Antibody-directed targeting of retroviral vectors via cell surface antigens.

Targeted stable transduction of specific cells is a highly desirable goal for gene therapy applications. We report an efficient and broadly applicable approach for targeting retroviral vectors to specific cells. We find that the envelope of the alphavirus Sindbis virus can pseudotype human immunodeficiency virus type 1- and murine leukemia virus-based retroviral vectors. When modified to contain the Fc-binding domain of protein A, this envelope gives a significant enhancement in specificity in combination with antibodies specific for HLA and CD4 relative to that without antibody. Unlike previous targeting strategies for retroviral transduction, the virus titers are relatively high and stable and can be further increased by ultracentrifugation. This study provides proof of principle for a targeting strategy that would be generally useful for many gene therapy applications.

Characterization of four new monoclonal antibodies that recognize mouse natural killer activation receptors.

With the aim of identifying natural killer (NK) activation receptors, we immunized BALB/c mice with (BALB/cxB6)F1 NK LAK cells and made B-cell hybridomas. These were screened for monoclonal antibody (MAb) reacting with an NK activation receptor by using an antibody-induced redirected lysis (AIRL) assay against FcR-bearing P815 targets. Four hybridomas, clones 1C10, 1F10, 2D10 and 4G4, were selected for further characterization. Protein G-purified MAbs from these clones activated both resting and IL-2 activated B6 or F1 NK cells in the AIRL assay. 1F10 MAb, but not the other three MAbs, could compete for the binding of anti-NK1.1 (PK136) MAb to F1 NK cells. The four MAbs were screened for their ability to bind to or activate NK cells from the mouse strains SJL/J, DBA/2, 129/J, C3H/J, and BALB.K. None showed activity except IC10, which could bind to and activate SJL/J NK cells. When members of the NKR-P1 family from both B6 mice (A, B, and C genes expressed) and SJL mice (only A and B genes expressed) were expressed in Jurkat cells and tested for their antibody reactivity, PK136 MAb was found to recognize B6 NKR-P1C and SJL/J NKR-P1B; IC10 MAb was found to recognize NKR-P1-A, -B and -C from B6, but not NKR-P1A or -B from SJL/J; and 1F10 MAb was found to react only with B6 NKR-P1C.

Lentivirus vector-mediated hematopoietic stem cell gene transfer of common gamma-chain cytokine receptor in rhesus macaques.

Nonhuman primate model systems of autologous CD34+ cell transplant are the most effective means to assess the safety and capabilities of lentivirus vectors. Toward this end, we tested the efficiency of marking, gene expression, and transplant of bone marrow and peripheral blood CD34+ cells using a self-inactivating lentivirus vector (CS-Rh-MLV-E) bearing an internal murine leukemia virus long terminal repeat derived from a murine retrovirus adapted to replicate in rhesus macaques. In vitro cytokine stimulation was not required to achieve efficient transduction of CD34+ cells resulting in marking and gene expression of the reporter gene encoding enhanced green fluorescent protein (EGFP) following transplant of the CD34+ cells. Monkeys transplanted with mobilized peripheral blood CD34+ cells resulted in EGFP expression in 1 to 10% of multilineage peripheral blood cells, including red blood cells and platelets, stable for 15 months to date. The relative level of gene expression utilizing this vector is 2- to 10-fold greater than that utilizing a non-self-inactivating lentivirus vector bearing the cytomegalovirus immediate-early promoter. In contrast, in animals transplanted with autologous bone marrow CD34+ cells, multilineage EGFP expression was evident initially but diminished over time. We further tested our lentivirus vector system by demonstrating gene transfer of the human common gamma-chain cytokine receptor gene (gamma(c)), deficient in X-linked SCID patients and recently successfully used to treat disease. Marking was 0.42 and.001 HIV-1 vector DNA copy per 100 cells in two animals. To date, all EGFP- and gamma(c)-transplanted animals are healthy. This system may prove useful for expression of therapeutic genes in human hematopoietic cells.

A murine leukemia virus (MuLV) long terminal repeat derived from rhesus macaques in the context of a lentivirus vector and MuLV gag sequence results in high-level gene expression in human T lymphocytes.

We constructed human immunodeficiency virus type 1 (HIV-1) vectors that will allow higher levels of gene expression in T cells. Gene expression under the control of an internal cytomegalovirus (CMV) immediate-early promoter in a self-inactivating lentiviral vector (CSCG) is 4- to 15-fold lower in T-cell lines (SUPT1 and CEMX174) than in non-lymphoid-cell lines (HeLa and 293T). This is in contrast to a Moloney murine leukemia virus (MoMLV)-based retrovirus vector (SRalphaLEGFP). We therefore replaced the internal CMV promoter of CSCG with three different murine oncoretroviral long terminal repeat (LTR) promoters-murine sarcoma virus (MSV), MoMLV (MLV), and the LTR (termed Rh-MLV) that is derived from the ampho-mink cell focus-forming (AMP/MCF) retrovirus in the serum of one rhesus macaque monkey that developed T-cell lymphoma following autologous transplantation of enriched bone marrow stem cells transduced with a retrovirus vector preparation containing replication-competent viruses (E. F. Vanin, M. Kaloss, C. Broscius, and A. W. Nienhuis, J. Virol. 68:4241-4250, 1994). We found that the combination of Rh-MLV LTR and a partial gag sequence of MoMLV (Deltagag(871-1612)) in CS-Rh-MLV-E gave the highest level of enhanced green fluorescent protein (EGFP) gene expression compared with MLV, MSV LTR, phosphoglycerate kinase, and CMV promoters in T-cell lines, as well as activated primary T cells. Interestingly, there was a further two- to threefold increase in EGFP expression (thus, 10-fold-higher expression than with CMV) when the Rh-MLV promoter and Deltagag(871-1612) were used in a self-inactivating-vector setting that has a further deletion in the U3 region of the HIV-1 LTR. These hybrid vectors should prove useful in gene therapy applications for T cells.

Ly-49CB6 NK inhibitory receptor recognizes peptide-receptive H-2Kb.

NK-mediated cytotoxicity involves two families of receptors: activating receptors that trigger lysis of the target cells being recognized and inhibitory receptors specific primarily for MHC I on the target cell surface that can override the activating signal. MHC I molecules on the cell surface can be classified into molecules made stable by the binding of peptide with high affinity or unstable molecules potentially capable of binding high affinity peptide (hence, peptide receptive) and being converted into stable molecules. It has been previously shown that the Ly-49A inhibitory receptor recognizes stable Dd molecules. We show in this study that the inhibitory receptor Ly-49CB6 recognizes peptide-receptive Kb molecules, but does not recognize Kb molecules once they have bound high affinity peptide.

The NKR-P1B gene product is an inhibitory receptor on SJL/J NK cells.

The mouse NKR-P1 family includes at least three genes: NKR-P1A, -B, -C. Neither surface expression nor function of the NKR-P1B gene product has previously been shown. Here, we demonstrate that the SJL/J allele of the NKR-P1B gene product is expressed on SJL/J NK cells, and is recognized by PK136 mAb. Interestingly, the same mAb does not recognize the NKR-P1B gene product of C57BL/6. We have also generated a novel mAb, 1C10, that recognizes an activation receptor on SJL/J NK cells. Activation of the NKR-P1B receptor-inhibited 1C10 mAb induced redirected lysis and recruited SHP-1, indicating that NKR-P1B is an inhibitory receptor. Therefore, the mouse NKR-P1 gene family, like the Ly49 family, includes both activation and inhibitory receptors.

NK cells can recognize different forms of class I MHC.

NK recognition and lysis of targets are mediated by activation receptor(s) whose effects may be over-ridden by inhibitory receptors recognizing class I MHC on the target. Incubation of normal lymphoblasts with a peptide that can bind to their class I MHC renders them sensitive to lysis by syngeneic NK cells. By binding to class I MHC, the peptide alters or masks the target structure recognized by an inhibitory NK receptor(s). This target structure is most likely an "empty" dimer of class I heavy chain and beta2m as opposed to a "full" class I trimer formed by binding of specific peptide that is recognized by CTL.

NK cells from human MHC class I (HLA-B) transgenic mice do not mediate hybrid resistance killing against parental nontransgenic cells.

We have investigated the capacity of human MHC class I HLA-B gene products, HLA-B27, -B7 (fully human), and -B7kb (human-mouse hybrid consisting of the alpha1 and alpha2 domains of HLA-B7, and the alpha3 and cytoplasmic domains of mouse H-2Kb), expressed on mouse NK cells during ontogeny to influence NK recognition of otherwise syngeneic mouse target cells. Despite a high level of surface expression of the transgene (comparable to that of endogeneous H-2DbKb molecules), the direct killing of YAC-1 targets, and the killing of P815 targets in a redirected lysis assay, the NK effectors of these transgenic mice could not mediate hybrid resistance-like killing of nontransgenic C57BL/6 target cells either in vitro or in vivo. Splenocytes from B6-B27 mice could be used to generate CTL lines against a B27-binding peptide, implying that T cells restricted by HLA-B27 developed during ontogeny. NK cells from B6-B27 could lyse B6-B27 Con A lymphoblasts pulsed with Db-binding peptide but not B27-binding peptides. Taken together, our results show that these human HLA-B transgene products cannot function as class I MHC "self" elements for mouse NK cells, even when present throughout ontogeny.

Mouse natural killer subsets defined by their target specificity and their ability to be separately rendered unresponsive in vivo.

NK cells from F1(AxB) hybrid mice are known to reject bone marrow grafts from either parent A or parent B (hybrid resistance). Using cold target competition in an in vitro hybrid resistance assay, we demonstrate in this work that parent A and parent B are killed by different NK subsets. As confirmation of the existence of these subsets, and to determine whether they undergo a selection/education process during NK cell ontogeny, we constructed bone marrow chimeras in which NK1.1-depleted bone marrow cells from F1 mice were allowed to mature in the microenvironment of either parent A or parent B. These F1-reconstituted chimeras were shown to have normal NK cell numbers and lytic ability in terms of YAC-1 killing and Ab-mediated redirected lysis. Using a three-color flow cytometry analysis in an in vivo hybrid resistance assay, we found that A targets, but not B targets, were less effectively removed by the F1 NK cells that develop in an A recipient (F1-->A chimeras) than in normal F1 or F1-->F1 chimeric mice. IL-2-activated killer cultures of cells from these chimeras, however, did not show a significant difference in the anti-parent killing response in an in vitro assay. These results suggest that NK cells exist in subsets and that self-reactive NK subsets can be rendered unresponsive by radioresistant host element(s). However, this unresponsiveness can be broken when the NK cells are removed from the in vivo environment.

Allo-skin graft rejection, tumor rejection and natural killer activity in mice lacking p56lck.

Mice lacking the p56lck molecule (lck -/-) have a profound block in the maturation of thymocytes and a greatly reduced number of peripheral mature T cells. To analyze further the functions of the T cells developed in lck -/- mice in vivo, we evaluated the ability of lck -/- mice to reject allo-skin grafts and methylcholanthrene (MCA)-induced syngeneic fibrosarcoma, and also examined the biological activity of lck -/- natural killer (NK) cells. Mice lacking p56lck failed to reject skin grafts from either MHC-disparate or minor-histocompatibility-different donors, even after they had been primed with donor spleen cells. They also failed to reject the MCA-induced immunogenic syngeneic fibrosarcoma, MC57X. NK activity in mice lacking p56lck was normal, and there were no differences in the NK cell activation induced by poly(I).poly(C) stimulation or interleukin-2 stimulation (lymphokine-activated killer induction) between mice lacking p56lck and their immunocompetent heterozygous littermates. NK cells lacking p56lck mediated a normal antibody-dependent cell-mediated cytotoxicity (ADCC) response. The results of this study indicate that the loss of p56lck severely impairs the effectors of the immune system which mediate the rejection of allo-skin grafts and syngeneic tumors. The normal NK activity in lck -/- mice suggests that p56lck is not required for the development and activation of NK cells.

The NK1.1 antigen in NK-mediated F1 antiparent killing in vitro.

NK cells in lethally irradiated F1(A x B) hybrid mice can reject parental A or B strain bone marrow cells, a phenomenon called "hybrid resistance." The recognition mechanism used by the NK cells remains unknown. Our laboratory has previously described an in vitro model for hybrid resistance, and we have used it here to test whether the NK surface marker, NK1.1, is involved in such recognition. We found that 1) an anti-NK1.1 mAb (PK136) inhibited F1 lymphokine-activated killer (LAK) antiparent lysis if the LAK expressed NK1.1. Other mAb, even a mAb such as 2B4 that recognizes the same LAK as PK136, did not produce inhibition. 2) The F(ab')2 fragment of PK136 also inhibited lysis. 3) F1 LAK generated from athymic nude mice were as effective antiparent killers as LAK from normal mice and were equally inhibitable by anti-NK1.1 mAb, strengthening the conclusion that killing is mediated by NK cells and not T cells. 4) As previously shown by others, addition of anti-NK1.1 mAb to a mixture of NK1.1+ LAK cells and NK-resistant FcR+ cells allowed lysis of the FcR+ cells via "redirected lysis," in which the anti-NK1.1 mAb binds to NK1.1 on the NK cells and FcR on the target cell. The ability of anti-NK1.1 mAb to inhibit direct lysis and enhance redirected lysis is most consistent with NK1.1 being a receptor involved in NK activation.

Hypofibrinolysis in patients with hypercoagulability: the roles of urokinase and of plasminogen activator inhibitor.

The prevalence of abnormalities of fibrinolysis in patients with venous thromboembolism is as yet unknown. Defined abnormalities include congenital dysfunction and deficiency of plasminogen, and probably impaired plasminogen activation secondary to elevated levels of plasminogen activator inhibitor type 1 (PAI-1) or to impaired release of tissue plasminogen activator (tPA). In this preliminary study, we analyzed plasma samples from 21 patients for whom an investigation for possible thrombophilia was requested. Twenty of the patients had venous thromboembolism, and one had arterial thrombosis at an early age. Two patients had deficiency of protein C or protein S, but no other recognized biochemical disturbances related to thrombophilia were identified. Patient samples and plasma from 25 normal controls were assayed for tPA activity, PAI-1 activity, and urokinase (uPA) activity and antigen. tPA activity and antigen were not significantly different in patients than in controls. PAI-1 activity was significantly greater in patients (P < 0.0001). uPA activity was not different in the two groups. However, uPA antigen was significantly reduced in patients compared to controls (P = 0.001). These data suggest that hypofibrinolysis leading to a risk of thrombosis may be caused not only by elevated PAI-1 activity but also by reduced total uPA concentration.

Characterization of immunotrap assays for urokinase plasminogen activator and its inhibitors and measurements of these molecules in human plasma and mouse macrophage in culture.

1. Immunotrap assays that can measure the activities of urokinase-type plasminogen activator (uPA) and its inhibitors (PAIs) were characterized. 2. Both human plasma and mouse macrophages in culture were found to contain much higher inhibitor activity than uPA-like activity. 3. The balance between pro- and anti-fibrinolytic activities was quantitatively changed in the murine macrophages after the injection of thioglycollate. uPA-like materials were synthesized by the macrophages and secreted to the conditioned medium continuously, while PAI activity was unchanged during the same time period.

Modulation of the plasminogen activation system in murine macrophages.

We have dissected the state of fibrinolytic balance in the C57/BL mouse macrophages, by means of immunotrap assays and zymography. We have monitored the individual changes of plasminogen activator (PA) and plasminogen activator inhibitor (PAI) activities of cellular lysates and secretions of these macrophages, after they were stimulated by various exogenous agents. The resident peritoneal macrophages were found to have very little PA but high level of PAI, and are therefore highly anti-fibrinolytic in nature. Upon stimulation by thioglycollate, PA activity increased and PAI activity decreased, thus raising the fibrinolytic balance in these macrophages. Upon incubation of resident or thioglycollate-activated macrophages by lipopolysaccharide (LPS), the PA level was depressed while the PAI level was increased, resulting in a large drop in the total fibrinolytic balance of the activated cells. When resident or thioglycollate-activated macrophages were incubated with the anti-inflammatory agent dexamethasone, the drug depressed both the expressions of PA and PAI, in the lysate and conditioned medium of both cell types. Thus cell-bound or secreted forms of macrophage PA and PAI activities were either increased or decreased in response to thioglycollate, LPS or dexamethasone challenge. The changes in PA and PAI resulted in different state of fibrinolytic balance in macrophages, and could be related to the different functions of these macrophages at different stages of their development.