Bone marrow transplantation across major histocompatibility barriers. V. Protection of mice from lethal graft-vs.-host disease by pretreatment of donor cells with monoclonal anti-Thy-1.2 coupled to the toxin ricin.

A new method has been devised to eliminate T cells from murine bone marrow grafts across major histocompatibility barriers and thus prevent graft-vs.-host disease (GVHD). The method utilizes a monoclonal antibody directed at the Thy-1.2 antigen but is complement independent. To make anti-Thy-1.2 toxic, the antibody is covalently linked to the toxin ricin. Ricin ordinarily binds, enters, and kills cells through receptors containing galactose. The hybrid protein, anti-Thy-1.2-ricin, can enter and kill cells via the Thy-1.2 receptor. In the presence of lactose the usual entry route for ricin is largely blocked and the hybrid is shown to be a highly selective reagent that is T cell specific in its inhibition of mitogen-stimulated splenocytes. We have used a model of severe and fatal GVHD where BALB/c splenocytes and bone marrow cells are given to irradiated C57BL/6 recipients. Over 90% of these mice die by day 70, exhibiting signs of GVHD. When donor cells are pretreated with 0.5 microgram/ml of anti-Thy-1.2-ricin plus 200 mM lactose before injection, 10 of 11 animals survive through day 70 without signs of GVHD. These studies demonstrate that ricin linked to monoclonal antibodies may have utility related to the prevention of GVHD in human bone marrow transplantation.

Use of a novel colony assay to evaluate the cytotoxicity of an immunotoxin containing pokeweed antiviral protein against blast progenitor cells freshly obtained from patients with common B-lineage acute lymphoblastic leukemia.

We report a novel colony assay for B-lineage progenitor cells in acute lymphoblastic leukemia (ALL). The primary plating efficiency of blast progenitors freshly obtained from common B-lineage ALL patients varied between 0.09 and 2.63%. Morphological, cytochemical, and immunological analyses of cells from day 7 colonies provided the evidence that they are B-lineage lymphoblasts. Immunological marker analyses of cultured blasts using BA-2 (anti-CD9), BA-3 (anti-CD10), BA-1 (anti-CD24), and B43 mAb have allowed us to define two distinct immunological groups. The first group had BA-2+, BA-3+, BA-1+, B43+ marker profiles, consistent with the phenotype of uncultured bone marrow blasts. The second group differed in that the cells in the blast colonies were BA-3 (anti-CD10)-negative, although many of the cells in the bulk population were BA-3+ before culture. Blasts from both groups were able to proliferate and form secondary colonies when recultured. A pan-B immunotoxin was synthesized by linking B43, a human B cell-specific mAb, to pokeweed antiviral protein (PAP). This study showed that B43-PAP can effectively eradicate leukemic progenitor cells freshly obtained from patients with common B-lineage ALL. B43-PAP eliminated greater than 99.96% of blast progenitors under conditions in which only minimal inhibition of normal bone marrow progenitor cells (CFU-GM, CFU-E, CFU-MK, CFU-GEMM) was observed. Our results establish that the surface determinant recognized by B43 is expressed on B-lineage progenitor cells in ALL, and that these cells are sensitive to PAP at the ribosomal level. To our knowledge, B43-PAP is the first IT to prove effective against common B-lineage ALL cells.

Design and modification of EGF4KDEL 7Mut, a novel bispecific ligand-directed toxin, with decreased immunogenicity and potent anti-mesothelioma activity.

BACKGROUND: Potency, immunogenicity, and toxicity are three problems that limit the use of targeted toxins in solid tumour therapy. METHODS: To address potency, we used genetic engineering to develop a novel bispecific ligand-directed toxin (BLT) called EGF4KDEL, a novel recombinant anti-mesothelioma agent created by linking human epidermal growth factor (EGF) and interleukin-4 (IL-4) to truncated pseudomonas exotoxin (PE38) on the same single-chain molecule. Immunogenicity was reduced by mutating seven immunodominant B-cell epitopes on the PE38 molecule to create a new agent, EGF4KDEL 7Mut. RESULTS: In vitro, bispecific EGF4KDEL showed superior anti-mesothelioma activity compared with its monospecific counterparts. Toxicity in mice was diminished by having both ligands on the same molecule, allowing administration of a 10-fold greater dose of BLT than a mixture of monomeric IL4KDEL and EGFKDEL. EGF4KDEL 7Mut, retained all of its functional activity and induced about 87% fewer anti-toxin antibodies than mice given the parental, non-mutated form. In vivo, intraperitoneal (IP) injection of the BLT showed significant (P<0.01) and impressive effects against two aggressive, malignant IP mesothelioma models when treatment was begun 14-16 days post tumour innoculation. CONCLUSION: These data show that EGF4KDEL 7Mut is a promising new anti-mesothelioma agent that was developed to specifically address the obstacles facing clinical utility of targeted toxins.

Anti-T-cell reagents for human bone marrow transplantation: ricin linked to three monoclonal antibodies.

Three new reagents that react against human T cells were synthesized by covalently linking the toxin ricin to monoclonal antibodies recognizing differentiation antigens on the surface of T lymphocytes. Each of these immunotoxins selectively inhibited T-cell proliferation when the cells were incubated in the presence of lactose. Multipotent human stem cells were inhibited only at much higher concentrations. Mixtures of all three immunotoxins were more effective than any one alone. These reagents have the potential for preventing graft-versus-host disease in man.

Genetically designing a more potent antipancreatic cancer agent by simultaneously co-targeting human IL13 and EGF receptors in a mouse xenograft model.

OBJECTIVE: Investigators are currently interested in the epidermal growth factor receptor (EGFR) and interleukin 13 receptor (IL13R) as potential targets in the development of new biologicals for pancreatic cancer. Attempts to develop successful agents have met with difficulty. The novel approach used here was to target these receptors simultaneously with EGF and IL13 cloned on the same bispecific single-chain molecule with truncated diphtheria toxin (DT(390)) to determine if co-targeting with DTEGF13 had any advantages. DESIGN: Proliferation experiments were performed to measure the potency and selectivity of bispecific DTEGF13 and its monospecific counterparts against pancreatic cancer cell lines PANC-1 and MiaPaCa-2 in vitro. DTEGF13 was then administered intratumourally to nude mice with MiaPaCa-2 flank tumours to measure efficacy and toxicity (weight loss). RESULTS: In vitro, bispecific DTEGF13 was 2800-fold more toxic than monospecific DTEGF or DTIL13 against PANC-1. A similar enhancement was observed in vitro when MiaPaCa-2 pancreatic cancer cells or H2981-T3 lung adenocarcinoma cells were studied. DTEGF13 activity was blockable with recombinant EGF13. DTEGF13 was potent (IC(50) = 0.00017 nM) against MiaPaCa-2, receptor specific and significantly inhibited MiaPaCa-2 tumours in nude mice (p<0.008). CONCLUSIONS: In vitro studies show that the presence of both ligands on the same bispecific molecule is responsible for the superior activity of DTEGF13. Intratumoural administration showed that DTEGF13 was highly effective in checking aggressive tumour progression in mice. Lack of weight loss in these mice indicated that the drug was tolerated and a therapeutic index exists in an "on target" model in which DTEGF13 is cross-reactive with native mouse receptors.

CD4(+) T cells tolerized ex vivo to host alloantigen by anti-CD40 ligand (CD40L:CD154) antibody lose their graft-versus-host disease lethality capacity but retain nominal antigen responses.

A major goal of the transplant field is to tolerize donor T cells to prevent graft-versus-host disease (GVHD) (1). We describe an ex vivo approach in which the blockade of CD40 ligand (CD40L:CD154):CD40 interactions, a pathway required for optimal T cell expansion, induces donor CD4(+) T cells to become tolerant to host alloantigens (2). High doses of tolerized cells did not cause GVHD lethality in vivo. T cells had intact responses to antigens not present during tolerization. Tolerance was long lived and not readily reversible in vivo. These data have significant implications for the use of tolerization approaches to prevent human GVHD.

The critical early proinflammatory events associated with idiopathic pneumonia syndrome in irradiated murine allogeneic recipients are due to donor T cell infusion and potentiated by cyclophosphamide.

We have hypothesized that lung damage occurring in the peri-bone marrow transplant (BMT) period is critical for the subsequent generation of idiopathic pneumonia syndrome (IPS), a major complication following human BMT. The proinflammatory events induced by a common pre-BMT conditioning regimen, cyclophosphamide (Cytoxan(R)) (Cy) and total body irradiation, were analyzed in a murine BMT model. Electron microscopy indicated that Cy exacerbated irradiation-induced epithelial cell injury as early as day 3 after BMT. Allogenicity was an important contributing factor to lung injury as measured by lung wet and dry weights and decreased specific lung compliance. The most significant pulmonary dysfunction was seen in mice receiving both allogeneic T cells and Cy conditioning. IPS was associated with an influx of T cells, macrophages, and neutrophils early post-BMT. Hydroxyproline levels were not increased, indicating that the injury was not fibrotic early post-BMT. As early as 2 h after chemoradiation, host macrophages increased in number in the lung parenchyma. Continued increases in macrophages occurred if splenic T cells were administered with the donor graft. The expression of costimulatory B7 molecules correlated with macrophage numbers. Frequencies of cells expressing mRNA for the inflammatory proteins TNF-alpha, IL-1beta, and TGFbeta were increased. Cy accelerated the upregulation of TGFbeta and increase in host macrophages. The exacerbation of macrophage activation and severity of IPS was dependent on allogeneic T cells, implicating immune-mediated mechanisms as critical to the outcome of IPS. This demonstration of early injury after BMT indicates the need for very early therapeutic intervention before lung damage becomes profound and irreversible.

Differential effects of the absence of interferon-gamma and IL-4 in acute graft-versus-host disease after allogeneic bone marrow transplantation in mice.

Graft-versus-host disease (GVHD), in which immunocompetent donor cells attack the host, remains a major cause of morbidity after allogeneic bone marrow transplantation (BMT). To understand the role of cytokines in the pathobiology of GVHD, we used cytokine knockout (KO) mice as a source of donor T cells. Two different MHC-disparate strain combinations were examined: BALB/c (H2(d)) donors into lethally irradiated C57BL/6 (H2(b)) recipients or C57BL/6 (H2(b)) donors into B10.BR (H2(k)) recipients. Donor cells were from mice in which either the interferon-gamma (IFN-gamma) or the IL-4 gene was selectively disrupted to understand the role of these cytokines in acute GVHD. In both strain combinations the same pattern was noted with regard to GVHD onset and morbidity. All mice exhibited the classic signs of acute GVHD: weight loss with skin, gut, and liver pathology resulting in morbidity and mortality. Surprisingly, donor cells obtained from mice lacking IFN-gamma gave rise to accelerated morbidity from GVHD when compared with cells from wild-type control donors. Similar results were obtained using normal donors when neutralizing antibodies to IFN-gamma were administered immediately after the BMT. These results suggest that IFN-gamma plays a role in protection from acute GVHD. In marked contrast, cells obtained from IL-4 KO mice resulted in protection from GVHD compared with control donors. Splenocytes from IFN KO mice stimulated with a mitogen proliferated to a significantly greater extent and produced more IL-2 compared with splenocytes obtained from IL-4 KO or control mice. Additionally, there was increased IL-2 production in the spleens of mice undergoing GVHD using IFN-gamma KO donors. These results therefore indicate, with regard to the TH1/ TH2 cytokine paradigm, the absence of a TH1-type cytokine can be deleterious in acute GVHD, whereas absence of a TH2 cytokine can be protective.

Elimination of clonogenic T-leukemic cells from human bone marrow using anti-Mr 65,000 protein immunotoxins.

Two anti-Mr 65,000 protein (p65) murine monoclonal antibodies, T101 and VIII-1, were conjugated to intact ricin. Toxicity of the resulting immunotoxins (IT) was measured against leukemic cell lines treated alone and in the presence of excess bone marrow using a highly sensitive colony inhibition assay. Cells were pretreated with IT in the presence of lactose to block the native binding of ricin. The IT proved to be potent cytotoxins for the p65-positive cell lines, CEM and MOLT-4. Treatment with T101-ricin (1000 ng/ml) inhibited clonogenic activity of these lines by more than 5.1 logs. Less than 1 log of the inhibition at this dose was due to nonspecific killing by IT. Notably, the presence of excess bone marrow did not reduce IT toxicity against the leukemic populations. Comparison of IT concentrations which inhibited 50% of clonogenic activity showed that T101-ricin was 140- to 540-fold and VIII-1-ricin was 12- to 192-fold more toxic to p65-positive than to p65-negative cell lines. Neither unconjugated anti-p65 nor IT prepared with an irrelevant antibody inhibited clonogenic activity. Blocking of IT toxicity by unconjugated antibody further demonstrated that the antibody moiety of the IT directed the selective toxicity. We found that T101-ricin was more toxic for CEM cells than was VIII-1-ricin, even though blocking studies indicated that the two antibodies bind to proximal or identical epitopes. This report is unique in that an IT was shown to specifically eliminate greater than 99.99% of leukemic cells from human bone marrow. These findings indicate the utility of T101-ricin as an in vitro reagent for autologous bone marrow transplantation in treatment of T-cell leukemia.

Tumor sponge implantation: an in vivo method for studying syngeneic, primary antitumor lymphocyte responses.

Sponge matrices surgically implanted in the s.c. space of the back of normal BALB/c mice were injected with a "regressor" dose of Moloney virus-induced BALB/c tumor cells. The kinetics of the generation of cytotoxic cells within the sponge was studied over a 22-day period in a short-term 51Cr release assay. Cytotoxic activity peaked on Day 16 and then declined to negligible levels by Day 22. No cytotoxicity was detectable when nontransformed BALB/c blast cells, Moloney leukemia virus-induced tumor (LSTRA) cells, or unrelated chemically induced tumor (EL4) cells were used as targets. When the cellular composition of implanted tumor sponges was examined on Day 16, it was found to be 30 to 40% myeloperoxidase-positive cells, 15 to 25% surface immunoglobulin-positive cells, and 40 to 50% theta-positive cells. Treatment with anti-Thy 1.2 plus complement eliminated the cytotoxic response on Day 16. The ratio of T-cells to tumor cells within the sponge was determined by immunofluorescence. Kinetic studies showed that the number of theta-positive cells increased well before cytolytic activity was detected, possibly reflecting increasing numbers of amplifier T-cells or cytotoxic cell precursors. A later decline in theta-positive cells correlated directly with decreased cytotoxicity. Furthermore, onset of cytotoxic activity also correlated with a decline in the percentage of Moloney murine sarcoma virus tumor cells within the sponge. Sponge cells isolated on Day 16 (peak cytotoxicity), mixed with lethal dosages of moloney murine sarcoma virus tumor cells, successfully neutralized the lethal challenge demonstrating the in vivo antitumor efficacy of these effector cells. Sponges were also implanted in mice which had been immunized with single injection of Moloney murine sarcoma virus cells. Inoculation of the sponge with tumor cells resulted in a second set response in which cytotoxic cells appeared much earlier than in unsensitized animals. Cells from spleen, lymph node, or peritoneal cavity of normal or presensitized animals with tumor sponge implants were not cytotoxic, suggesting a highly localized response.

Antitumor protection from the murine T-cell leukemia/lymphoma EL4 by the continuous subcutaneous coadministration of recombinant macrophage-colony stimulating factor and interleukin-2.

Combined continuous s.c. coadministration of macrophage-colony stimulating factor (M-CSF) plus interleukin-2 (IL-2) by osmotic pump protected mice given i.v. injections of a lethal dose of EL4 T-cell leukemia/lymphoma. Antitumor protection was significantly greater than that afforded by treatment with either cytokine alone. Since neither IL-2 receptors nor M-CSF receptors were expressed on EL4, the antitumor effect was likely attributed to murine effector cells. To determine how M-CSF+IL-2 provided this effect, we performed immunophenotypic and functional analyses as well as in vivo depletion studies of putative antitumor effector cells. Splenic phenotyping experiments revealed that the highest levels of macrophages and natural killer cells were observed in mice given the cytokine combination rather than either M-CSF or IL-2 alone. In vivo depletion of natural killer cells ablated the antitumor protective effect of M-CSF and IL-2. T-cells were also important for M-CSF+IL-2 efficacy, since adult thymectomy/T-cell depletion significantly inhibited the ability of cytokine coadministration to protect against EL4. Coadministration of the 2 cytokines significantly elevated in vivo levels of CD3+CD4+, CD3+CD8+, CD3+NK1.1+ T-cells, and CD3+CD25+ (activated) T-cells, and elevated anti-EL4 cytotoxic T-cell activity measured in vitro. Although WBC counts and fluorescence-activated cell sorter studies showed that M-CSF+IL-2 treatment significantly elevated neutrophils, s.c. delivery of granulocyte-colony stimulating factor at doses sufficient to induce neutrophilia was unable to confer anti-EL4 protection. These studies indicate that macrophages, T-cells, and natural killer cells are all important in the M-CSF+IL-2 anti-EL4 response. The superior antitumor effect of this cytokine combination along with the ability of M-CSF to diminish the toxicity of IL-2 in this model suggests that further investigations into the clinical potential of this combination treatment are warranted.

Cytotoxic effects of anti-CD5 radioimmunotoxins on human tumors in vitro and in a nude mouse model.

An immunoconjugate, consisting of both toxin and radionuclide on the same antibody molecule, was synthesized by cross-linking the phytotoxin ricin to the T101 monoclonal antibody recognizing the CD5 cluster expressed on normal and malignant T-cells. The hybrid molecule was then labeled with iodine-125 by an iodine monochloride procedure. This radioimmunotoxin (RIT), which selectively bound to the CD5-positive CEM human leukemia cell line, was selectively inhibitory to antigen-positive cells in protein synthesis inhibition assays. RIT was only 3.0-7.8-fold less toxic and was 1.1-1.6-fold slower than unlabeled immunotoxin in inhibiting protein synthesis. Because of the radionuclide moiety, the RIT also provided information related to biodistribution and pharmacokinetics. Four days following intratumoral injection, more than 125-fold greater activity was found in CEM tumors implanted in nude mice as compared to normal tissues. The mean blood half-life for RIT was 25.7 h and for radiolabeled antibody, 91.3 h. Intratumoral injections of RIT selectively induced regression of established CEM tumors. To our knowledge, these studies are the first to demonstrate that a single immunoconjugate can combine the advantages of both a catalytic toxin and radionuclide for cancer therapy.

Combined immunochemotherapy of human solid tumors in nude mice.

In vivo immunochemotherapy of human solid tumors was studied in a nude mouse model. Large tumors (3 to 6 cm3) were induced by s.c. injection of the acute lymphoblastic leukemia T-cell line CEM. Transient tumor inhibition could be achieved by intratumoral injection of an intact-ricin immunotoxin that specifically recognizes a determinant CD5 (T,p67) expressed on the cell surface. Injection of the in vitro active cyclophosphamide congener mafosfamid had little effect on the progression of tumor growth. A combination regimen of immunotoxin and mafosfamid induced the most dramatic antitumor effect; a 72 to 100% reduction in tumor volume was observed within 3 to 4 days posttreatment. However, tumors relapsed within 5 to 13 days. Persistent, tumor regression was observed only when protocols using successive injections of combined immunotoxin/mafosfamid were used. All seven mice undergoing this treatment had a precipitous decrease in tumor size, and 86% survived greater than 30 days posttreatment. No residual tumor was detectable on Day 30 in five of seven mice. Regression was partly attributed to the selective activity of immunotoxin, since successive injections of an irrelevant control immunotoxin coupled to ricin in combination with mafosfamid did not reduce tumor size. Thus, we have demonstrated that a combination of anticancer chemotherapy and immunotoxin therapy yielded a greater antitumor effect than either therapy alone.

Antitumor activity of L6-ricin immunotoxin against the H2981-T3 lung adenocarcinoma cell line in vitro and in vivo.

The monoclonal antibody L6 recognizes a determinant that is expressed on lung, breast, colon, and ovarian carcinomas and is present only at trace levels in normal tissues. L6 was covalently linked to intact ricin by a thioether bond to produce an immunotoxin (IT). Gel analysis revealed that this IT was heterogeneous, but mostly one monoclonal antibody molecule linked to one ricin molecule. The L6-ricin IT selectively bound and was selectively toxic to L6-positive H2981-T3 adenocarcinoma cells in protein synthesis inhibition assays in which lactose was added to block the native ricin binding site. Clonogenic studies showed that 1 microgram/ml L6-ricin could inhibit about 99.99% of H2981-T3 growth in a limiting dilution assay, even in the presence of a 20-fold excess of human bone marrow cells. Treatment of bone marrow cells with the same dose of L6-ricin resulted in the growth of ample numbers of bone marrow progenitor cells (colony-forming units-mixed, colony-forming units granulocyte/macrophage, and blast-forming units erythroid) after 14 days. We also evaluated the antitumor effect of L6-ricin administered intratumorally with lactose against established H2981-T3 tumors in a nude mouse model. Thirty % of the tumor-bearing animals responded completely to single-dose treatment, while 60% gave partial responses. The in vivo effects were not absolutely specific, since irrelevant anti-CD5 IT also induced tumor regression in this model (10% responded completely, while 30% gave partial responses). However, irrelevant IT gave higher systemic toxicity (50% mortality) than L6-ricin (23% mortality). The nonspecific activity of IT was possibly due to Fc binding, which was demonstrated in vitro, or due to ricin B-chain binding. Ricin alone was too toxic for sustained tumor protection. Unconjugated L6 had no antitumor effect. The data suggest that L6-ricin may be useful for in vitro purging of autologous bone marrow from patients with solid tumors and marrow involvement and for in vivo regional therapy of L6-positive carcinomas.

Cytotoxic effect of anti-Mr 67,000 protein immunotoxins on human tumors in a nude mouse model.

We have studied the potential use of immunotoxins (ITs) for therapeutic treatment of human tumors in an experimental model of human neoplasia. We tested intact ricin IT for its antitumor activity against established tumors. CEM, a human T-cell leukemia line expressing an Mr 67,000 cell surface antigen, and Daudi, a human B-cell lymphoma line which does not express the antigen, were found to be consistently tumorigenic in nude mice. ITs were synthesized using T101, a high-affinity monoclonal antibody reacting with the Mr 67,000 protein determinant and intact ricin. We have shown for the first time that established CEM solid tumors in nude mice will regress following intratumoral injection of T101-ricin IT, while Daudi tumors will not. Selective activity of T101-ricin is dependent on systemic i.v. administration of lactose and local intratumoral injection of the T101-ricin IT with lactose. Intact ricin ITs require the presence of lactose to block native ricin binding and render them antigen specific when linked to monoclonal antibody. Killing of target was cell specific since (a) nonspecific (irrelevant) ITs did not cause the regression of CEM tumors, and (b) injection of large amounts of free T101 antibody prior to T101-ricin IT blocked antitumor activity. Selectivity was not absolute, since regression occurred in one of six animals given irrelevant IT, and blocking was observed in two of four mice. Intratumoral IT treatment with 1 or 2 micrograms of T101-ricin IT plus lactose was not harmful to mice in contrast to intratumoral ricin treatment, which killed all treated tumor-bearing mice at a dose of 0.3 micrograms. Without i.v. injection of lactose, intratumoral injection of T101-ricin IT was also effective in eliminating established tumors. However, this treatment did not result in the selective elimination of tumor, since Daudi tumors also regressed following T101-ricin IT treatment. IT, made with ricin A chain only (T101-A chain IT), was also tested against established CEM tumors. We found that high dosages of T101-A chain IT did not destroy CEM tumors when injected intratumorally, even in the presence of activating agents such as NH4Cl or the carboxylic ionophore X-537 A. In contrast, in vitro experiments demonstrated that T101-A chain IT plus activating agents had potent and selective cytotoxic effect against CEM cells. We conclude that ITs are specifically toxic to established tumors. Although selectivity is not absolute, ITs exhibit potential as a new class of antitumor reagents.

Radiotherapy in mice with yttrium-90-labeled anti-Ly1 monoclonal antibody: therapy of the T cell lymphoma EL4.

Yttrium-90 is a potent beta-emitting radionuclide with potential for therapy of lymphoma. A monoclonal antibody against Ly1, the murine homologue of human CD5, was labeled with 90Y and found to selectively bind to Ly1-positive, radiation-sensitive, EL4 mouse lymphoma cells. When tested in this aggressive model of T cell lymphoma, in vivo studies in C57BL/6 mice showed that a single 140-microCi i.p. dose of 90Y-anti-Ly1, given 1 day after i.v. injection of a lethal dose of EL4 cells, resulted in significant but transient improvement in survival. Protection was selective, since a 90Y-labeled irrelevant control antibody did not prolong survival. Biodistribution studies showed that protection was likely limited by inadequate localization of labeled antibody to tumor. Importantly, labeled anti-Ly1 specifically localized in the immunological tissue (spleen and thymus) and lowered the WBC count, perhaps limiting the tolerated dose. Myelosuppression, which is considered one of the major side effects associated with 90Y usage, was not a lethal complication, since WBC counts recovered in mice given a 140-microCi dose of 90Y-anti-Ly1 without EL4 cells and 100% of these animals survived. The maximum tolerated dose was less than 200 microCi. Despite the high localization of 90Y-anti-Ly1 in spleen, splenectomies of tumor-injected mice did not improve the antitumor efficacy of radiolabeled antibody. Further evidence for inadequate delivery of radionuclide to tumor was shown when external total-body irradiation was given to mice given injections of a lethal dose of EL4 tumor cells. Comparison of internal and external irradiation studies indicated that the partially protective effect of 140 microCi 90Y-Ly1 was equivalent to external radiation of only 100-200 cGy. Because this model reflects the current clinical limitations of radiolabeled antibodies for therapy, including partial antitumor efficacy, delivery of labeled anti-T cell antibodies to the immune system, and low maximum tolerated dose, the model may be useful for examining strategies which could increase the tolerated dose and therapeutic efficacy.

Radiotherapy in mice with yttrium-90-labeled anti-Ly1 monoclonal antibody: therapy of established graft-versus-host disease induced across the major histocompatibility barrier.

A monoclonal antibody recognizing Ly1, the murine homologue of CD5, was labeled with 90Y. In vivo biodistribution studies showed that 90Y-anti-Ly1 selectively localized in lymphoid tissue. Groups of B10,BR mice (H-2k) were lethally irradiated and given major histocompatibility complex-disparate C57BL/6 (H-2b) bone marrow and spleen cells to induce graft-versus-host disease (GVHD). Eight days later, mice with active GVHD were administered a single i.p. injection of 50 microCi90Y-anti-Ly1. Fifty % of these mice were alive 2 months after treatment. Long term (greater than 4-month) survival was significantly higher than in phosphate-buffered saline-treated mice. Survival was slightly improved in groups of mice receiving control irrelevant antibody labeled with 90Y or mice receiving free 90Y. However, survival in these groups was not significantly different from the phosphate-buffered saline-treated control group. The improved survival was supported by data showing improved mean animal weight. An anti-GVHD effect was confirmed by histopathological analysis. Unlabeled anti-Ly1 monoclonal antibody at comparable doses to 90Y-anti-Ly1 was not effective. Animals that died following 50-microCi treatment did not die of radiation toxicity, since all mice receiving 50 microCi 90Y-anti-Ly1 plus syngeneic bone marrow survived. The window of therapy was narrow in our studies, since 100 microCi 90Y-anti-Ly1 did not confer any survival advantage. Animals that did survive long term were studied for evidence of alloengraftment and found to have high levels of circulating donor mononuclear cells. 90Y-Anti-Ly1 localized in the spleen, thymus, liver, kidney and bone marrow but not in the bowel, lung, muscle, or skin. Animals given similar doses of free 90Y, 90Y-anti-Ly1, or labeled irrelevant antibody eliminated free 90Y fastest, followed by 90Y-anti-Ly1 and then labeled irrelevant antibody. Hematological analysis of peripheral blood from 90Y-anti-Ly1-treated mice showed reduction in total WBC counts, absolute lymphocyte numbers, and absolute neutrophil numbers on day 24 after treatment. Myelosuppression recovered by day 38. These findings indicate that Ly1-positive cells are involved in the effector phase of GVHD and that radiolabeled antibodies may be useful as cell-specific probes for studying the GVHD network. 90Y-Anti-Ly1 protected recipients long term from lethal GVHD, and the fact that it had a rather remarkable inhibitory and selective effect on the lymphoid system of mice suggests that these agents may have broader application in the field of transplantation.

Successful treatment of human acute T-cell leukemia in SCID mice using the anti-CD7-deglycosylated ricin A-chain immunotoxin DA7.

The study of new therapeutic approaches for refractory human leukemia has been hampered by the lack of relevant in vivo models with disseminated disease, particularly T acute lymphoblastic leukemia (T-ALL). In the present study we evaluated methods for establishing and therapy of a human T-ALL cell line (MT-ALL) in 73 SCID mice. MT-ALL is a T-cell receptor alpha/beta +, CD3+, and CD7+ leukemia cell line, derived from a patient with refractory disease and early death. Injection of 5 x 10(7) MT-ALL cells i.v. caused disseminated human leukemia in hematopoietic and nonhematopoietic organs in 100% of SCID mice (n = 9) leading to death or terminal disease at 65 to 70 days after a uniform clinical course. To study possible therapeutic approaches for disseminated leukemia we utilized an immunotoxin, DA7, constructed by chemically linking the mouse IgG2b anti-CD7(3A1E) monoclonal antibody which recognizes a pan-T-cell marker expressed on almost all T-cell leukemias to deglycosylated ricin A-chain, a catalytic plant toxin and inhibitor of protein synthesis. Administration of DA7 led to greater than 5 log kill of clonogenic MT-ALL cells in vitro and selectively inhibited protein synthesis. DA7 was administered to mice at a dose of 10 micrograms/mouse/day for 5 consecutive days starting 8 days after i.v. inoculation of leukemia. The immunotoxin therapy resulted in significant long term survival over 348 days compared to untreated or control mice treated with anti-CD7 antibody and deglycosylated ricin A-chain which were all dead by day 70 (P less than 0.001). Even after more than 11 months there was no evidence of disease in 82% of the DA7 treated animals. SCID mice given i.p. injections (n = 9) developed an i.p. tumor mass but demonstrated metastasis outside the peritoneum with disseminated leukemia in hematopoietic and nonhematopoietic organs, a finding different from most conventional nude mouse models. The leukemia was fatal in 100% and killed the animals at 68-95 days. SCID mice given i.p. injections of MT-ALL completely responded to therapy with DA7, resulting in survival of 100% of the animals (n = 10) at 216 days (P less than 0.001 compared to untreated animals). Anti-CD7 antibody, deglycosylated ricin A-chain, and a control anti-melanoma immunotoxin (IND1-RTA) showed no therapeutic effect. We conclude that DA7 is an effective in vivo therapeutic agent against human MT-ALL in the SCID mouse system, suggesting potential usefulness for therapy of humans with poor prognosis T-cell leukemia.

Retroviral immunotoxin gene therapy of acute myelogenous leukemia in mice using cytotoxic T cells transduced with an interleukin 4/diphtheria toxin gene.

The potential benefit of immunotoxin (IT) for cancer therapy has mostly been limited by the fact that only a small portion of injected dose ever reaches the cancer target. Thus, we set out to determine whether antigen-specific CTLs would be suitable vehicles to deliver IT to the site of cancer cells in vivo. A retroviral vector was constructed for gene therapy with (interleukin 4) IL-4 positioned downstream of its 20-amino-acid leader sequence that permitted cotranslational protein synthesis of IT along with truncated diphtheria toxin (DT). IL-4 was chosen as a ligand based on the expression of IL-4 receptor (IL-4R) on most acute myeloid leukemia cases. The first-time expression and secretion of a cytokine/DT fusion toxin was accomplished in mammalian NIH.3T3 cells, and then a retroviral vector was assembled. The retroviral IT was used to transiently transduce T15, a CD8+ CTL T cell line that specifically recognized C1498 (a lethal murine acute myeloid leukemia cell line). Transduced T15 T cells expressed intracellular DT and IL-4 as determined by immunofluorescence. Secreted IT supernatants collected from T15 had enzymatic activity and killed IL-4R+ C1498 cells, but not IL-4R- EL4 cells. Intravenous injection of transduced T15, but not nontransduced T15, into mice with s.c. tumors significantly inhibited tumor growth. In contrast, systemic therapy with a bacterial preparation of the same IL-4 IT given at its maximum tolerated dose did not protect. Retroviral IT-treated mice showed no sign of the renal or hepatic toxicity that is common to this class of IT. Together, these data indicate that retroviral IT may solve problems relating to systemic IT therapy by delivering reagent more directly to the site of cancer in vivo and may impart new anticancer defense mechanisms to antigen-specific T cells.

Heterodimeric Bispecific Single Chain Variable Fragments (scFv) Killer Engagers (BiKEs) Enhance NK-cell Activity Against CD133+ Colorectal Cancer Cells.

BACKGROUND: Natural killer (NK) cells are potent cytotoxic lymphocytes that play a critical role in tumor immunosurveillance and control. Cancer stem cells (CSC) initiate and sustain tumor cell growth, mediate drug refractory cancer relapse, and express the well-known surface marker CD133. METHODS: DNA fragments from two fully humanized single chain fragment variable (scFv) antibodies recognizing CD16 on NK-cells and CD133 on CSC were genetically spliced forming a novel drug, 16 × 133 BiKE that simultaneously recognizes these antigens to facilitate an immunologic synapse. The anti-CD133 was created using a fusion protein prepared by fusing DNA fragments encoding the two extracellular domains of CD133. Immunization of mice with the resulting fusion protein generated a unique antibody that recognized the molecular framework and was species cross-reactive. RESULTS: In vitro chromium-51 ((51)Cr) release cytotoxicity assays at both high and low effector:target ratios demonstrated the ability of the heterodimeric biological drug to greatly enhance NK-cell killing of human Caco-2 colorectal carcinoma cells known to overexpress CD133. The tumor associated antigen specificity of the drug for CD133 even enhanced NK-cell cytotoxicity against the NK-resistant human Burkitt's lymphoma Daudi cell line, which has less than 5 % CD133 surface expression. Flow cytometry analysis revealed increases in NK-cell degranulation and Interferon-γ production upon co-culture with Caco-2 targets in the presence of the drug. CONCLUSION: These studies demonstrate that the innate immune system can be effectively recruited to kill CSC using bispecific antibodies targeting CD133 and that this anti-CD133 scFv may be useful in this bispecific platform or perhaps in the design of more complex trispecific molecules for carcinoma therapy.