to the tumor microenvironment by a recombinant antibody-IL-2 fusion protein (huKS1/4-IL-2) with gene therapy by the murine chemokine MIG (CXCL9) markedly reduced s.c. tumor burden and decisively suppressed dissemination of experimental lung metastases of CT26-KSA colon carcinoma in syngeneic BALB/c mice. This combined therapy significantly prolonged the life span of these mice 3-4-fold by concurrently delivering MIG and IL-2 to the tumor site and thereby achieving chemoattraction of T cells together with their activation. The antitumor effect obtained was mediated predominantly by MHC class I antigen-restricted CD8(+) T cells with help from MHC class II antigen-restricted CD4(+) T lymphocytes. In addition, the MIG chemokine also induced angiostatic effects in the tumor vasculature. Taken together, this combination of MIG chemokine gene therapy with tumor-targeted cytokine IL-2 provides an approach for the rational design of novel cancer immunotherapy modalities.

The huKS1/4-IL2 fusion protein, directed against the human epithelial cell adhesion molecule (huEpCAM) has been shown to induce a strong CD8+ T-cell-dependent, natural killer (NK) cell-independent, antitumor response in mice bearing the huEp-CAM-transfected CT26 colon cancer CT26-EpCAM. Here we investigate the effectiveness of huKS1/4-IL2 against CT26-Ep21.6, a subclone of CT26-EpCAM, expressing low levels of MHC class I. In vitro antibody-dependent cellular cytotoxicity (ADCC) assays in the presence of huKS1/4-IL2 demonstrate that murine NK cells from spleen and blood can kill CT26-Ep21.6 significantly better than they kill CT26-EpCAM. NK-mediated ADCC of CT26-EpCAM can be enhanced by blocking the murine NK cell-inhibitory receptor, Ly-49C. A potent in vivo antitumor effect was observed when BALB/c mice bearing experimental metastases of CT26-Ep21.6 were treated with huKS1/4-IL2. The depletion of NK cells during huKS1/4-IL2 treatment significantly reduced the antitumor effect against CT26-Ep21.6. Together our in vitro and in vivo data in the huEp-CAM-transfected CT26 models indicate that the amount of MHC class I expressed on the tumor target cell plays a critical role in the in vivo antitumor mechanism of huKS1/4-IL2 immunotherapy. A low MHC class I level favors NK cells as effectors, whereas a high level of MHC class I favors T cells as effectors. Given the heterogeneity of MHC class I expression seen in human tumors and the prevailing T-cell suppression in many cancer patients, the observation that huKS1/4-IL2 has the potential to effectively activate an NK cell-based antitumor response may be of potential clinical relevance.

Eradication of established colon carcinoma metastases is a major goal for adjuvant immunotherapy of this disease. This was accomplished in a murine model by targeting IL-2 to the tumor microenvironment with a recombinant Ab-IL-2 fusion protein (huKS1/4-IL-2). The generation of a long-lived protective immunity was demonstrated by a 10- to 14-fold increase in CTL precursor (pCTL) frequency and induction of genes encoding Th1 cytokines, followed by the generation of tumor-specific CD8+ T effector cells, some of which differentiated into long-lived T memory cells. The frequency of pCTL correlated with enhanced immune protection against tumor cell challenge, and long-lived T cell memory was maintained in syngeneic SCID mice in the absence of tumor Ag. Tumor cell challenge of these SCID mice, concomitant with a boost of two noncurative doses of huKS1/4-IL-2 fusion protein, resulted in the generation of primed CD8+ T effector cells with concurrent release of Th1 cytokines. These events culminated in the complete rejection of the tumor cell challenge and prevention of pulmonary metastases. Taken together, the data suggest that T cell memory against colon carcinoma can be maintained in the absence of Ag.

Recombinant antibody-cytokine fusion proteins are immunocytokines that achieve high cytokine concentrations in the tumor microenvironment and thereby effectively stimulate cellular immune responses against malignancies. The activation and expansion of immune effector cells, such as CD8+ T lymphocytes, by interleukin-2 immunocytokines resulted in the eradication of established pulmonary and hepatic metastases of murine melanoma and colorectal carcinoma in syngeneic mouse models. These immunocytokines were equally effective in eliminating established bone marrow and liver metastases of murine neuroblastoma by activating natural killer cells. The effective eradication of metastases by immunocytokines resulted in significant prolongation in life span of mice over that of controls receiving equivalent mixtures of antibody and interleukin-2, which failed to reduce the growth of disseminated metastases. Proof of concept was established, indicating that immunocytokine-induced activation and expansion of immune effector cells in the tumor microenvironment can effectively eradicate established tumor metastases. This promising new approach to cancer immunotherapy may lead to clinical applications that improve treatment of cancer patients with minimal residual disease in an adjuvant setting.

The antiganglioside GD2 monoclonal antibody 14G2a (Ab1) served as an immunogen to generate the anti-idiotype (anti-Id) 1A7 (IgG1,kappa), which mimics GD2 both antigenically and biologically. Anti-Id 1A7 induced anti-GD2 antibodies in mice and rabbits. In this preclinical study, a pair of cynomolgus monkeys, immunized with 1A7 that had been mixed with QS-21 adjuvant, produced anti-anti-Id antibodies (Ab3), which reacted with the GD2-positive melanoma cell line M21/P6 cells but not with GD2-negative LS174-T cells. The Ab3 shared Ids with mAb 14G2a (Ab1), as demonstrated by their ability to inhibit binding of 1A7 to this Ab1. The Ab3 bound specifically to purified GD2 antigen and competed with the Ab1 14G2a in binding to a GD2-positive melanoma cell line or to purified GD2, suggesting that Ab1 and Ab3 may bind to the same epitope and may behave as an Ab1-like antibody (Ab1'). The isotype of the GD2-specific antibodies was mostly IgG in nature. The specificity of the antibodies for GD2 was further confirmed by dot blot analysis. These antisera also specifically lysed GD2-positive target cells in an antibody-dependent cellular cytotoxicity assay. The induction of anti-GD2 responses in monkeys did not cause any apparent side effects, despite the fact that GD2 antigen is expressed by many normal tissues of these animals. Taken together, these results suggest that anti-Id 1A7 can induce GD2-specific antibodies in nonhuman primates and can thus serve as a potential network antigen for triggering active anti-GD2 antibodies in patients with GD2-positive neuroectodermal tumors.

IL-12 is a complex cytokine in both its structure and its range of biologic activities. Fusions of this heterodimeric molecule with an intact antitumor Ab were made to test the feasibility and efficacy of targeting IL-12 to tumors to elicit a local immune response. Fusion proteins composed of the human p35 and p40 subunits had IL-12 bioactivities that were nearly as potent on human immune cells as the rIL-12 standard, but were inactive on mouse cells. Hybrid IL-12 fusion proteins composed of mouse p35 and human p40, fused to Ab, were capable of inducing IFN-gamma, but were much less active on mouse spleen cells than a mouse IL-12 standard. Despite this relatively low activity, the hybrid fusion protein was as effective in a SCID mouse model as a fully active Ab-IL-2 fusion protein in eliminating established pulmonary metastases of CT26 colon carcinoma. Specific targeting of a human IL-12 fusion protein to metastatic prostate carcinoma xenografts was also shown to be effective in SCID mice transplanted with human lymphocyte-activated killer cells. These results demonstrate the importance of directing this potent cytokine to the tumor microenvironment and suggest an important alternative to systemic IL-12 administration or gene therapy for increasing its therapeutic index.

We initiated a clinical trial for patients with advanced malignant melanoma treated with an anti-idiotype antibody that mimics the disialoganglioside GD2. We report the clinical and immune responses of the first 12 patients entered into this trial. Patients received 1-, 2-, 4-, or 8-mg doses of the anti-idiotype antibody mixed with 100 microg of QS-21 adjuvant every other week, four times, and then monthly. Twelve patients have been on trial for 2-23 months, and all of them have generated immune responses. Patients were removed from the study if they demonstrated disease progression. Hyperimmune sera from all 12 patients revealed an anti-anti-idiotypic Ab3 response, as demonstrated by the inhibition of Ab2 binding to Ab1 by patients' immune sera. To further test the anti-anti-idiotypic response, patients' Ab3 antibodies were affinity purified on Sepharose 4B columns containing adsorbed immunizing anti-idiotype immunoglobulin. Purified Ab3 of all patients studied inhibited binding of Ab1 to a GD2-positive cell line. Purified Ab3 also inhibited binding of Ab1 to purified GD2, in a manner comparable to equal quantities of purified Ab1. The patient Ab3 was truly an Ab1' because it specifically bound to purified disialoganglioside GD2. The isotypic specificity of the Ab3 antibody was predominantly IgG, with only minimal IgM. The predominant IgG subclass was IgG1, with approximately equal quantities of IgG2, IgG3, and IgG4. These Ab3 antibodies reacted specifically with tumor cells expressing GD2 by immune flow cytometry and immunoperoxidase assays. Five patients' Ab3 antibodies studied for antibody-dependent cellular cytotoxicity were positive. One patient had a complete clinical response, with resolution of soft tissue disease, and six patients had stable disease, ranging from 9 to 23 months, and are being continued on vaccine therapy. Toxicity consisted of local reaction at the site of the injection, including induration and pain that generally resolved within a few days. Mild fever and chills were observed in 75% of the patients but rarely required acetaminophen. There was no additional toxicity, including abdominal pain that was previously seen with infusion of murine monoclonal anti-GD2 antibody. Current trials include patients with stage III melanoma and small cell lung cancer. Future trials will attempt to enhance the antitumor response by the addition of interleukin 2, granulocyte macrophage colony-stimulating factor, and other cytokines, together with the 1A7 vaccine.

Targeted interleukin-2 (IL-2) therapy with immunocytokines (i.e. antibody-cytokine fusion proteins) is effective in eradicating established hepatic and pulmonary metastases of melanoma in animal model systems. The effector mechanisms responsible for this antitumor effect in syngeneic, immunocompetent mice involves mainly CD8+ T cells. This was clearly indicated by immunohistochemical analyses, in vivo depletion studies and cytotoxicity tests. Such CD8+ T cells, isolated from tumor-bearing mice after immunocytokine therapy, exerted a major histocompatibility complex class I-restricted cytotoxicity against the same tumor in vitro. Because of this cellular immune response, antibody-directed IL-2 therapy can even address established metastases displaying extensive heterogeneity in the expression of the targeted antigen. The effector mechanisms induced by immunocytokines facilitate partial regressions of large subcutaneous melanoma exceeding more than 5% of the body weight. These results demonstrate the ability of immunocytokines to induce a T-cell-dependent host immune response capable of eradicating established melanoma metastases in clinically relevant organs and offers an effective, new tool for immunotherapy of malignant melanoma.

We conducted a Phase IB trial of antidisialoganglioside chimeric 14. 18 (ch14.18) antibody and interleukin 2 (IL-2) to determine the maximal tolerated dose (MTD), immunological effects, antitumor effects, and toxicity of this treatment combination. Twenty-four melanoma patients received immunotherapy with ch14.18 antibody and a continuous infusion of Roche IL-2 (1.5 x 10(6) units/m2/day) given 4 days/week for 3 weeks. The ch14.18 antibody (dose level, 2-10 mg/m2/day) was scheduled to be given for 5 days, before, during, or following initial systemic IL-2 treatment. The ch14.18 MTD was 7.5 mg/m2/day, and 15 patients were treated with the ch14.18 MTD. Immunological effects included the induction of lymphokine-activated killer activity and antibody-dependent cellular cytotoxicity by peripheral blood mononuclear cells. In addition, serum samples obtained following ch14.18 infusions were able to facilitate in vitro antibody-dependent cellular cytotoxicity. Antitumor activity included one complete response, one partial response, eight patients with stable disease, and one patient with50% decrease of hepatic metastases in the face of recurrence of a s.c. lesion. Dose-limiting toxicities were a severe allergic reaction and weakness, pericardial effusion, and decreased performance status. Most patients treated at the MTD had abdominal, chest, or extremity pain requiring i.v. morphine. One patient had an objective peripheral neuropathy. This IL-2 and ch14.18 treatment combination induces immune activation in all patients and antitumor activity in some melanoma patients. We are attempting to enhance this treatment approach by addition of the anti-GD3 R24 antibody to this IL-2 and ch14.18 regimen.

The murine monoclonal antibody (MoAb) 14.G2a recognizes GD2, a disialoganglioside expressed in tumors of neuroectodermal origin, and facilitates antibody dependent cellular cytotoxicity (ADCC) in vitro. When given in vivo, interleukin-2 (IL-2) can increase ADCC by enhancing the activity and number of circulating lymphocytes.Thirty-three pediatric patients with GD2 positive malignancies, ranging in age from 2 to 17 years (median, 9.9 years), received IL-2 and 14.G2a in this Phase I/IB study of the Children's Cancer Group (CCG) and were monitored for toxicities and response to therapy. Seven of these patients also received granulocyte-macrophage-colony stimulating factor.The maximum tolerated dose (MTD) of 14.G2a with IL-2 was 15 mg/m2/day. The most prevalent Grade 3-4 toxicities were generalized pain (n = 14 [42%]) and fever without documented infection (n = 17 [52%]). IL-2 was thought to be the causative agent in most cases of fever. Toxicities attributed to 14.G2a included pain, allergic or anaphylactic reactions, and rash. Human antimouse antibodies were demonstrated in 9 of 21 evaluated patients. One patient with neuroblastoma had a partial response, and one patient with osteosarcoma had a complete response. Immunocytology demonstrated that the number of neuroblastoma cells in bone marrow decreased in three patients.The murine MoAb 14.G2a was well tolerated at the MTD and appeared to have some antitumor activity. Further development of this approach will involve additional engineered forms of the antibody as well as testing in the adjuvant and minimal residual disease setting.

Our understanding of the role of matrix degrading proteases in cancer has dramatically expanded over the last two decades. From correlative observations linking proteases to cancer progression, we have accumulated evidence supporting a causal role for proteases in various steps of tumor progression and have become increasingly aware of the complex interactions that exist among proteases. Specific natural inhibitors of these proteases have also been identified and their role as potent cytostatic agents in cancer has been suggested. In this article some of the concepts on the role of proteases in cancer are discussed and examples of cooperation between matrix metalloproteinases and the plasmin/plasminogen activators system are presented. The role of protease inhibitors such as tissue inhibitor of metalloproteinases-2 (TIMP-2) and plasminogen activator inhibitor-2 (PAI-2) as inhibitors of tumor growth, invasion and metastasis is discussed.

Cytotoxic effector cells interact with target cells through various mechanisms. CTLs use the antigen-specific T cell receptor, whereas Fc receptor-positive natural killer cells use this receptor to interact with antibody-coated target cells. We evaluated the tumor-binding and lymphocyte-activating capability of a recombinant fusion protein consisting of a tumor-selective human/mouse chimeric anti-ganglioside GD2 antibody (ch14.18) and recombinant human interleukin-2 (IL2) (ch14.18-IL2). This fusion protein bound specifically to GD2-positive melanoma and neuroblastoma tumor cell lines, and its IL2 component stimulated in vitro proliferation of an IL2-dependent cell line, as well as peripheral blood mononuclear cells, in healthy control individuals and in cancer patients receiving continuous infusion of IL2. The IL2 presented by the fusion protein, when bound to tumor cells, induced proliferation of IL2-responsive cells as well as a comparable amount of soluble IL2 did. This suggests that localization of IL2 at the site of contact between tumor and effector cells is an effective way of presenting this cytokine to IL2-responsive cells. The ch14.18-IL2 fusion protein also mediated antibody-dependent cellular cytotoxicity with Fc receptor-positive effector cells to an extent similar to ch14.18. These results, together with those of previous studies documenting antitumor efficacy against human tumor xenografts in SCID mice and GD2-positive murine tumors in immunocompetent syngeneic mice, suggest that the ch14.18-IL2 fusion protein should be tested in Phase I and II trials in patients with GD2-positive tumors.

Boron neutron capture therapy (BNCT) is based on the nuclear capture reaction that occurs when 10B, a stable isotope, is irradiated with low energy neutrons to produce high linear energy transfer (LET) alpha particles and recoiling 7Li nuclei. In order for BNCT to be successful in treating cancer, approximately 10(9) boron atoms must be delivered per tumor cell to sustain a lethal 10B, (n,a) 7Li capture reaction. In the present study, we have produced and characterized a bispecific antibody (BsAbB8), which was reactive with both human glioma and melanoma cell lines, as well as with a variety of polyhedral borane anions (PBA). The affinity constants (KA) of BsAb-B8 with D-54 MG and M21 cells were 3.49 and 2.57 x 10(8) M-1, respectively, which were almost identical to those of the parental mAb 9.2.27 with these cell lines. In vivo tumor localizing properties were studied in nude mice bearing subcutaneous xenografts of the D-54 MG glioma. Following intravenous injection of 131I-labeled BsAb-B8, 3.4 +/- 0.2% of the injected dose/g was detected in the tumor at 24 hours, and then slowly declined to 2.0 +/- 0.4% at 96 hours compared to 1.34 +/- 0.07% and 0.03 +/- 0.01%, respectively, for normal mouse IgG. Based on the assumption that all the tumor cell antigenic receptor sites could be saturated, the following calculations have been carried out. The maximum concentration of BsAb-B8 that could be delivered to 1 g of D-54 MG glioma cells would be 99.6 micrograms, which could bind 71.7 ng of a PBA. However, since at least 500 x more boron would be required per gram of tumor to sustain a lethal 10B (n,a) 7Li capture reaction, a macromolecule containing -10(3)-10(4) boron atoms rather than a low molecular weight PBA would be required to deliver this amount. Such boron containing macromolecules have been synthesized by us, and future studies should provide information on the feasibility of using them in combination with BsAb-B8 to deliver the requisite amount of 10B.

Antibody-cytokine fusion proteins can target biologically active cytokines to various tumor sites, achieving local concentrations sufficient to induce host immune responses leading to tumor elimination. Here, we demonstrate the therapeutic efficacy of a tumor-specific antibody-lymphotoxin fusion protein (ch225-LT) on xenografted pulmonary metastases of human melanoma. In vitro studies indicated a direct cytotoxic effect of such construacts on melanoma cells via the induction of apoptosis, as demonstrated by cell cycle analysis and DNA fragmentation. However, ch225-LT lacked any therapeutic effect in immune deficient C.B17 scid/beige and scid/scid mice, indicating the insufficiency of this direct mechanism in vivo. In contrast, in athymic nu/nu mice, ch225-LT completely inhibited outgrowth of the xenografted tumor. This therapeutic effect was accompanied by infiltrations of CD45+, Mac-1+, and asialo-GM1+ cells into the tumor; B220+ cells were present in the surrounding tissue and the periphery of the tumor. The functional role of asialo-GM2+ cells was confirmed by in vivo depletion studies. Our data indicate that an antibody-lymphotoxin fusion protein effectively inhibits the growth of disseminated melanoma metastases by mechanisms that function in the absence of mature T cells, but require B, NK, and other asialo-GM1+ cells.

Doxorubicin (DOX) was conjugated to a monoclonal antibody (mAb 425) directed against the human epidermal growth factor receptor. The immunoreactivity of these conjugates, with an average of six to eight molecules of DOX per antibody, was largely conserved, and their in vitro cytotoxicity against metastatic human melanoma cells (M24 met) was improved over that of free DOX. An evaluation of the therapeutic efficacy of mAb 425-DOX indicated that this immunoconjugate suppressed the growth of primary and secondary M24 met tumors in mice with severe combined immunodeficiency and prolonged the life span of these animals, whereas an equivalent dose of free DOX was ineffective. Conjugation of DOX to an irrelevant mAb also increased its antitumor effect over that of equivalent amounts of free drug but to a lesser extent than that achieved by the mAb 425-DOX conjugate. These results demonstrate targeted delivery and striking antitumor activity of DOX immunoconjugates in a preclinical model of spontaneous, metastatic human melanoma that was insensitive to free DOX.

In this study, we examined the effect of expression of tissue inhibitor of metalloproteinases-2 (TIMP-2) on the growth and dissemination of a highly metastatic human melanoma cell line (M24net). M24net melanoma cells express a number of matrix metalloproteinases (MMPs), including gelatinase A and B (MMP-2 and MMP-9) and interstitial collagenase (MMP-1) (A. M. P. Montgomery et al., Cancer Res., 53: 693-700, 1993). The activity of these proteases was effectively down-regulated by transfecting M24net cells with complementary DNA-encoding human TIMP-2. Overexpression of TIMP-2 markedly reduced melanoma growth in the skin of immunodeficient mice but did not prevent these highly malignant cells from spontaneously metastasizing to the lungs and lymph nodes of inoculated mice. We provide a mechanism to account for the growth inhibitory property of TIMP-2 based on its ability to regulate M24net cell growth in three-dimensional interstitial collagen. In the presence of this matrix, M24net cells assume a differentiated morphology and have a reduced growth rate. We present evidence that overexpression of TIMP-2 increases the susceptibility of M24net cells to growth inhibition and morphological differentiation by occluding interstitial collagen.

Cutaneous melanoma is the most rapidly increasing malignant disease in the United States today. Currently, no effective therapeutic regimens are available, particularly after metastasis has occurred. Among solid tumors, however, malignant melanoma is considered one of the more suitable candidates for immunotherapy. One of our goals was to develop a metastatic human melanoma model in mice that mimics clinical disease and furthermore, to demonstrate that this model is useful for studying approaches to immunotherapy. We describe here the effects of antibody therapy in a model for metastasizing human melanoma, using monoclonal antibodies directed against the EGF-receptor.

Mice with severe combined immunodeficiency (scid) provide an excellent model for studying interactions between human tumor cells and effector cells of the immune system. Because these animals lack functional B and T lymphocytes, they can accept human tumor xenografts and transfer of human effector cells. Here, we determined the ability of a human melanoma-specific, cytotoxic T-cell line (CTL) in suppressing the growth of spontaneously metastasizing human melanoma cells M24 met (HLA-A11, A33) in scid mice. This CTL line was highly cytotoxic and restricted by HLA-A11 against M24 met melanoma cells in vitro but poorly cytotoxic when tested against a human melanoma cell line that did not express HLA-A11. In order to evaluate the efficacy of this CTL line against M24 met melanoma cells in vivo, randomized groups of animals were given injections of either RPMI culture medium, interleukin 2 (IL-2), CTLs, or CTLs + IL-2. IL-2, per se, did not significantly reduce tumor metastases; however, injection of melanoma-specific, HLA-A11 restricted CTLs into scid mice, 1 day postexcision of the previously induced primary tumor, markedly reduced the number of metastatic foci in the lung and decreased metastatic involvement in lymph nodes. The combination of these CTLs with IL-2 proved even more effective, since almost all lung metastases were eradicated and metastatic involvement in both axillary and inguinal lymph nodes was substantially reduced. Our results indicate that these human CTLs maintain their ability for specific killing of metastasizing melanoma cells in scid mice. Our data suggest that reconstitution of scid mice with a specific group of effector cells (step-wise scid/hu) may be helpful for in vivo evaluation of potentially useful cancer immunotherapy modalities.

Constitutive overexpression of both urokinase and matrix metalloproteinase (MMP) activity is frequently observed in individual malignant tumors. In this study we describe the combined contribution of these distinct enzyme systems to the invasive phenotype of a highly metastatic human melanoma cell line (M24met). M24met cells were found to secrete a spectrum of MMPs, including interstitial collagenase, type IV collagenases (M(r) 92,000 and 72,000 progelatinases), and stromelysin. Urokinase, but not tissue-type plasminogen activator, was detected in M24met-conditioned media and on cell surfaces. The contribution of these enzymes to extracellular matrix dissolution was determined by exploiting specific inhibitors, namely tissue inhibitor of the metalloproteinases-2 and plasminogen activator inhibitor-2. Due to the coexpression of urokinase and MMP-dependent activity, M24met cells were observed to degrade multiple components of the extracellular matrix and to significantly degrade both interstitial and basement membrane matrices. Urokinase-dependent removal of matrix glycoprotein was observed to precede MMP-dependent collagenolysis as a prerequisite rate-limiting step. We present evidence which suggests that this temporal relationship is imposed by the structural architecture of the matrix such that matrix glycoprotein serves to protect associated collagen from MMP-dependent degradation. In addition to mediating significant collagenolysis, MMP activity was further implicated in the dissolution of matrix tropoelastin. Urokinase/plasmin activity was not found to be required for MMP-zymogen activation.

A conjugate of 4-desacetylvinblastine-3-carboxyhydrazide (DAVLBHY) and the glioma-reactive monoclonal antibody (mAb) 9.2.27 induced long-term suppression of tumor growth in athymic nude mice engrafted with U87MG human glioma cells. In vitro, DAVLBHY had the strongest antiproliferative activity (inhibitory concentration at which incorporation of [3H]thymidine is at 50% of untreated control is 2.0 x 10(-9) M) of seven cytotoxic drugs tested and so was chosen for conjugation to mAb 9.2.27, which reacts specifically with the core protein of chondroitin sulfate proteoglycans found in human glioblastomas. After conjugation of DAVLBHY to the carbohydrate residues of mAb 9.2.27 it retained its full binding capacity. For in vivo studies, DAVLBHY and several conjugate derivatives were evaluated by using two dosages of i.v. injections, each starting 2 days after s.c. tumor inoculation. The control tumors reached a volume of nearly 3000 mm3 within 30 days. Tumor growth was delayed by about 20 days with four i.v. injections of 0.5 mg/kg 9.2.27-DAVLBHY, which was slightly superior to the unconjugated drug. Moreover, 9.2.27-DAVLBHY produced a highly significant suppression of growth so that the average tumor volume was only 3% of that observed in untreated controls after 28 days. Four injections of this conjugate at a larger dose, 2.0 mg/kg, prevented recurrence of the tumors for 130 days in all animals tested, thus demonstrating a significant increase in the therapeutic index, since the unconjugated drug provided limited inhibition of tumor growth for only 40 days. The specificity of the antitumor effect was demonstrated in a comparison with the control conjugate, KS1/4-DAVLBHY, which despite partial tumor suppression had only a transient effect. The specific antitumor effect of 9.2.27-DAVLBHY was unexpected, since the target antigen is expressed at a relatively low density (40,000 sites/cell) on U87MG glioma cells.

The preceding article focused on some novel approaches for the adjuvant treatment of human melanoma and neuroblastoma with mAbs against antigens preferentially expressed on these tumors. It should be emphasized that the major goal of the immunotherapy modalities described here is to apply them in an adjuvant setting for the treatment of micrometastases. The major aim is to decrease the rate of development of metastases in a setting of very low tumor burden and ultimately achieve a prolongation in life span. The combination of powerful modern technologies achieving genetic engineering of mAbs, resulting in more human-like molecules, will lead to a reevaluation of these reagents alone or in combination with molecularly defined cytokines and growth factors for the immunotherapy of cancer. The initial, albeit anectodal, findings, of phase I clinical trials mentioned in this article lead to cautious optimism that immunotherapy may find a place and will eventually contribute to the adjuvant treatment of cancer.

Human glioblastomas (five of five), the most malignant astroglial-derived tumors, specifically express a chondroitin sulfate proteoglycan that is recognized by monoclonal antibody 9.2.27 and localized to the glioma cell surface, proliferating endothelial cells, and the perivascular extracellular matrix within the tumor bed. In contrast, the expression of this proteoglycan in normal adult neocortex and white matter is limited to the smooth muscle of small arteries, while normal glia, endothelial cells, and endothelial cell basement membranes are nonreactive. Moreover, two anaplastic astrocytomas, representing medium-grade astroglial-derived tumors, fail to react with monoclonal antibody 9.2.27. In culture, glioblastoma and capillary brain endothelial cells specifically synthesize a 250-kDa core protein and a high-molecular-mass chondroitin sulfate proteoglycan, recognized by monoclonal antibody 9.2.27. These data suggest a correlation between the expression of this chondroitin sulfate proteoglycan on proliferating brain capillary endothelial cells and the malignant phenotype of astroglial cells. The prominent perivascular localization of chondroitin sulfate proteoglycan makes it a marker for both proliferating brain capillary endothelial cells and the most malignant transformed astroglial cells, thus providing an ideal target for the immunotherapy of glioblastoma.

The human melanoma cell line M24met metastasizes spontaneously from s.c. tumors to multiple distant sites in mice with severe combined immunodeficiency. Metastasis to lymph nodes and lungs is found in 100% of the animals. M24met has an undifferentiated phenotype and extra copies of the short arm of chromosome 7. This cell line expresses the epidermal growth factor receptor, and 425.3, a monoclonal antibody to the epidermal growth factor receptor, binds to 291,000 receptor molecules per M24met cell with a KD of 2.3 x 10(-10) M. This antibody has no effect on the proliferation of M24met cells under tissue culture conditions and does not mediate effector cell or complement-dependent cytotoxicity of these cells in vitro. However, treatment of established s.c. M24met tumors in mice with severe combined immunodeficiency with monoclonal antibody 425.3 specifically suppresses spontaneous metastasis of these tumors. Total doses of 4, 2, and 1 mg antibody per mouse decrease the number and size of melanoma metastases and prolong the life span of treated animals. Treatment with 4 mg of the F(ab')2 fragment of monoclonal antibody 425.3 does not influence M24met melanoma metastasis, implying a significant contribution of the Fc portion to the antimetastatic effect of this antibody.

Monoclonal antibodies (mAB) with tumor specificity are able to enhance the immunological specificity of interleukin 2 (IL-2)-activated lymphokine activated killer (LAK) cells. Antibodies may also be used to broaden the range of tumor types susceptible to immune mediated cytotoxicity by the activated LAK cells. In these studies, mAB with relative tumor specificity were used to target immunologically activated effector cells in an in vitro antibody dependent cell mediated cytotoxicity (ADCC) assay. The mAB included: 3F8 and 14.G2a, which are both specific for neuroblastoma and melanoma and recognize ganglioside GD2, and mAB ING-1, a mouse-human chimeric antibody with constant regions from human IgG1 and kappa chains and variable regions from a mouse mAB that binds to a broad range of human adenocarcinomas. Each of these mAB was able to mediate ADCC with fresh effector cells and antibody binding targets. When peripheral blood mononuclear cells were obtained from cancer patients prior to and following in vivo therapy with interleukin 2, a significant increase was noted in ADCC activity by peripheral blood mononuclear cells obtained following IL-2 therapy. Inclusion of IL-2 in the medium during the cytotoxic assay with mAB further boosted ADCC. The total activity seen was often greater than the sum of the independent LAK activity and standard ADCC activity. The cells responsible for this ADCC had the CD16+ Fc receptor. Combining IL-2 with mAB in clinical tumor therapy may lead to a wider range of tumor types being responsive to immunotherapy and may also enhance the efficacy of therapy by specifically targeting activated effector cells to tumor cells recognized by mAB. Our results provide strong support for the testing of these hypotheses in clinical trials by combining in vivo treatment with IL-2 and mAB able to mediate ADCC.

Monoclonal antibody KS1/4 recognizes an epitope expressed on the cell surface of human adenocarcinoma cells and certain epithelia. Western blotting analyses of tumor cell extracts utilizing KS1/4 reveal staining of a major Mr 40,000 band and a minor Mr 42,000 band. Both components are also detectable in KS1/4 immunoprecipitates of L-[35S]methionine- and D-[3H]glucosamine-labeled human lung tumor cell extracts. When synthesis occurs in the presence of tunicamycin or when the immunoprecipitates are treated with peptide:N-glycosidase F, a single polypeptide component (Mr 37,000) is precipitated. Immediately following translation, digestion of Mr 40,000 and Mr 42,000 glycoproteins with endo-beta-N- acetylglucosaminidase H also yields a single polypeptide component at Mr 37,000. However, over a 3-h period beginning at 10 min posttranslation, a Mr 39,000 major component and a Mr 41,000 minor component gradually appear in the endo-beta-N-acetylglucosaminidase H digests as the Mr 37,000 component gradually disappears. Analysis of tryptic glycopeptides derived from the Mr 40,000 and 42,000 components suggests that the two components differ by the addition of one extra oligosaccharide to the Mr 42,000 component. Nonequilibrium pH gradient electrophoresis/sodium dodecyl sulfate-polyacrylamide gel electrophoresis analysis of KS1/4 immunoprecipitates resolves each of the two components into multiple spots. Digestion of the KS1/4 immunoprecipitates with neuraminidase prior to two-dimensional analysis or immunoprecipitation of short pulse-labeled extracts reduces the number of spots to three each at the Mr 40,000 and Mr 42,000 positions. Digestion of the KS1/4 immunoprecipitates with peptide:N-glycosidase F, immunoprecipitation of extracts labeled in the presence of tunicamycin, or endo-beta-N-acetylglucosaminidase H digestion of immunoprecipitates of short pulse-labeled extracts prior to two-dimensional analysis results in a single series of Mr 37,000 spots, suggesting that the polypeptide portions of the Mr 40,000 and Mr 42,000 components may be identical. Endo-beta-N-acetylglucosaminidase H digestion of KS1/4 immunoprecipitates of short pulse-labeled extracts, followed by nonequilibrium pH gradient electrophoresis, V8 protease digestion, and polyacrylamide gel electrophoresis revealed an apparently identical set of polypeptides derived from each of the three Mr 37,000 spots, suggesting that the three spots derive from highly similar polypeptides.

The role of Interferon-gamma (IFN-gamma) in the immunotherapy of neuroblastoma was investigated. In vitro experiments showed that IFN-gamma augments the cytotoxicity of Natural Killer (NK) cells and of interleukin 2 (IL-2)-activated NK (LAK) cells against neuroblastoma target cells. Incubation of the neuroblastoma cells with IFN-gamma resulted in an increased susceptibility of these target cells to NK and LAK cells. Additionally, the IFN-gamma-treated neuroblastoma cells showed an increased susceptibility to the antibody-dependent cellular cytotoxicity (ADCC). In patients who have been treated with continuous infusions of IL-2, IL-2-induced secretion of IFN-gamma was detected by measuring the elevation of the 2-5 A synthetase activity in peripheral mononuclear cells or the 2-5 A oligoadenylates in the serum, although IFN-gamma itself was not detectable. From these results we conclude that IFN-gamma may play an important role in the immunotherapy of neuroblastoma in combination with IL-2 and/or with monoclonal antibodies.

Human urine was shown to be a good source for the isolation of immunologically functional HLA-A9 antigens. The use of complex solubilization procedures can be avoided since the antigens are present in soluble form and are not complexes with membrane fragments. Purification in excess of 400-fold could be achieved by the application of cellulose ion exchange chromatography, isoelectric focusing, and acrylamide gel electrophoresis. The purified HLA-A9 antigen is composed of a glycoprotein of m.w. 38,000 and β2-microglobulin, a peptide of m.w. 12,000. HLA-A9 antigens isolated from urine proved to be immunologically functional since they not only reacted specifically with anti-HLA-A9 alloantibody but also elicited anti-HLA-A9 xenoantibodies. These antibodies when covalently attached to Sepharose 4B specifically bound HLA-A9 antigens isolated from both serum and urine.

Xenoantisera to human melanoma cells and to partially purified melanoma-associated antigens were coupled to protein A-bearing Staphylococcus Aureus or protein A-Sepharose and used as immunoadsorbents for the indirect immunoprecipitation of intrinsically radiolabeled proteins released into culture medium from various cultured human tumor and nontumor cell lines. These radiolabeled immunoprecipitates when analyzed by SDS-PAGE revealed highly reproducible molecular profiles of proteins and glycoproteins released by various cultured tumor lines and control cells into their spent culture media. A comparison of molecular profiles together with data indicating the binding specificity of known xenoantisera produced against human melanoma cells or their extracts led to the discovery of 2 macromolecules that are associated with human melanoma cells; a glycoprotein with a subunit m.w. of 240,000 (240K) and a single-chain glycoprotein of 94,000 (94K) m.w. also found in association with human carcinoma cells are described in this report.

Immunization of mice with a plasma membrane-enriched fraction from human malignant melanoma cells and subsequent generation of hybridomas resulted in the isolation of an IgG1 monoclonal antibody, 155.8, that recognizes chondroitin sulfate proteoglycans. By cell binding analysis, 155.8 was shown to react with seven of eight cultured melanoma cell lines, but not with a variety of lymphoblastoid cell lines or cultured tumor cells derived from other solid tumor types. Indirect immunoprecipitation of the 155.8 antigen from intrinsically labeled melanoma cells revealed a glycoprotein of Mr = 250,000 and a sulfated molecule of Mr greater than 400,000. The antigen was identified as a chondroitin sulfate type A/C proteoglycan synthesized by melanoma cells on the basis of its sensitivity to chondroitinase ABC digestion and the identification of sulfated glycosaminoglycans released from the antigen immunoprecipitated by 155.8. The determinants recognized by antibodies 155.8 and 9.2.27, another anti-chondroitin sulfate proteoglycan, immunoprecipitate only a proteoglycan from high density cesium chloride gradient fractions, (1.487 g/liter); however, they immunoprecipitate a free glycoprotein of Mr = 250,000 from low density fractions (1.317 g/liter). This demonstrated that the 155.8 and 9.2.27 determinants, both of which reside on the glycoprotein of Mr = 250,000, are also present in the proteoglycan, suggesting that this glycoprotein is the proteoglycan core protein. Monoclonal antibody 155.8 reacts with a determinant on the core protein distinct from that recognized by 9.2.27. Proteoglycans bearing 155.8 determinants are distributed on the surface of cultured melanoma cells in a punctated fashion that apparently resolves to short, filamentous structures at high magnification. Immunohistochemical analysis demonstrated that 155.8-defined proteoglycans are found in freshly biopsied melanoma tissue, suggesting that these antigens are also synthesized in vivo by melanoma cells.

Antigens solubilized from cultured human lymphoid cells WI-L2 and RPMI 1788 were partially purified by ultracentrifugation on a KBr gradient. These antigens injected into rabbits produced xenoantibodies which after absorption with melanoma cells became specific to B cell antigens. Three such xenoantisera were submitted to the Second Histocompatibility Workshop of the Americas and reacted much like alloantisera to B cell antigens against a large panel of B peripheral lymphocytes and cells from patients with chronic lymphocytic leukemia. Xenoantisera to B cell antigens inhibited the mixed lymphocyte reaction, but did not affect the mitogenic activity of phytohemagglutinin or the functional properties of C3 receptors, monkey red blood cell receptors, or T cell receptors.

A procedure is described for purifying to homogeneity the alpha and beta chains of HLA-DR antigen from the homozygous human lymphoid cell line LG-2. The procedure involves a combination of lectin affinity as well as classical and high pressure liquid chromatography. From 100 g of packed cells, 8 mg (235 nmol) of alpha chain and 7 mg (205 nmol) of beta chain were isolated.

The shedding of 2 tumor-specific glycoprotein antigens from human melanoma cells into spent culture medium was selectively inhibited by nontoxic doses (0.5 microgram/ml) of tunicamycin, an inhibitor of N-asparagine-linked glycosylation. The inhibition of shedding of these 2 antigens with m.w. of 240K and 94K is complete within 24 hr after addition of tunicamycin. During this time interval, our data suggest that these glycosylated cell surface antigens are replaced by their nonglycosylated forms. Removal of tunicamycin or addition of N-acetyl glucosamine restores shedding of these melanoma-associated antigens with initially reduced glycosylation. This same selective inhibition of shedding was observed with cultures adapted to grow in high doses (2.5 microgram/ml) of tunicamycin that otherwise killed > 98% of the cells upon first exposure. In contrast to other glycoproteins found in spent culture medium of melanoma cells, the shedding of melanoma-associated antigens is strictly dependent on glycosylation.

In recent years extensive investigations have been directed toward determining the biologic and chemical profiles of soluble HL-A alloantigens isolated from a variety of tissue sources. During the course of such studies, the in vitro reactivity of these materials was established serologically by their ability to inhibit specifically cytotoxic alloantisera (1). The soluble HL-A antigens have been shown to have further immunologic significance as they can induce specifically blastic transformation of histoincompatible but not of isogeneic leukocytes in vitro (2). Chemical studies thus far suggest that HL-A antigens are essentially protein in nature and apparently consist of single polypeptide chains with a molecular weight of 31,000 (3). The data concerning the immunogenic properties of soluble HL-A alloantigens thus far are scanty and conflicting. Batchelor (4) reported that rabbits immunized with partially purified soluble HL-A substances produced only species-specific lymphocytotoxic antibodies.

Antibodies produced in B10.D2 mice against soluble lymphocyte membrane antigens of B10.A (H-2a) mice reacted only with lymphocytes of the strains carrying the Irk region, i.e., B10.A(2R), B10.K, B10.BR, B10.HTT, AQR, A.TE, C3H, and CBA; they did not react with cells of strains carrying different Ir regions, i.e., B10.A(4R), B10, B10.M, A.SW, DBA/1. It is therefore concluded that the antigen detected with these antibodies is apparently controlled by the Ir region of the H-2 complex. The antigen is present on some T lymphocytes and absent on B lymphocytes. Its presence or absence seems to correlate with MLC and GVH reactivity.

Human histocompatibility loci control the expression of antigenic factors which elicit an immune response resulting in the destruction of allografts. These HL-A alloantigens located on the surface of all nucleated cells thus play an important role in the graft rejection phenomenon. Recently there have also been reports linking the occurrence of certain HL-A determinants with the susceptibility to several diseases (1). In order to investigate these phenomena at the molecular level, there have been several efforts to isolate HL-A antigens in soluble form to facilitate their thorough biologic and chemical characterization. Since it is practically impossible to isolate sufficient amounts of antigens for this purpose from genetically uniform tissue sources, there has been an increased utilization of cultured human lymphocytes (2). To gain an understanding of the expression of HL-A antigenic determinants on the surface of both synchronized and non-synchronized cultured cells during the cell cycle, we applied a quantitative microabsorption technique and an indirect immunofluorescence test.

Summary Amino acid analyses were performed on relatively homogeneous guinea pig transplantation antigens, which had been liberated by exposure to sound and purified by gel filtration and by discontinuous electrophoresis on polyacrylamide gel. Comparison of the amino acid compositions of the antigens extracted from two histoincompatible lines of guinea pigs revealed significant differences in their content of serine, alanine, leucine, isoleucine and valine. These data suggest that transplantation antigens have allotypic specificities related to protein structure.

Water soluble HL-A antigens isolated from a cultured human lymphocyte cell line derived from the peripheral blood of a normal donor inhibited cytotoxic reactions of monospecific anti-HL-A antisera in a pattern identical to that found upon direct and absorption typing of the donor's peripheral lymphocytes. The possibility of obtaining large quantities of cells from mass cultures now makes the chemical and biologic characterization of HL-A antigens a reality.