Natural killer inhibitory substance produced by the peritoneal cells of patients with ovarian cancer.

Peritoneal cells obtained from 8 patients with minimal residual ovarian cancer produced a substance during in vitro culture that markedly inhibited the expression of natural killer (NK) cell-mediated lysis. Its molecular weight was less than 2,000, the same size as the NK-inhibiting substance (NK-IS), a similar NK-suppressive molecule produced by the peritoneal cells of rats. Human NK-IS suppressed the expression of antibody-dependent cell cytotoxicity as well as NK lysis, but it had no effect on erythrocyte-rosette formation and was not cytotoxic to peripheral blood lymphocytes or cell fractions enriched for large granular lymphocytes. NK-IS inhibited lysis mediated by interferon-activated lymphocytes and completely prevented NK activation when used in a preincubation. During intraperitoneal immunotherapy with Corynebacterium parvum, an agent that can activate peritoneal cytotoxic effectors, the production of NK-IS by peritoneal cells decreased considerably. Human peritoneal cells produce an NK-IS similar to the peritoneal cells of rats, and this material may create an environment within the peritoneal cavity that is permissive to the growth of NK-sensitive tumor cells.

Leukemia-associated antigens detected by heterologous antisera.

An antiserum was produced by immunization of rabbits with the membrane fraction of a lymphoblastoid cell line, RPMI 4265. This antiserum reacted against leukemia-associated antigens on immature blast cells of 24 patients with acute leukemia (13 myeloblastic, 11 lymphoblastic). No reactivity was observed against morphologically normal blood mononuclear cells from patients in remission, cells from normal control subjects and patients with unrelated disorders, phytohemagglutinin-induced lymphoblasts, or normal bone marrow cells. Reactivity against leukemia cells was not reduced by absorption with fetal tissues. These findings were consistent with the presence of tumor-associated antigens on leukemia cells. The antigens were detectable neither during hematologic remission nor on cells from patients with unrelated diseases.

Effects of different fractions of Corynebacterium parvum on the cytotoxic T-cell response to alloantigens in mice.

The effects of several biochemically derived fractions of Corynebacterium parvum and chemically treated intact organisms on the generation of cytotoxic T-lymphocytes (CTL) were assessed in a tumor allograft model with the use of C57BL/6 and DBA/2 mice. The acid-modified, active and inactive fragmented preparations and pyridine extract and residue were all capable of inhibiting primary spleen cell allocytotoxicity. Only the active fragmented, pyridine residue and unfractionated preparations caused splenomegaly and prevented the secondary in vitro generation of CTL. Periodate treatment, acid modification, and pyridine extraction abrogated the ability of C. parvum to activate suppressor macrophages. Although incapable of activating suppressor macrophages, the phenol and pyridine extracts significantly enhanced suppressor T-cell activity. The effect of whole C. parvum organisms on T-cell-mediated cytotoxicity is thus extremely complex and due to the direct activation or inhibition of various cell types that interact with each other.

Deficiency of antibody-dependent cellular cytotoxicity and mitogen-induced cellular cytotoxicity effector cell function in patients with acute myelogenous leukemia in remission.

Patients with acute myelogenous leukemia in remission have pronounced deficiency in antibody-dependent cellular cytotoxicity (ADCC) and mitogen-induced cellular cytotoxicity. The deficiency in ADCC was partly explained by reduction in the number of circulating effector cells (Fc receptor-bearing cells) demonstrable at a time when white blood cell and platelet counts were normal. These cytotoxic functions, as well as the circulating numbers of T-cells and Fc receptor-bearing cells were further decreased by the administration of monthly cycles of combination chemotherapy with 1-beta-D-arabinofuranosylcytosine and 6-thioguanine. Following each cycle of chemotherapy, progressive recovery of these functions occurs during the third and fourth weeks with occasional increases above base line in patients in whom chemotherapy is withheld for longer than five weeks. In selected patients recovery of one cytotoxic function preceded the other, indicating that these functions are mediated by different effector cells. Administration of a single dose of daunomycin i.v. had no effect in either of these cytotoxic functions or in the circulating numbers of lymphocytes. The decrease in ADCC effector cell function induced by phase cycle-specific agents correlated with the level of reactivity exhibited by patients after achieving bone marrow and clinical remission. Patients showing low levels of reactivity postremission experienced highest degree of depression. In two patients, complete abrogation of ADCC effector function was demonstrated with minimal recovery even six weeks after stopping chemotherapy. These findings indicate that effector cells in ADCC and mitogen-induced cellular cytotoxicity are highly susceptible to phase cycle-specific agents, and their recovery takes longer that of other lymphoid and nonlymphoid populations.

Recognition by human and rabbit sera of common antigens to leukemia blast cells, peripheral blood B-lymphocytes, and monocytes.

A human serum (obtained from a multiparous and multiple-transfused patient with chronic myelogenous leukemia) and a rabbit antiserum (obtained by immunization with papain extracts from a B-lymphoblastoid cell line) showed reactivity against antigenic specificities (different from HLA) expressed on peripheral blood B-lymphocytes, unmarked lymphocytes, and monocytes. These antigenic determinants were expressed on myeloblasts and lymphoblasts from patients with acute leukemia (during the active phase of their disease) and on B-lymphoblastoid cell lines and lymphocytes from patients with chronic lymphocytic leukemia. Purified peripheral blood T-lymphocytes, mitogen (phytohemagglutinin)-activated T-lymphocytes, and lymphoblasts (with T-cell characteristics) obtained from patients with acute lymphoblastic leukemia or established lymphoblastoid cell lines lacked these antigenic specificities. Absorption experiments indicate that the antigen(s) detected on normal mononuclear cell populations, leukemia cells, and B-lymphoblastoid cell lines were either identical or highly cross-reactive.

Inhibition of natural killer cell activity by a soluble substance released by rat peritoneal cells.

We describe here a soluble substance released by nonadherent cells from the peritoneal cavity of W/Fu rats that markedly inhibits the activity of mouse, rat, and human natural killer (NK) cells. The NK-inhibiting substance (NK-IS) has low molecular weight (less than 1000), is heat resistant (100 degrees for 15 min), and is insensitive to nonspecific proteases. NK-IS is produced in the presence of indomethacin (1 to 10 micrograms/ml), suggesting it is not prostaglandin. The inhibitory effect was seen on unstimulated as well as on cells activated in vivo or in vitro by Corynebacterium parvum. The activity of cells mediated antibody-dependent cell cytotoxicity (K-cells) was also inhibited by NK-IS although to a lesser degree. In sharp contrast, the substance had little effect on lysis mediated by murine or human cytotoxic T-lymphocytes. Production of NK-IS from rat peritoneal cells was significantly greater than by spleen cells. since the peritoneal cavity is relatively deficient in base-line NK activity compared to spleen, these data suggest that NK-IS may play an in vivo role in the expression of NK cytotoxicity.

Role of inflammatory neutrophils in antitumor effects induced by intraperitoneal administration of Corynebacterium parvum in mice.

We studied the role of inflammatory neutrophils in the antitumor effects that follow i.p. injection of Corynebacterium parvum (1400 micrograms) into C3HeB/FeJ mice challenged with the murine ovarian teratocarcinoma. Peritoneal neutrophils, obtained from mice 6 hr after injection of C. parvum, exerted significant antitumor effects when injected admixed with murine ovarian terato-carcinoma cells into the peritoneal cavities of normal mice. Treatment of recipient mice with whole-body irradiation or repeated injections of silica prevented the antitumor effects, indicating that neutrophils were activating a second effector mechanism in recipient mice. Peritoneal cells obtained at 24 or 72 hr or at 7 or 11 days following C. parvum injection were considerably less effective in activation of this effector mechanism. Heat-killed C. parvum (6 hr)-induced neutrophils activated antitumor responses, but thioglycolate-induced cells were without effect. Antitumor responses in mice receiving peritoneal neutrophils were not due to simple transfer of C. parvum organisms in the inocula. These results indicate that inflammatory neutrophils, elicited into the peritoneal cavity by injection of C. parvum, play an important role in the activation of subsequent antitumor effects.

Antigens shared by leukemic blast cell and lymphoblastoid cell lines detected by lymphocyte-dependent antibody.

Lymphocyte-dependent antibodies (LDA's) directed against antigenic determinants present on lymphoblastoid cell lines as well as human leukemia blast cells were demonstrated in heterologous antisera obtained by immunizing rabbits with a membrane fraction from RPMI-4265 (a lymphoblastoid cell line derived from a patient with chronic myelogenous leukemia). LDA was present at high titers against B-lymphoblastoid, myelomonocytic, and stem cell lines. The T-lymphoblastoid cell line MOLT-4, however, did not react. LDA was demonstrated against acute myelogenous as well as lymphoblastic leukemia cells. The reactivity was not directed against phytohemagglutinin-induced blastoid antigens, fetal antigens, or fetal calf serum. Absorptions with lymphoblastoid cell lines removed all LDA reactivity. Similar results were obtained by absorbing the rabbit antiserum with acute lymphoblastic and/or acute myelogeneous leukemia cells. These findings indicate the presence of cross-reactive antigens between lymphoblastoid cell lines and leukemia cells. Furthermore, cross-reactivity between acute lymphoblastic and acute myelogenous leukemia cells was demonstrated.

Intraperitoneal administration of human recombinant interferon-alpha in patients with ovarian cancer: effects on lymphocyte phenotype and cytotoxicity.

Eleven patients with persistent Stage III ovarian cancer, documented at second look laparotomy, received i.p. human recombinant interferon-alpha (5-50 x 10(6) units/week). Prior to immunotherapy, patients' peritoneal cell lymphocytes (PCLs) contained decreased proportions of Leu-7+ and Fc-receptor+ cells and almost nondetectable natural killer (NK) and antibody-dependent cell cytotoxic (ADCC) activity. In contrast, patients' peripheral blood lymphocytes (PBLs) contained normal proportions of lymphocyte subsets and cytotoxic activity compared to control donor PBLs. During therapy, there was a concurrent increase in PCL Leu-7+ cells and NK lysis. Both peaked predictably at 24 h after each treatment, regardless of the dose injected, and usually returned to baseline by Day 7 of each weekly cycle. PCL NK enhancement was striking, usually increasing from 2-6% (effector:target ratio, 25:1) to over 30% lysis. Enhancement of PCL ADCC was less impressive. PCLs of several patients developed lytic activity towards NK-resistant Raji targets. During therapy, patients' PBLs demonstrated: (a) modestly enhanced NK lysis at Day 4 of each cycle, and; (b) no development of Raji lysis. These data clearly demonstrate the efficacy of i.p. interferon in activation of peritoneal NK activity. However, increased NK lysis did not correlate with individual tumor responses in this cohort of patients.

Synergistic effects of combination sequential immunotherapies in a murine ovarian cancer model.

The antitumor effects of Corynebacterium parvum in a murine ovarian teratocarcinoma model depend upon a sequential activation of neutrophils and macrophages within the peritoneal cavity. We studied the sequential administration of biological response modifiers that independently activate each phase of the response. Tumor-challenged mice treated by i.p. injection of a pyridine-extracted fraction of cell-free Propionibacterium acnes (PA-PE, 1400 micrograms) demonstrated prolonged survival in less than 20% of the cases. An i.p. injection of a detoxified Salmonella endotoxin (DSE) preparation (150 micrograms) had no effect on tumor outgrowth. However, i.p. treatment with PA-PE (1400 micrograms), followed by 150 micrograms of DSE 1 day later, resulted in long-term survival (greater than 100 days) in 40 to 60% of mice. This antitumor effect was only evident when PA-PE was administered first (before DSE) and optimal when DSE was administered 24 h after PA-PE. The synergistic antitumor effect could be duplicated when tumor-challenged mice were first treated i.p. with peritoneal polymorphonuclear leukocytes, elicited by injection of PA-PE, and then treated with DSE 18 h later. These data indicate that appropriately timed injection of biological response modifiers with complementary effects can result in a synergistic prevention of tumor growth.

Immunotherapy with biochemically dissociated fractions of Propionibacterium acnes in a murine ovarian cancer model.

The antitumor effect of two strains of Propionibacterium acnes (PAI and PAII) and chemically derived fractions from the whole bacterial cell were studied using a murine ovarian teratocarcinoma (MOT) model. When injected i.p. in high doses (700 to 1400 micrograms/mouse), both strains produce survival of a significant proportion of tumor-bearing mice (30 to 90%). On a weight to weight basis, however, PAI was significantly more effective than PAII. PAI and PAII were extracted using pyridine, which yielded four fractions, i.e., pyridine-extracted strains PAI and PAII (PA-PEI and PA-PEII, respectively) which are composed of the cell wall material extracted by pyridine, and the residues of PA-PEI and PA-PEII (PA-RI and PA-RII, respectively) which are composed of the residue material following the chemical extraction. The chemical composition of PA-PEI was different from that of PA-PEII (the latter had proportionately three times as many carbohydrates and one-third of the protein content of the former) and so were their antitumor properties in the MOT model. PA-PEI had markedly reduced antitumor effect when compared to the untreated cell on a per weight basis. Furthermore, curability was only seen when using a high dose (1400 micrograms/mouse). By contrast, the cell wall components extracted by pyridine from PAII (PA-PEII) had powerful antitumor effects, i.e., greater than 50% of mice given 1400-micrograms injections survived. The material contained in PA-PEII was further fractionated on the basis of its organic solubility in chloroform:methanol solvent. The water-soluble and solvent-insoluble fractions retained most of the antitumor effects of PA-PEII, while the water-insoluble and solvent-soluble fractions were only moderately effective, suggesting that the active moiety(ies) was associated with the nonlipid components of this fraction. Both residue fractions (PA-RI and PA-RII) were as effective on a per weight basis in controlling the growth of 10(5) tumor inoculum as were whole untreated cells. However, periodate oxidation of PA-RI resulted in complete loss of its antitumor effects. When surviving mice that had no evidence of tumor persistence following a tumor challenge (10(5) MOT cells) and i.p. treatment with PA were subsequently rechallenged with 10(4) tumor cells, survival was significantly prolonged, as compared to tumor-challenged (10(4) MOT) naive mice. In addition, 10 to 20% of these rechallenged mice had complete eradication of the tumor inoculum (no evidence of disease for greater than 120 days).(ABSTRACT TRUNCATED AT 400 WORDS)

Studies of lymphocyte-dependent antibodies to leukemia-associated antigens using frozen stored leukemia target cells.

The ability of frozen stored leukemia blast cells to participate in a lymphocyte-dependent antibody assay (LDA) is demonstrated. Frozen stored acute myelogenous and lymphocytic blast cells retained antigenicity and high viability (larger than or equal to 80%). High titer LDA reactivity against frozen stored leukemic blasts was demonstrated with heterologous (rabbit) and homologous (leukemic patient) antisera. Similarly, we demonstrated the ability of frozen stored peripheral blood mononuclear cells to mediate effectively the destruction of antibody-coated leukemia blast cells stored frozen in liquid nitrogen before use. These findings will facilitate the study of LDA and antibody-dependent cellular cytotoxicity in human leukemia patients using a completely autogeneic system.

Regulation of cytotoxic reactivity to minor histocompatibility antigens by administration of Corynebacterium parvum.

B10.BR(H-2k) mice were primed with H-2-identical allogeneic CBA/J(H-2k) spleen cells and restimulated in vitro 14 days later to generate specific secondary cytotoxic lymphocytes. A single intravenous injection of primed mice with 700 micrograms of Corynebacterium parvum 7 days after alloimmunization markedly inhibited the subsequent secondary in vitro generation of cytotoxic cells. In addition, regulatory spleen cells were detected in alloimmunized C-parvum-injected mice that prevented the restimulation of primed control spleen cells. Suppressive activity could not be abrogated by treating regulatory cells with anti-theta antibody and complement or by removing phagocytic cells, but it was overcome by treatment with mitomycin C. Unfractionated regulatory cells suppressed responses in an antigen nonspecific fashion. However, cells remaining after carbonyl and iron treatment (nonphagocytic) could no longer suppress responses to third party alloantigens while maintaining significant suppression of anti-CBA responses. These data suggest the generation of two distinct suppressor cell types that can control the cytotoxic response to minor histocompatibility antigens: an antigen-nonspecific phagocytic cell and an antigen specific nonphagocytic, non-theta-bearing cell.

Lymphocyte activation by recombinant interleukin-2 in ovarian cancer patients.

Peripheral blood lymphocytes from 42 patients and peritoneal cavity lymphocytes from eight patients with advanced ovarian carcinoma were tested for lymphokine-activated killer lymphocyte cytotoxicity against several ovarian carcinoma lines before and after exposure to recombinant interleukin-2 in vitro for 3-5 days. Only four of 42 (9.5%) of peripheral blood lymphocytes and zero of eight (0%) of peritoneal cavity lymphocytes had spontaneous cytotoxicity (greater than 20%) against the ovarian carcinoma lines. After in vitro exposure to recombinant interleukin-2, 41 of 42 (98%) of patients' blood lymphocytes showed a two- to fivefold increase in cytotoxicity against K562 (a lymphoblastoid human target), and 40 of 42 (95%) demonstrated lymphokine-activated killer cytotoxicity (greater than 20%) to the ovarian carcinoma lines. Lymphokine-activated killer activity against fresh allogeneic cell lines was variable, although most patients' peripheral blood lymphocytes (70%) had significant cytotoxicity. By contrast, incubation of patients' peritoneal cavity lymphocytes with recombinant interleukin-2 in vitro did not result in the generation of lymphokine-activated killer cell activity against K562 or ovarian cell lines. Peritoneal lymphocytes did produce lymphokine-activated killer cells in the presence of OK432 in half of the patients tested. The presence of autologous serum during recombinant interleukin-2 activation with blood lymphocytes had an augmentative effect on the resulting lymphokine-activated killer cytotoxicity in two of 20 patients, a suppressive effect in four of 20, and no effect in the other 14 of 20 patients tested.(ABSTRACT TRUNCATED AT 250 WORDS)

Lymphocyte cytotoxicity in the peritoneal cavity and blood of patients with ovarian cancer.

After an intensive course of combination chemotherapy, 16 patients with minimal residual ovarian cancer that was documented at second-look laparotomy, had an indwelling Tenckhoff catheter placed and underwent multiple peritoneal saline lavages. Lymphocyte-enriched populations from the peritoneal cavity and peripheral blood were obtained by density gradient centrifugation and examined for cell-surface phenotype and a variety of immune functions, including natural killer cytotoxicity and antibody-dependent cell-mediated cytotoxicity. Phenotypic characterization revealed that peritoneal lymphocytes consisted primarily of T cells and cells bearing receptors for the crystallizable fragment of immunoglobulin G (IgG) (crystallizable fragment-receptor), and contained a very low number of B cells. Peritoneal natural killer lymphocyte cytotoxicity and antibody-dependent cell-mediated cytotoxicity were very low in all but two patients. Incubation of peritoneal lymphocytes with Corynebacterium parvum and interferon in vitro did not result in augmented cytotoxicity against susceptible targets. Supernatants from cultured peritoneal cells of all patients markedly inhibited natural cytotoxic activity of normal donor blood lymphocytes. These results suggest that lymphocytes collected from the peritoneal cavity of patients with minimal residual ovarian cancer are deficient in natural and antibody-dependent cytotoxic effector function. This deficiency may influence the host's ability to control the spread and proliferation of tumor cells in the peritoneal cavity.

Cycloheximide-induced modulation of TNF-mediated cytotoxicity in sensitive and resistant ovarian tumor cells.

The mechanism of sensitivity and resistance of various ovarian carcinoma lines to recombinant tumor necrosis factor (rTNF)-mediated cytotoxicity has been investigated using a 24-h 51Cr-release assay. The cell line PA-1 is sensitive to TNF in a dose-dependent manner, whereas the cell line SKOV-3 is resistant to TNF even at high concentrations. The simultaneous addition of TNF and cycloheximide (CHX) in the assay converted the resistant SKOV-3 line into a sensitive line, but no detectable change was observed with PA-1. rTNF inhibited DNA, RNA, and protein synthesis of the sensitive PA-1 line, whereas it had no effect on SKOV-3. This finding was not due to differences in the expression of TNF receptors, as both cell lines expressed equivalent numbers of receptors. The addition of CHX to TNF resulted in suppression of DNA, RNA, and protein synthesis in both the sensitive and the resistant cell lines. Pretreatment of the cell line with TNF for 3 h and subsequent washing resulted in significant cytotoxicity of the sensitive PA-1 line and some cytotoxicity against SKOV-3. However, if the cells were pretreated with CHX for 3 h followed by rTNF for 24 h, a significant decrease in cytotoxicity was observed in both cell lines. Under these conditions, there was no significant inhibition of DNA, RNA, or protein synthesis. Pretreatment of cells for 24 h with TNF and 24 h with CHX resulted in augmentation of the cytotoxicity of PA-1 and SKOV-3, whereas pretreatment for 24 h with CHX followed by 24 h with TNF resulted in no cytotoxicity. Cells pretreated with CHX for 24 h showed poor binding of [125]I-TNF and poor internalization, whereas cells pretreated for 24 h with TNF showed marked enhancement of internalization. The sensitivity of freshly derived ovarian carcinoma lines to TNF and CHX demonstrated that TNF-resistant cells became more sensitive if treated with CHX. These results demonstrate the potential use of metabolic inhibitors in increasing the sensitivity of fresh ovarian tumor cells to TNF.

Methods for studying intestinal sensitivity and compliance: in vitro studies of balloons and a barostat.

The aim of this study was to compare in vitro various methods for recording intestinal sensitivity and compliance. Relationships between volume and pressure were determined in segments of penrose tubing and pig gut ("artificial intestine') using pressure increments of 2 mmHg (0-24 mmHg). We tested two direct methods of distension of the entire segments (by syringe inflation and the Mayo barostat); we also used three different balloon devices for indirect distension (a 10 cm polyethylene barostat bag, a 10 cm latex condom balloon and a 6 cm latex condom balloon). Maximal distending diameters of the recording systems were measured by injecting from 0 to 160 mL of air. The elastic properties of the balloons were also tested by distensions in air and in rigid tubes. All recording systems accurately detected a lesser compliance of the penrose drain as compared to pig gut. In absolute terms, only the compliance measured with a polyethylene barostat bag distended with a syringe was not different from the compliance of the segment as measured directly. The bellows of our barostat and the latex balloons had significant intrinsic compliances which interfered with the recorded pressure-volume curves. On the other hand, highly compliant plastic bags recorded most faithfully the compliance of artificial gut and that of non-compliant rigid tubes. For comparable volumes of distension, external diameters were larger with the 6 cm latex balloon than with the 10 cm latex balloon or the 10 cm polyethylene barostat balloon. A polyethylene bag distended with a non-compliant air injector (syringe) reflected most accurately the pressure-volume relationships of tubular structures. The different maximal diameters assumed by the three distending devices may explain, in part, why lower volumes of distension are required to elicit symptoms with smaller distending balloons in vivo.

Augmentation of cytotoxicity of lymphokine activated killer cells on ovarian tumor cells by various biological response modifiers.

The effect of combining lymphokine activated killer (LAK) cells with either recombinant interleukin-2 (rlL-2), recombinant interferon-alpha (rIFN-alpha), recombinant tumor necrosis factor (rTNF) or streptococcal preparation OK-432 were assessed, using Raji, 2 kinds of cultured ovarian lines (PA-1 and SKOV-3), and 7 fresh ovarian tumor lines. LAK cells were generated by culturing peripheral blood lymphocytes (PBL) with rIL-2 for 3-5 days. The simultaneous combination of LAK cells with rIFN alpha or OK-432 augmented the cytotoxicity of LAK cells. The susceptibility of tumor cells to LAK cells also increased after the pretreatment of tumor cells with OK-432. These results suggest that the simultaneous injection of LAK cells with rIFN-alpha or OK-432 and the intralesional injection of OK-432 prior to the adoptive transfer of LAK cells may be a beneficial combination treatment for LAK treatment.