Citations

Target structures important for natural killer (NK) cell recognition of virally infected cells are not well defined. Since major histocompatibility complex (MHC) class I molecules bind viral peptides during acute infection, we evaluated whether an interaction between MHC and virus might influence the susceptibility of infected cells to NK cell-mediated lysis. To control for MHC class I expression on target cells, we used either HLA class I-deficient C1R cells or C1R sublines expressing transfected HLA class I gene products. Human NK cells were unable to preferentially lyse class I-deficient C1R cells after infection with herpes simplex virus (HSV). In contrast, HLA class I transfectants were significantly more susceptible to NK cell-mediated cytotoxicity after HSV infection. This occurred for HSV-infected C1R cells expressing any of the three HLA class I gene products tested (i.e., HLA-B27, HLA-A3, or HLA-Aw68), indicating that NK cell recognition in this system does not require "self" MHC and is not unique for a single haplotype. Productive HSV infection is required for the increased killing, since inoculation with UV-inactivated virus did not lead to increased lysis. In addition, since HSV infection of the transfectants did not significantly alter the level of class I expression, the change in susceptibility appears to be due to qualitative changes in the target structures on HSV-infected, HLA class I+ targets. These results demonstrate a role for MHC class I in regulating NK cell-mediated killing of virus-infected cells.

Purpose Preclinical studies of mafosfamide, a preactivated cyclophosphamide analog, were performed to define a tolerable and potentially active target concentration for intrathecal (IT) administration. A phase I and pharmacokinetic study of IT mafosfamide was performed to determine a dose for subsequent phase II trials. Patients and Methods In vitro cytotoxicity studies were performed in MCF-7, Molt-4, and rhabdomyosarcoma cell lines. Feasibility and pharmacokinetic studies were performed in nonhuman primates. These preclinical studies were followed by a phase I trial in patients with neoplastic meningitis. There were five dose levels ranging from 1 mg to 6.5 mg. Serial CSF samples were obtained for pharmacokinetic studies in a subset of patients with Ommaya reservoirs. Results The cytotoxic target exposure for mafosfamide was 10 μmol/L. Preclinical studies demonstrated that this concentration could be easily achieved in ventricular CSF after intraventricular dosing. In the phase I clinical trial, headache was the dose-limiting toxicity. Headache was ameliorated at 5 mg by prolonging the infusion rate to 20 minutes, but dose-limiting headache occurred at 6.5 mg dose with prolonged infusion. Ventricular CSF mafosfamide concentrations at 5 mg exceeded target cytotoxic concentrations after an intraventricular dose, but lumbar CSF concentrations 2 hours after the dose were less than 10 μmol/L. Therefore, a strategy to alternate dosing between the intralumbar and intraventricular routes was tested. Seven of 30 registrants who were assessable for response had a partial response, and six had stable disease. Conclusion The recommended phase II dose for IT mafosfamide, administered without concomitant analgesia, is 5 mg over 20 minutes.

Although diverse signaling events are initiated by stimulation of multichain immune recognition receptors on lymphocytes, it remains unclear as to which specific signal transduction pathways are functionally linked to granule exocytosis and cellular cytotoxicity. In the case of natural killer (NK) cells, it has been presumed that the rapid activation of protein kinase C (PKC) enables them to mediate antibody-dependent cellular cytotoxicity (ADCC) and "natural" cytotoxicity toward tumor cells. However, using cloned human NK cells, we determined here that Fc receptor stimulation triggers granule release and ADCC through a PKC-independent pathway. Specifically, pretreatment of NK cells with the selective PKC inhibitor, GF109203X (using concentrations that fully blocked phorbol myristate acetate/ionomycin-induced secretion) had no effect on FcR-initiated granule release or ADCC. In contrast, FcR ligation led to the rapid activation of phosphatidylinositol 3-kinase (PI 3-kinase), and inhibition of this enzyme with the selective inhibitor, wortmannin, blocked FcR-induced granule release and ADCC. Additional experiments showed that, whereas FcR-initiated killing was wortmannin sensitive and GF109203X insensitive, natural cytotoxic activity toward the tumor cell line K562 was wortmannin insensitive and GF109203X sensitive. Taken together, these results suggest that: (a) PI 3-kinase activation induced by FcR ligation is functionally coupled to granule exocytosis and ADCC; and (b) the signaling pathways involved in ADCC vs natural cytotoxicity are distinct.

Crosslinking of the low affinity immunoglobulin G (IgG) Fc receptor (Fc gamma R type III) on natural killer (NK) cells initiates antibody-dependent cellular cytotoxicity. During this process, Fc gamma R stimulation results in the rapid activation of phospholipase C (PLC), which hydrolyzes membrane phosphoinositides, generating inositol-1,4,5-trisphosphate and sn-1,2-diacylglycerol as second messengers. We have recently reported that PLC activation after Fc gamma R stimulation can be inhibited by a protein tyrosine kinase (PTK) inhibitor. Based on the paradigm provided by the receptor tyrosine kinases, we investigated whether PLC-gamma 1 and/or PLC-gamma 2 are expressed in NK cells, and whether the PLC-gamma isoforms are tyrosine phosphorylated in response to Fc gamma R stimulation. Immunoblotting analyses with PLC-gamma 1- and PLC-gamma 2-specific antisera demonstrate that both isoforms are expressed in human NK cells. Furthermore, Fc gamma R crosslinking triggers the tyrosine phosphorylation of both PLC-gamma 1 and PLC-gamma 2 in these cells. Phosphorylation of both isoforms is detectable within 1 min, and returns to basal level within 30 min. Pretreatment with herbimycin A, a PTK inhibitor, blocked the Fc gamma R-induced tyrosine phosphorylation of PLC-gamma 1 and PLC-gamma 2, and the subsequent release of inositol phosphates. These results suggest that Fc gamma R-initiated phosphoinositide turnover in human NK cells is regulated by the tyrosine phosphorylation of PLC-gamma. More broadly, these observations demonstrate that nonreceptor PTK(s) activated by crosslinkage of a multisubunit receptor can phosphorylate both PLC-gamma isoforms.

The Rac1 guanine nucleotide exchange factor, Vav, is activated in hematopoietic cells in response to a large variety of stimuli. The downstream signaling events derived from Vav have been primarily characterized as leading to transcription or transformation. However, we report here that Vav and Rac1 in natural killer (NK) cells regulate the development of cell-mediated killing. There is a rapid increase in Vav tyrosine phosphorylation during the development of antibody-dependent cellular cytotoxicity and natural killing. In addition, overexpression of Vav, but not of a mutant lacking exchange factor activity, enhances both forms of killing by NK cells. Furthermore, dominant-negative Rac1 inhibits the development of NK cell–mediated cytotoxicity by two mechanisms: (a) conjugate formation between NK cells and target cells is decreased; and (b) those NK cells that do form conjugates have decreased ability to polarize their granules toward the target cell. Therefore, our results suggest that in addition to participating in the regulation of transcription, Vav and Rac1 are pivotal regulators of adhesion, granule exocytosis, and cellular cytotoxicity.

Natural killer (NK) cells are named based on their natural cytotoxic activity against a variety of target cells. However, the mechanisms by which sensitive targets activate killing have been difficult to study due to the lack of a prototypic NK cell triggering receptor. Pharmacologic evidence has implicated protein tyrosine kinases (PTKs) in natural killing; however, Lck-deficient, Fyn-deficient, and ZAP-70–deficient mice do not exhibit defects in natural killing despite demonstrable defects in T cell function. This discrepancy implies the involvement of other tyrosine kinases. Here, using combined biochemical, pharmacologic, and genetic approaches, we demonstrate a central role for the PTK Syk in natural cytotoxicity. Biochemical analyses indicate that Syk is tyrosine phosphorylated after stimulation with a panel of NK-sensitive target cells. Pharmacologic exposure to piceatannol, a known Syk family kinase inhibitor, inhibits natural cytotoxicity. In addition, gene transfer of dominant-negative forms of Syk to NK cells inhibits natural cytotoxicity. Furthermore, sensitive targets that are rendered NK-resistant by major histocompatibility complex (MHC) class I transfection no longer activate Syk. These data suggest that Syk activation is an early and requisite signaling event in the development of natural cytotoxicity directed against a variety of cellular targets.

Previous pharmacologic and genetic studies have demonstrated a critical role for the low molecular weight GTP-binding protein RhoA in the regulation of cell-mediated killing by cytotoxic lymphocytes. However, a specific Rho family guanine nucleotide exchange factor (GEF) that activates this critical regulator of cellular cytotoxicity has not been identified. In this study, we provide evidence that the Rho family GEF, Vav-2, is present in cytotoxic lymphocytes, and becomes tyrosine phosphorylated after the cross-linking of activating receptors on cytotoxic lymphocytes and during the generation of cell-mediated killing. In addition, we show that overexpression of Vav-2 in cytotoxic lymphocytes enhances cellular cytotoxicity, and this enhancement requires a functional Dbl homology and Src homology 2 domain. Interestingly, the pleckstrin homology domain of Vav-2 was found to be required for enhancement of killing through some, but not all activating receptors on cytotoxic lymphocytes. Lastly, although Vav and Vav-2 share significant structural homology, only Vav is able to enhance nuclear factor of activated T cells–activator protein 1–mediated gene transcription downstream of the T cell receptor. These data demonstrate that Vav-2, a Rho family GEF, differs from Vav in the control of certain lymphocyte functions and participates in the control of cell-mediated killing by cytotoxic lymphocytes.