Human granulocyte generation of hydroxyl radical.

Human granulocytes were capable of oxidizing 2-keto-4 thiomethylbutyric acid to ethylene during phagocytosis or membrane perturbation. The reaction required hydrogen peroxide and superoxide and in addition was inhibited by various hydroxyl radical (OH) scavengers. These observations represent direct evidence for the generation of OH by human granulocytes. Further, inhibition of ethylene generation by azide and cyanide suggests that OH generation in granulocytes may be linked to myeloperoxidase.

Human blood monocytes: stimulators of granulocyte and mononuclear colony formation in vitro.

Human blood monocytes in a feeder layer or by use for conditioning medium produced a colony-stimulating factor capable of stimulating the in vitro growth of colonies of granulocytes and mononuclear cells from human and murine marrow. Lymphocytes and neutrophils did not stimulate colony formation, and medium conditioned by neutrophils was inhibitory. This suggests that the monocyte may control granulocyte proliferation and maturation.

Red cells coated with immunoglobulin G: binding and sphering by mononuclear cells in man.

Human monocytes, macrophages, and certain lymphocytes bind firmly to red cells coated with immunoglobulin G, whether or not it is acting as antibody. Monocyte binding is specific for cells coated with immunoglobulin G and is inhibited specifically by this immunoglobulin or its Fc-fragment in solution. Although not involving serum complement and not usually a prelude to erythrophagocytosis, this binding causes rapid morphological injury to red cells, as manifested by their sphering, increased osmotic fragility, deformation, and fragmentation. It is inferred that mononuclear cells have specific surface receptors for immunoglobulin G and that these provide a critical phase of the mechanism in vivo, whereby red cells or other particles coated with antibody are apprehended and destroyed.

Human monocyte antibody-dependent cell-mediated cytotoxicity to tumor cells.

Previous investigations of mononuclear cell antibody-dependent cell-mediated cytotoxicity (ADCC) toward tumor cells suggest that K lymphocytes and not monocytes are active in this cytotoxic reaction. This report, however, demonstrates that human monocytes are able to carry out ADCC toward three different human tumor cell lines (CEM T lymphoblasts, Raji bone marrow-derived (B) lymphoblasts, and HeLa cells). The cytolytic event was found to be temperature dependent and rapid, with most of the lysis occurring in the first 4 h of incubation. The extent of lysis was directly related to the number of monocytes (effector cells) and to the degree of antibody sensitization of the target cells. The antibody-dependent cell contact-mediated nature of the cytolytic event was confirmed by inhibition with competing nonspecific monomeric immunoglobulin and by the ability of monocytes in "innocent bystander" experiments to lyse antibody-coated targets but not nonantibody-coated target cells. Evidence that monocytes were clearly the effector cells in the monocyte preparations included the observation that preincubation of effector cells with opsonized zymosan particles abolished ADCC by monocytes, but had little effect on lymphocyte ADCC. Furthermore, no evidence for Fc receptor K lymphocyte contamination of the monocyte preparations was found using antibody-coated target cells that were selectively lysed by lymphocytes but not monocytes. We suggest that ADCC toward tumor cell targets may prove to be a useful assay of monocyte function in normal and disease states.

Evidence for hydroxyl radical generation by human Monocytes.

A number of highly reactive oxygen species have been implicated in the oxygen-dependent mechanisms involved in bactericidal activity of phagocytic leukocytes. Hydrogen peroxide and superoxide, two agents known to occur during phagocytosis, are thought to interact to generate hydroxyl radical, singlet oxygen, and other potentially reactive molecules. Using an assay system of ethylene generation from methional, cell preparations of human monocytes were demonstrated to generate hydroxyl radical or a similar agent during phagocytosis of zymosan particles. The generation of ethylene was impaired by agents which reduce superoxide or hydrogen peroxide concentrations as well as by agents reported to be hydroxyl radical scavengers. The ethylene generation did not appear to be dependent on myeloperoxidase in that azide enhanced ethylene generation. Monocytes from a patient with chronic granulomatous disease failed to generate ethylene during phagocytosis. This assay technique may be useful in exploring the metabolic events integral to the bactericidal and inflammatory activity of phagocytic leukocytes.

Effects of corticosteroids on human monocyte function.

This report examined the effect of corticosteroids in vitro on human peripheral blood monocytes, essential cells in both immune and nonimmune cellular defense mechanisms. Monocyte chemotaxis in response to sera, Escherichia coli filtrate, and lymphokine chemotactic factor was markedly reduced (P < 0.01) by hydrocortisone succinate (HCS) at 16 mug/ml. Methylprednisolone succinate and unesterified hydrocortisone produced similar impairment of monocyte chemotaxis while two drugs which unmodified do not enter cells, hydrocortisone phosphate (HCP) and cortisone acetate, had no effect on chemotaxis. HCS also significantly impaired monocyte random migration at 16 mug/ml. Monocyte bactericidal activity was reduced by HCS at 16 mug/ml (P < 0.01)) but was not affected by HCP even at 120 mug/ml. In comparison, HCS did not alter granulocyte chemotaxis even at 500 mug/ml, and bactericidal activity was reduced at 16 mug/ml (P < 0.01). Monocyte phagocytosis of cryptococci was reduced only 20% (P < 0.05) at 120 mug/ml. HCS at 120 mug/ml did not alter monocyte base-line or postphagocytic hexosemonophosphate shunt activity, viability by trypan blue exclusion, adherence to tissue culture flasks, or surface binding of IgG globulin. These corticosteroid-induced defects in monocyte function may contribute to reduced cellular defense during corticosteroid therapy.

Role of hydrogen peroxide in neutrophil-mediated destruction of cultured endothelial cells.

Human neutrophils stimulated with phorbol myristate acetate were able to destroy suspensions or monolayers of cultured human endothelial cells. Neutrophil-mediated cytotoxicity was related to phorbol myristate acetate concentration, time of incubation and neutrophil number. Cytolysis was prevented by the addition of catalase, while superoxide dismutase had no effect on cytotoxicity. The addition of the heme-enzyme inhibitors, azide or cyanide, markedly stimulated neutrophil-mediated damage while exogenous myeloperoxidase failed to stimulate cytolysis. Neutrophils isolated from patients with chronic granulomatous disease did not destroy the endothelial cell targets while myeloperoxidase-deficient neutrophils successfully mediated cytotoxicity. Endothelial cell damage mediated by the myeloperoxidase deficient cells was also inhibited by catalase but not superoxide dismutase. The addition of purified myeloperoxidase to the deficient cells did not stimulate cytotoxicity. Glucose-glucose oxidase, an enzyme system capable of generating hydrogen peroxide, could replace the neutrophil as the cytotoxic mediator. The addition of myeloperoxidase at low concentrations of glucose oxidase did not increase cytolysis, but at the higher concentrations of glucose oxidase it stimulated cytotoxicity. The destruction of endothelial cells by the glucose oxidase-myeloperoxidase system was inhibited by the addition of hypochlorous acid scavengers. In contrast, neutrophil-mediated cytolysis was not effectively inhibited by the hypochlorous acid scavengers. Based on these observations, we propose that human neutrophils can destroy cultured human endothelial cells by generating cytotoxic quantities of hydrogen peroxide.

Immunologic specificity of transfer factor.

This study examined the immunologic specificity of transfer factor using a chromatographically purified transfer factor preparation. The specificity of transfer was examined utilizing immunity to keyhole limpet hemocyanin (KLH) and tuberculin. Transfer factor prepared from a donor immune to KLH successfully transferred KLH skin test reactivity to 10 out of 10 recipients. In contrast, comparable amounts of transfer factor from two donors not immune to KLH failed to transfer immunity to KLH in 11 recipients despite evidence for successful transfer of tuberculin reactivity. Unlike prior studies with a variety of antigens, the immunity to KLH in recipients of KLH immune transfer factor appeared comparable to that of the donor since both could be elicited with the same skin test antigen dose. These observations indicate that transfer factor can initiate a specific immune response to an antigen not previously encountered by the recipient and that in certain circumstances this immune response can be comparable to that of the donor. These observations on specificity and potency of transfer factor have important implications for the clinical use of this material.

Immune function of successfully treated lymphoma patients.

Immunologic function was evaluated in 12 patients with Hodgkin's disease and 5 patients with lymphocytic lymphoma who had been successfully treated with either chemotherapy, radiation therapy, or both of these modalities 3-42 mo previously. Only two of the patients were found to have total anergy to a battery of six recall skin test antigens and all were responsive to skin testing with phytohemagglutinin. However, 10 of 16 patients were unable to develop delayed cutaneous hypersensitivity to either of the neoantigens dinitrochlorobenzene or keyhole limpet hemocyanin. Four other patients developed reactivity to only one of these neoantigens for a total of 14 of 16 (88%) of the patients demonstrating some impairment in neoantigen response. Total lymphocyte, T-lymphocyte, B-lymphocyte, and null cell numbers, as well as serum immunoglobulins were quantitatively normal. Monocyte numbers, chemotaxis, and Fc receptor activity were normal. Monocyte staphylocidal activity at 60 min was modestly depressed and candidacidal activity was depressed in those receiving both chemotherapy and radiation therapy. Spontaneous (unstimulated) lymphocyte [3H]thymidine incorporation was low in the patients as a group and lymphoblastic transformation to specific antigens was impaired in 11 of 17 patients who had positive skin test reactions to the same antigen. Highly significant suppression of lymphoblastic transformation was noted after stimulation by the mitogens phytohemagglutinin, pokeweed, and concanavalin-A. The greatest impairment of mitogen response was seen in those patients receiving both chemotherapy and radiation therapy. These data demonstrate specific impairments of neoantigen processing, lymphocyte function, and to a lesser extent monocyte function in successfully treated patients with lymphoma. These impairments may contribute to the increased incidence of infections and second primary malignancies in these patients.

Mouse/human chimeric antibodies to a tumor-associated antigen: biologic activity of the four human IgG subclasses.

Variable region genes from mouse monoclonal antibody 17-1A (gamma 2a kappa) with specificity for human gastrointestinal malignancies have been paired with human immunoglobulin constant region genes (for heavy and light chains) to produce mouse/human chimeric immunoglobulin molecules (chIgG) for each of the four human IgG subclasses. Mouse 17-1A and the four chIgG bound similarly to two human colon cancer cell lines and had comparable binding affinities. The chIgG1 and chIgG3 molecules mediated lymphocyte and monocyte antibody-dependent cell-mediated cytotoxicity (ADCC) to colon cancer tumor cell lines comparable to that of the parent murine 17-1A. The chIgG2 and chIgG4 molecules were able to mediate ADCC to colon cancer cell lines but were clearly inferior to the chIgG1 and chIgG3 reagents. None of the chIgG antibodies or the murine 17-1A was able to mediate complement lysis of colon cancer cell lines. These studies demonstrate the ability to produce all four human IgG subclass chimeric molecules which retain biologic activity. We have confirmed the subclass preferences of human lymphocyte and monocyte Fc receptors for human IgG subclasses previously determined by studies with monomeric or aggregated IgG. These data may aid in the selection of chimeric antibodies for in vivo trials.

Immune response to L2 guinea pig leukemia-specific antigens.

Strain 2 guinea pigs were immunized with the LE-L2C cell line and challenged with either LE-L2C or BZ-L2C, a subline of L2C leukemia deficient in la gene product and C3 receptor. The LE-L2C-immune animals were completely protected from a challenge of 5 X 10(6) cells (100 times the lethal dose) of either cell line, and the delayed skin test responses for both lines of injected cells were equivalent. Thus established in vivo immune recognition of the leukemia antigen was similar for both cell lines. After repeated immunization of strain 2 animals with BZ-L2C cells, animals were challenged with 3 X 10(5) viable cells; 9 of 13 animals survived. A subsequent 5 x 10 (6) challenge lowered survival to 8 of the 13 animals. The skin test response to BZ-L2C developed slowly during the immunization period, but survivors of the viable cell challenge exhibited good responses to skin testing with BZ-L2C or LE-L2C leukemic cells. Thus the BZ-L2C cell line possesses a leukemia-specific antigen, but the immunogenicity of this mutant line is decreased when compared to that of the LE-L2C line.

Phase I trial of multiple large doses of murine monoclonal antibody CO17-1A. I. Clinical aspects.

Twenty-five patients with metastatic gastrointestinal adenocarcinoma received one to four infusions (400 mg) of murine monoclonal antibody CO17-1A. Eleven patients had mild gastrointestinal symptoms, and one had a transient flushing episode. Two of five who received three weekly infusions had readily reversible anaphylactic reactions at the time of the third infusion (day 15). There were no other toxic effects. One patient had a complete remission and is surviving at greater than 104 weeks, and four had stable disease. The median survival for the whole group was 57 weeks. In general, the antibody infusions were well tolerated but had modest antitumor effects.

Phase I trial of multiple large doses of murine monoclonal antibody CO17-1A. II. Pharmacokinetics and immune response.

Twenty-five patients with metastatic gastrointestinal adenocarcinoma received one to four infusions of large doses (400 mg) of murine monoclonal antibody CO17-1A (17-1A). The pharmacokinetics of 17-1A at the time of first, second, third, or fourth infusion were not statistically different; plasma half-lives were 15.0 +/- 1.7 hours (n = 5), 15.1 +/- 1.8 (n = 10), 25.3 +/- 6.2 (n = 3), and 14.4 +/- 1.8 (n = 5), respectively. Most patients had an antibody response to 17-1A, with peak levels occurring 15-22 days after infusion. The presence of serum antibody to 17-1A at the time of the second or third infusion did not significantly alter the pharmacokinetics of this large dose of antibody. Four of 25 patients failed to develop an antibody response, but this did not correlate with the amount of 17-1A administered. The administration of four doses of 400 mg over 1 week provided continuously circulating 17-1A for 10 days.

Lack of T-cell immunity in humans with preexisting anti-mouse immunoglobulin reactivity.

We have identified five normal individuals without known exposure to mouse immunoglobulin with "preexisting" human anti-mouse antibody (HAMA) against a panel of four mouse monoclonal antibodies and polyclonal mouse IgG. Competitive inhibition by polyclonal mouse IgG of HAMA binding to monoclonal antibody coated beads demonstrates the mouse constant region specificity of these sera. Lack of such inhibition by polyclonal human IgG eliminates polyclonal rheumatoid factors as an explanation. Lymphoblastic transformation studies in these five persons failed to demonstrate a memory T-cell response to the four mouse monoclonals or polyclonal mouse IgG. Positive controls included T-cell responses to streptolysin O in these five individuals as well as responses to monoclonal antibody D612 in three patients following treatment with that antibody. This lack of T-cell immunity to mouse IgG suggests that "preexisting" HAMA is the product of inadvertent cross-reactivity with murine constant region by antibodies directed against other antigens. Therefore, "preexisting" HAMA should pose no risk of anamnestic or allergic response in patients considered for murine monoclonal therapy.

Characterization of a messenger RNA polynucleotide vaccine vector.

We have constructed mRNA transcripts encoding luciferase and human carcinoembryonic antigen (CEA) which are capped, polyadenylated, and stabilized by human beta-globin 5' and 3' untranslated regions. The mRNA construct encoding human CEA directed CEA expression in mouse fibroblasts in vitro following liposome-mediated transfection. The luciferase encoding mRNA transcripts mediated luciferase expression in vivo following i.m. injection. Based on the demonstration of protein expression in vitro and in vivo, the feasibility of using such a vector as a tumor vaccine was examined. In this pilot study, seven mice received 50 micrograms mRNA transcripts encoding CEA twice weekly for 5 weeks by i.m. injection followed by challenge with syngeneic, CEA-expressing tumor cells. This dose and schedule "primed" an immune response to CEA. Five of seven mRNA-immunized mice demonstrated anti-CEA antibody 3 weeks after tumor challenge whereas control mice had no evidence of antibody response. This strategy might be particularly useful to induce an immune response to a proto-oncogene product or growth factor which poses a risk of inducing malignant transformation consequent to prolonged protein expression.

Immune response to a carcinoembryonic antigen polynucleotide vaccine.

We have constructed a DNA plasmid encoding the full length complementary DNA for human carcinoembryonic antigen (CEA) driven by the cytomegalovirus early promoter/enhancer (plasmid DNA encoding human CEA) and demonstrated that this plasmid can function as a polynucleotide vaccine. This polynucleotide vaccine induced humoral and/or cellular immune responses specific for human CEA in all 5 immunized mice. Lymphoblastic transformation data with the use of enriched T-cell populations detected the presence of CEA-specific memory T-cells in 3 of 5 mice. Lymphocytes from 2 of 5 mice had interleukin 2/interleukin 4 release in response to CEA. CEA specificity was confirmed by the absence of reactivity to a control antigen and lack of CEA reactivity among mice vaccinated with a control plasmid encoding chloramphenicol acetyltransferase. Four of 5 mice vaccinated with plasmid DNA encoding human CEA demonstrated anti-CEA antibody responses. This immune response compared favorably with a positive control group of mice immunized with vaccinia-CEA by a dose and schedule previously shown to induce immunoprotection and therapy against a human CEA expressing syngeneic murine colon carcinoma model. Studies are ongoing to establish the construct, dose, and schedule to elicit optimal CEA-specific immune response as well as immunoprotection and therapy against human CEA expressing syngeneic murine adenocarcinoma models.

Evidence for the elevation of serum carcinoembryonic antigen and tumor-associated glycoprotein-72 levels in patients administered interferons.

Sera were collected from 111 patients diagnosed with adenocarcinoma or nonadenocarcinoma malignancies who received different schedules of interferon (IFN)-gamma or IFN-beta ser alone or in combination. Serum carcinoembryonic antigen (CEA) and tumor-associated glycoprotein-72 (TAG-72) antigen levels were measured to determine whether interferon could enhance the tumor shedding and, thereby, the serum level of either tumor antigen. Less than 10% of the sera samples from patients diagnosed with nonadenocarcinoma malignancies (e.g., hairy cell leukemia, melanoma) had positive titers of TAG-72 or CEA, and interferon neither increased nor resulted in the appearance of either tumor antigen in those sera. In contrast, 59.2% and 75.4% of the patients with adenocarcinoma had positive serum levels of TAG-72 and CEA, respectively, prior to interferon. IFN-gamma and IFN-beta ser alone or in combination significantly increased serum TAG-72 or CEA in approximately 65% of those patients. The results suggest that interferon administration to patients with adenocarcinoma can result in increased serum levels of selected tumor-associated antigens used in the diagnosis of malignancy. These preliminary findings may be important in the development of new strategies to obtain more sensitive tumor antigen serum assays for the diagnosis and monitoring for disease progression of adenocarcinoma.

A phase I clinical trial of murine monoclonal antibody D612 in patients with metastatic gastrointestinal cancer.

In a phase I study, 21 patients with metastatic adenocarcinoma of the gastrointestinal tract received the murine monoclonal antibody D612. This antibody is directed at a M(r) 48,000 antigen restrictively expressed on tumors of the gastrointestinal tract and to a limited degree on normal gastrointestinal mucosa. Patients received total doses of 10-180 mg/m2 administered as single or multiple doses of 1-100 mg/m2 over an 8-day period. Dose-limiting toxicity was secretory diarrhea. A single dose of 100 mg/m2 exceeded guidelines for maximal tolerated dose. Higher total doses were achieved in subsequent patients by using repeated administration of lower doses. Three of five patients receiving 60 mg/m2 for 3 doses (180 mg/m2 total dose) experienced grade 3 diarrhea and could not complete the prescribed course. The dose of 40 mg/m2 administered on days 1, 4, and 8 (total dose, 120 mg/m2) has been selected as the dose for phase II studies. The pharmacokinetics of D612 is best described by a one-compartment model with a mean t1/2 of 48 +/- 3 h (SEM). Eighteen of 21 patients developed human anti-mouse antibody (HAMA). Patients who developed high levels of HAMA demonstrated a more rapid clearance of the day 8 dose than those who developed low levels of HAMA. In all patients studied, a component of HAMA was directed at the D612 variable region. With one exception, serum from all patients with detectable antibody to the D612 variable region demonstrated anti-paratope reactivity. Thirty-four % of known metastatic sites demonstrated uptake of radiolabeled D612. There were no objective antitumor responses in this phase I trial. The antitumor effect of D612 in vitro has been shown to be potentiated by interleukin 2 and recombinant human macrophage colony-stimulating factor. A phase II study of D612 administered in combination with cytokines that enhance human effector function is presently ongoing.

Phase II trial of murine monoclonal antibody D612 combined with recombinant human monocyte colony-stimulating factor (rhM-CSF) in patients with metastatic gastrointestinal cancer.

In a Phase II study, 14 patients with metastatic gastrointestinal cancer received the mAb D612 (40 mg/m2, days 4, 7, and 11) in combination with recombinant human monocyte colony-stimulating factor [(rhM-CSF) 80 micrograms/kg/days 1-14]. The combined treatment was well tolerated and resulted in characteristic biological activity associated with each of the agents. Thus, 10 of 14 patients experienced D612-associated secretory diarrhea, which responded to the prostaglandin inhibitor Indomethacin in 5 of 7 patients. rhM-CSF therapy was associated with peripheral monocytosis (peak absolute monocyte count, 1444 +/- 394/mm3) and thrombocytopenia (nadir count, 78 +/- 10/mm3). Monocyte surface marker analysis revealed a high baseline expression of CD16+ cells in our patient population with an additional increase with rhM-CSF therapy. We observed a correlation between the degree of thrombocytopenia and the pretreatment CD16+ monocyte count. Of the plasma cytokines assayed, serum Neopterin demonstrated the most consistent increase during rhM-CSF therapy. There was a significant difference in the half-life of the first and last dose of D612 (35.8 +/- 2 versus 27 +/- 2.9 h; P < 0.05). Eleven of fourteen patients developed low-moderate levels of anti-D612 antibody. Despite the observed biological activity of both rhM-CSF and D612 and the previously described in vitro synergy, no clinical antitumor responses were observed in this Phase II study.

Phase I clinical and pharmacological study of suppression of human antimouse antibody response to monoclonal antibody L6 by deoxyspergualin.

Development of human antimouse antibody (HAMA) is a major limiting factor in the application of murine mAb for clinical use. A novel immunomodulatory drug, deoxyspergualin (DSG), has shown potential to suppress antimouse antibody response in preclinical model systems. We conducted a Phase I trial to determine the effect of DSG on HAMA response to murine mAb L6 administered to patients with advanced cancers (in previous trials, this antibody elicited HAMA in two-thirds of the treated patients). L6 mAb was administered at a fixed dose of 200 mg/m2 on days 1-5. DSG was administered at doses of 50 mg/m2 [dose level (dl) 1] or 150 mg/m2 (dls II and III) on days 1-7. Treatment courses were repeated every 6 weeks (dls I and II) or every 3 weeks (dl III). HAMAs were quantitated by a commercially available ELISA assay (ImmuSTRIP; anti-isotypic antibodies) and a radiometric assay (antiisotypic and anti-idiotypic antibodies). Pharmacokinetics of L6 and DSG was also studied in all consenting patients. Among 24 evaluable patients, 2 patients developed detectable HAMAs using the ELISA (one each at dls I and II) after a median follow-up of 122 days (P = 0.0001 as compared to historical controls). Even in the two patients who developed HAMA, the HAMA levels were quite low (160 and 181 ng/ml; historical experience, 70-38,744 ng/ml). The radiometric assay detected anti-L6 antibodies in 13 patients (4, 6, and 3 at dls I-III, respectively) after a median of 82 days. The median highest anti-L6 antibody level was 129 ng/ml (range, 21-2150). The highest anti-L6 antibody level at dl III was only 44 ng/ml. The results suggest suppression of anti-idiotypic response also. No clinical antitumor activity was observed, and no significant changes in T4/T8 subsets or immunoglobulins occurred (suggesting a lack of generalized immunosuppression). We conclude that DSG can suppress HAMA response to L6. A starting dose of 150 mg/m2/day is recommended for Phase II trials to confirm this observation.