Ability of cell-sized beads bearing tumor cell membrane proteins to stimulate LAK cells to secrete interferon-gamma and tumor necrosis factor-alpha

We recently reported that lymphokine activated killer (LAK) cells were stimulated to release both interferon-gamma (IFN-gamma) and tumor necrosis factor-alpha (TNF-alpha) when stimulated by a variety of tumor cells. We proposed then that the released cytokines may play a role in mediating tumor cell regression in vivo. In this paper, we provide further information on the nature of the signals, provided by the tumor cells (K562 erythroleukemia), that stimulate LAK cells to secrete IFN-gamma and TNF-alpha. Using a previously published protocol for coating tumor-membrane molecules onto cell-sized hydrophobic beads (also called pseudocytes), we demonstrate that the signal provided by the tumor cell is membrane associated. Beads coated with K562 membranes stimulated LAK cells to release IFN-gamma and TNF-alpha. The pretreatment of these beads with trypsin and sodium periodate eliminated the ability of these pseudocytes to stimulate cytokine release in LAK cells. The glycoproteins that stimulate LAK cells to secrete IFN-gamma and TNF-alpha were further enriched by their ability to bind concanavalin A (Con A, Jack Bean). To determine if the tumor-associated molecules that stimulate LAK cells to release IFN-gamma and TNF-alpha are also the molecules involved in mediating tumor cell lysis, we tested the ability of the Con A binding and nonbinding proteins to inhibit the LAK cell-mediated lysis of K562 cells. Our results demonstrate that molecules that inhibited LAK cell-mediated cytotoxicity were not enriched by Con A. These results are therefore consistent with the conclusion that different sets of tumor-associated molecules are involved in the stimulation of LAK cells to secrete cytokine and in the induction of LAK cells to mediate tumor cell cytolysis.     

Mechanisms for generating cell membrane antigens that are recognized by cytotoxic T lymphocytes

Studies with many viruses have revealed that viral specific protein synthesis is an obligatory step in generating antigens on target cells for antiviral cytotoxic T lymphocytes. This has been most clearly demonstrated with DI particles, virions that are structurally complete but lack infectious RNA. Adsorption of such particles onto target cell membranes does not render these cells susceptible to lytic attack by antiviral effector cells, unless some viral protein synthesis transpires. However, some viruses, such as Sendai virus, circumvent the requirement for viral protein synthesis via fusion of the viral envelope with the target cell membrane, a process mediated by a specialized fusion protein. Once inserted into the lipid bilayer, it is likely that viral components and self H-2 noncovalently associate so that the complex can be recognized by antiviral cytotoxic T cells. This idea is supported by the demonstration that viral proteins and H-2 containing membrane proteins, incorporated into reconstituted membrane vesicles or liposomes are recognized by cytotoxic T cells. These data further show that native rather than altered viral and H-2 molecules are the moieties recognized. Associations between antigen and H-2 have been detected by a variety of techniques and in some cases are not random but selective; that is, viral antigens perferentially associate with some H-2 alleles and not others. In summary, these findings indicate that although viral antigens are present in the mature virions, these components are not recognized by antiviral killer cells until integrated into the plasma membrane. This may be achieved either through direct fusion of the viral envelope with the target cell or following viral protein synthesis and insertion of viral antigens into the plasma membrane.     

Distinct and non-cross-reactive epitopes are recognized on B16 melanoma by LAK cells and anti-B16 monoclonal antibodies.

Two rat anti-B16 melanoma monoclonal antibodies (MoAb), designated IB16-6 and IB16-8, recognize an epitope expressed with high density on the surface of B16 parental cells and B16-F1, F10, F10FLR, and BL6 sublines. The purpose of this study was to define by means of cytolytic and clonogenic assays whether these MoAbs reacted with the same or distinct determinants as those recognized on B16 targets by lymphokine-activated killer (LAK) cells. Using 125I-labeled antibody and Scatchard analysis, the affinity constant (KA) of IB16-6 was determined to range from 5.6 to 9.4 x 10(8) liter/M and the number of receptor sites per B16 cell was 4.8 x 10(4) to 2.5 x 10(5). The effects of anti-B16 MoAb on LAK activity were determined by either preincubating 51Cr-labeled B16 target cells with varying concentrations of MoAb, followed by the cytolytic assay, or exposing unlabeled B16 cells to MoAb, and then carrying out a 10-day clonogenic assay. Over a wide range of antibody concentrations, IB16-6 and IB16-8 had minimal effects on LAK activity, and even at MoAb concentrations up to 1 mg there were no changes in target cell sensitivity or colony-forming ability. Enzymatic treatment of B16 melanoma cells with either trypsin or pronase completely removed the epitope recognized by MoAb IB16-6 but did not alter B16 sensitivity to LAK cells. These observations indicate that the LAK recognition unit was distinct from the epitope reactive with MoAb IB16-6 and that the B16 determinant(s) recognized by LAK cells is resistant to proteolytic enzymes. The molecular structure of each of these remains to be determined.     

Inability of EDTA to prevent damage mediated by cytolytic T-lymphocytes.

To analyze the nature of the target cell determinants recognized and bound by killer lymphocytes during lymphocyte-mediated cytolysis (LMC), the specific binding of serologically active tumor cell membrane fractions to cytotoxic T lymphocytes has been investigated. Particulate membrane fractions and soluble antigen preparations (extracted by papain or 3 M KCl) from tumor target cells were tested for their ability to inhibit the destruction of intact ⁵¹Cr-labeled target cells by killer lymphocytes in vitro. The effect of papain-solubilized tumor cell antigen on the binding of killer lymphocytes to tumor cell monolayers was also evaluated. Direct assays to determine the extent of binding of unlabeled or radioiodinated soluble antigen (extracted by papain or deoxycholate) to cytotoxic lymphocytes were carried out. In marked contrast to their serological activity, all of these particulate and soluble preparations failed to inhibit LMC or bind to killer lymphocytes in an immunologically specific way. It is suggested that killer lymphocytes recognize and bind to an antigenic complex whose organization is dependent upon the integrity of the target cell membrane.     

Antibody-dependent cellular cytotoxicity mediated by murine lymphocytes activated in recombinant interleukin 2

The incubation of murine splenocytes in recombinant interleukin 2 (RIL 2) gives rise to lymphokine-activated killer (LAK) cells that can lyse fresh, NK-resistant tumor cells but not normal cells in 4-hr 51Cr-release assays. Lysis by this IL 2-activated cell population was enhanced up to 100-fold by prior reaction of target cells with specific antisera reactive with antigens on the target cells. This antibody-dependent cellular cytotoxicity (ADCC) also resulted in lysis of fresh normal target cells, which are not usually susceptible to LAK lysis. The ADCC was evident after 24 hr of incubation of splenocytes in RIL 2, but peak lytic activity was reached after 3 to 4 days of incubation. The concentrations of RIL 2 needed for the in vitro activation of the effectors in order to attain maximal ADCC was between 100 and 3000 U/ml and parallel the IL 2 concentrations required to generate LAK cells. ADCC mediated by IL 2-activated splenocytes was completely blocked by anti-FcR monoclonal antibodies. Although antisera directed against MHC antigens were used in most experiments, anti-B16 monoclonal antibodies have also shown the ability to induce ADCC mediated by RIL 2-activated syngeneic and allogeneic cells. Treatment of the precursor splenocyte populations with anti-asialo GM1 and complement eliminated the direct LAK activity and the antibody-dependent cytotoxicity, suggesting that both direct and indirect tumor cell lysis may be caused by the same effector cell. ADCC mediated by LAK cell populations represents another possible mechanism for the in vivo therapeutic effects of these cells.     

Target cell lysis by cytotoxic T lymphocytes redirected by antibody-coated polystyrene beads

Cytotoxic T lymphocytes were found to mediate rapid lysis of target cells not normally recognized in the presence of small polystyrene beads coated with a combination of anti-T3 and antitarget cell antibodies. Lysis was not seen with beads bearing one of these antibodies alone, nor with a mixture of two types of beads each coated with a single antibody. The effector cells mediating this lysis include long term allospecific human CTL, and both human and mouse CTL clones recognizing mouse class I MHC Kb Ag. TNP-modified mouse tumor cells, a human lymphoblastoid line, and human red cells were found to be good targets for this cytotoxicity. Polystyrene beads with diameters of 3 to 15 mu caused target lysis, with a dose-response curve which typically went through a maximum and declined at high bead numbers. Maximal bead-redirected lysis by CTL was less efficient than that mediated by soluble antibody heteroconjugates of the same two antibodies. Bead-redirected target lysis was calcium dependent. These results are interpreted as a form of bystander lysis induced by the beads, since the target cell membrane is not directly crosslinked to the region of CTL activation. These observations thus favor a mechanism of lysis involving the polarized secretion of a locally acting lytic agent by CTL.     

A fast-acting cytotoxin derived from Con A-activated porcine leukocytes. II. Mechanism of target cell lysis

Serum-free culture supernatants of Con A-stimulated pig leukocytes contain cytotoxins (PCT) which have a fast-acting lytic effect on a variety of murine lymphomas. Mathematical analysis revealed that the cytodestructive reaction which leads to the immediate release of 51Cr from labeled target cells follows "single hit"/"first order" kinetics. The release of labeled compounds was found to be size dependent, such that low molecular isotope (86Rb) was released at faster rates than high molecular 51Cr-labeled complexes. The 51Cr release was strongly reduced by lowering the temperature to 4 degrees C. PCT-mediated target cell lysis could be competed by cold target cells and the factor could be absorbed by intact cells, plasma membranes and artificial liposomes. Neither the presence of EDTA or of various monosaccharides, nor exposure of target cells to trypsin, metabolic inhibitors, and cytoskeletal antagonists altered the target cells susceptibility to PCT-mediated lysis. Labeling of target cell DNA and subsequent exposure of these target cells to PCT revealed that target cell DNA is degraded into low-molecular-weight split products which is similar to that seen during T-cell mediated target cell lysis. The analysis of the lytic event mediated by PCT thus revealed similarities to T-cell mediated target cell lysis. PCT was clearly distinct from the well-known cytolytic agents lymphotoxin (LT), tumor necrosis factor (TNF), or complement (C'). Taken together, the described characteristics make PCT a unique-acting cytolytic cytokine with anticancer activity and suggest further research concerning its mode of action and its possible therapeutic application.     

Mechanism of target cell recognition by CD3- LGL. I. Development of a monoclonal antibody to a K562-associated target cell antigen.

In an attempt to identify the target recognition molecule(s) involved in the interaction between CD3- large granular lymphocyte (LGL) and a tumor cell target, monoclonal antibodies (mAb) to NK-susceptible K562 tumor cell membrane glycoproteins were developed. After screening by ELISA for reactivity to K562 membrane glycoproteins, two monoclonal antibodies were identified (mAb 35 and mAb 36). One of the monoclonal antibodies (mAb 36) was found to inhibit conjugation between LGL and K562 target cells and also to inhibit lysis of K562 by LGL. Upon further testing, mAb 36 also inhibited the binding between LGL and other NK-susceptible target cells, e.g., Daudi and Molt 4. In contrast, mAb 35, even though binding to K562, did not inhibit the binding of LGL to tumor targets and therefore was used as an isotype control. When mAb 36 was utilized as an affinity matrix, bound proteins specifically inhibited CD3- LGL-K562 conjugation. Experiments involving tunicamycin treatment of tumor target cells demonstrated that mAb 36 recognized a carbohydrate moiety rather than the protein core. Therefore, these data suggested that the target cell recognition molecule which is recognized by mAb 36 appears to be a membrane carbohydrate-associated molecule.     

Binding of active matrilysin to cell surface cholesterol sulfate is essential for its membrane-associated proteolytic action and induction of homotypic cell adhesion

Regulation of cell surface molecules by matrix metalloproteinases (MMPs), as well as MMPs-catalyzed degradation of extracellular matrix, is important for tumor invasion and metastasis. Our previous study (Kioi, M., Yamamoto, K., Higashi, S., Koshikawa, N., Fujita, K., and Miyazaki, K. (2003) Oncogene 22, 8662-8670) demonstrated that active matrilysin specifically binds to the surface of colon cancer cells and induces notable cell aggregation due to processing of the cell membrane protein(s). Furthermore, these aggregated cells showed a dramatically enhanced metastatic potential. To elucidate the mechanism of matrilysin-induced cell aggregation, we attempted to identify the matrilysin-binding substance on the cell surface. Here, we demonstrate that cholesterol sulfate on the cell surface is a major matrilysin-binding substance. We found that active matrilysin bound to the cell membrane and cholesterol sulfate incorporated into liposomes with similar affinities. Treatment of colon cancer cells with beta-cyclodextrin significantly reduced not only matrilysin binding to the cell surface but also matrilysin-dependent proteolysis and cell aggregation. Interestingly, replenishment of cholesterol sulfate, but not cholesterol, neutralized the effects of beta-cyclodextrin. Taken together, it is likely that binding of matrilysin to cholesterol sulfate facilitates the matrilysin-catalyzed modulation of cell surface proteins, thus inducing the cancer cell aggregation.     

The human lymphokine-activated killer cell system. V. Purified recombinant interleukin 2 activates cytotoxic lymphocytes which lyse both natural killer-resistant autologous and allogeneic tumors and trinitrophenyl-modified autologous peripheral blood lymphocytes

Culture of human peripheral blood lymphocytes (PBL) in purified natural or recombinant interleukin 2 in the absence of exogenous antigen or mitogen causes the differentiation of nonlytic precursor cells into lymphokine-activated killers (LAK). A titration of purified Jurkat IL-2 (BRMP, FCRC, NIH) IL-2 showed that the relatively low concentration of 5 U/ml was optimal for LAK activation. When the responding PBL were pretreated with either mitomycin C or gamma irradiation, LAK activation did not occur, indicating that proliferation, in addition to differentiation, is required. The spectrum of target cells susceptible to LAK lysis in a 4-hr chromium-51-release assay includes fresh NK-resistant tumor cells and trinitrophenyl (TNP)-modified autologous PBL. Unmodified PBL are not lysed. Cold target inhibition studies indicated that LAK lysis of autologous TNP-PBL is totally inhibited by fresh tumors cells, and that tumor lysis is inhibited by TNP-PBL. Additionally, allogeneic tumors totally inhibit lysis of autologous tumor cells in other cold target studies. These results demonstrate that the lytic activity expressed by LAK is not HLA restricted, is not limited to tumor cells, and is "polyspecific" as indicated by the cross-reactive recognition of multiple target cell types in these cold target inhibition studies.     

Decorated surfaces by biofunctionalized gold beads: application to cell adhesion studies

We describe a simple but versatile method to decorate solid surfaces randomly with colloidal gold particles to which ligands of cell receptors can be coupled to generate local attraction sites for the control of cell adhesion. A self-assembled monolayer of (3-mercaptopropyl)trimethoxysilane was deposited on glass slides. Gold beads were anchored to the functionalized surface through the sulfur group. We characterized the gold bead distribution on the functionalized surface with reflection interference contrast microscopy. The gold beads were functionalized with a disulfide-coupled cyclic pentapeptide containing an arginine-glycine-aspartic acid (RGD) tripeptide sequence which is selectively recognized by integrin receptors alpha(V)beta(3) of endothelial cells. A blocking layer of bovine serum albumin was adsorbed onto the surface to prevent non-specific binding of the cells. We demonstrate that the RGD-functionalized colloidal gold beads act as local attraction centers, mediating rapid cell anchoring on a substrate impeding cell adhesion in the absence of attraction centers. Surprisingly, microinterferometry shows that after a time delay of about 1 h, the regions of the cell surface between the gold beads form close contacts with the substrate, which is attributed to strong van der Waals attraction after escape of repeller molecules from the contact surface.     

Antibodies to adhesion molecules inhibit the lytic function of MHC-unrestricted cytotoxic cells by preventing their activation.

We evaluated the effect of the antibodies to adhesion molecules CD2, CD11a/CD18 (LFA-1), and CD56 (N-CAM) on MHC-unrestricted cytotoxicity mediated by polyclonal NK cells and LAK cells or by CD3+ or CD3- cytolytic cell clones against a panel of tumor cell targets selected according to expression or absence of the corresponding ligands. We show that (i) antibodies to CD11a/CD18 and, to a lesser extent, antibodies to CD2 inhibit target cell lysis, whereas anti-CD56 antibodies exert little if any effect; (ii) in a model system using polyclonal NK/LAK cells as effectors and K562 or HL60-R (NK-resistant) cells as targets, inhibition of cytotoxicity occurs without a significant impairment of effector to target cell binding; (iii) the cytotoxic function of CD3+ or CD3- cytotoxic cell clones is inhibited differentially by antibodies to adhesion molecules; (iv) conjugates formed in the presence of antibodies which inhibit target cell lysis display a significant reduction of target to effector cell contact surface; and (v) this may lead to defective activation of effector cells, as indicated by lack of redistribution of the microtubular apparatus. We conclude that (i) MHC-unrestricted cytotoxicity is regulated by a number of molecular interactions that span far beyond our present knowledge and that it is strictly dependent on the surface phenotype of the effector cell and of the target cell; (ii) in certain types of effector/target cell interactions, antibodies to adhesion molecules do not prevent conjugate formation but reduce the extent of cell-to-cell surface contact which, in turn, leads to defective activation of the effector cell and, therefore, to inhibition of target cell lysis.     

Tumor-infiltrating lymphocyte secretion of IL-6 antagonizes tumor-derived TGF-beta 1 and restores the lymphokine-activated killing activity

IL-6 is a multifunctional cytokine that regulates cell growth, differentiation, and cell survival. Many tumor cells produce TGF-beta1, which allows them to evade CTL-mediated immune responses. IL-6 antagonizes TGF-beta1 inhibition of CD3 cell activation. However, whether IL-6 restores NK activity, which also is suppressed by TGF-beta1, is not known. We used canine transmissible venereal tumor (CTVT), which produces TGF-beta1, as a model to determine whether IL-6 restores lymphokine-activated killer (LAK) activity. During the progression phase, CTVT cells stop expressing MHC molecules. During the regression phase, the number of surface MHC molecules increases dramatically on about one-third of tumor cells. Tumor cells that stop expressing MHC should be targeted by NK cells. In this study, we found that TGF-beta1 secreted by CTVT cells suppressed LAK cytotoxicity. Interestingly, tumor-infiltrating lymphocytes (TIL) isolated from regressing CTVT secrete high concentrations of IL-6 and antagonize the anti-LAK activity of tumor cell TGF-beta1. TIL also produce IL-6 during progression phase, but the concentration is too low to block the anti-LAK activity of TGF-beta1. There is probably a threshold concentration of IL-6 needed to reverse TGF-beta1-inhibited LAK activity. In addition, in the absence of TGF-beta1, IL-6 derived from TIL does not promote the activity of LAK. This new mechanism, in which TIL manufacture high concentrations of IL-6 to block tumor TGF-beta1 anti-LAK activity, has potential applications in cancer immunotherapy and tumor prognosis.     

Specific reversal of cytolytic T lymphocyte – target cell interaction

A functional assay is described that measures the reversal of specific cytolytic T cell (CTL)-target cell binding. Binding of ⁵¹Cr-labeled P815 cells was stable in suspension but could be readily reversed by the addition of unlabeled P815 cells. The reversal of CTL-tumor cell and CTL-spleen cell binding was H-2 specific; only cells of the same H-2 type as the bound target cell could induce reversal. In all cases, tumor cells were substantially more efficient than spleen cells in inducing specific reversal.     

Heterogeneity of the lymphokine-activated killer cell phenotype.

Lymphokine-activated killer cells (LAK) are functionally defined by their ability to mediate the MHC-unrestricted lysis of a range of tumor targets, while sparing normal cells. They can also lyse TNP-modified normal syngeneic lymphoblasts. We show here that lysis of TNP-modified targets is mediated by CD8+ LAK in a self-MHC-restricted manner, whereas lysis of tumor targets is largely by CD8- LAK and is MHC-unrestricted. LAK generated from the immune-deficient strains Balb/c nude and C.B-17 scid lyse tumor targets as effectively as LAK from normal mice but do not lyse TNP-modified normal targets. Further, lysis of TNP-modified targets, but not tumors, can be inhibited by antibody to the T cell receptor complex. These experiments strongly suggest that recognition of TNP-modified targets is not accomplished by the same mechanism as that of tumors. Rather, they are consistent with recognition of TNP-modified targets by CD8+ LAK cells being mediated via recognition through the T cell receptor.     

Assessment of human natural killer and lymphokine-activated killer cell cytotoxicity against Toxoplasma gondii trophozoites and brain cysts

Because previous work has suggested that NK cells may be important in host resistance against the intracellular parasite Toxoplasma gondii we examined whether human NK cells and lymphokine-activated killer (LAK) cells have activity against trophozoites and cysts of this organism in vitro. A method to radiolabel Toxoplasma trophozoites with 51Cr was developed and direct cytotoxic activity was determined by using modifications of the standard 51Cr release assay. Viability of 51Cr-labeled trophozoites assessed by both methylene blue staining and trypan blue exclusion was greater than 90%. Significantly more 51Cr was released by anti-Toxoplasma antibody and C than by antibody in the absence of C. Incubation of trophozoites with freshly isolated human NK cells or NK cells activated with either rIL-2 or rIFN-alpha did not result in significant release of 51Cr (specific lysis was 0 to 2.3%). In contrast, the average specific lysis of radiolabeled trophozoites by LAK cells was significant (specific lysis was 7.8% +/- 1.1, p less than 0.01). In a series of separate experiments, preincubation of radiolabeled trophozoites with heat-inactivated normal or Toxoplasma antibody-positive human serum increased the cytotoxicity of LAK cells from a mean specific lysis of 15% +/- 4.5 to 39% +/- 8.5, respectively (p less than 0.05), as assessed by 51Cr release. Because previous work has shown that radioisotope release from parasites may be nonspecific, separate experiments were performed to determine the cytotoxicity of LAK cells against antibody-coated trophozoites by using ethidium bromide-acridine orange staining to assess effector cell damage. LAK cells had a mean specific lysis of 51% against antibody-coated trophozoites by ethidium bromide-acridine orange staining. Preincubation with heat-inactivated Toxoplasma-antibody positive human serum did not increase activity of rIL-2-activated NK cells against 51CR-labeled trophozoites. Neither human NK cells (freshly isolated or activated by rIL-2 or rIFN-alpha) nor LAK cells were cytotoxic for purified preparations of cysts of Toxoplasma isolated from the brains of chronically infected mice.     

Cytometrically coherent transfer of receptor proteins on microporous membranes.

A new method (Freeze-Transfer) is described for performing high-resolution immunocytochemistry for soluble cell proteins on frozen sections of biological tissues that involves thaw-mounting frozen tissue sections directly onto the surface of nitrocellulose thin films instead of directly onto glass slides. This technically straight-forward change in methodology resulted in chromogenic immunocytochemical assays for Her-2 and EGF receptors that were 1-2 orders of magnitude more sensitive while still fully utilizing the diagnostic resolving power of light microscopy. The effects of membrane pore size and surface chemistry on the resolution and intensity of Her-2 signal suggest that the enhanced sensitivity of Freeze-Transfer was caused by the cytologically coherent transfer of target molecules normally lost from cut surfaces of cells mounted on nonporous glass during assay.     

A murine very late activation antigen-like extracellular matrix receptor involved in CD2- and lymphocyte function-associated antigen-1-independent killer-target cell interaction

CD2 and lymphocyte function-associated antigen (LFA)-1 are well known as T cell adhesion molecules involved in killer-target cell interactions. However, our recent study revealed that molecule(s) other than CD2 and LFA-1 might be involved in the lymphokine-activated killer (LAK) cell cytotoxicity against certain target cells. In order to characterize such unknown molecules, we established a mAb (RMV-7) which could inhibit CD2/LFA-1-independent LAK cell cytotoxicity and binding to target cells at the effector site. The Ag identified by RMV-7 appeared on splenic T cells late after mitogenic stimulation and was a noncovalently linked heterodimer composed of a 140-kDa alpha-chain and a 95-kDa beta-chain. RMV-7 blocked LAK cell binding to fibronectin (FN), fibrinogen, and vitronectin but not that to laminin or type IV collagen, indicating that the RMV-7-defined molecule is a unique extracellular matrix receptor for FN, fibrinogen, and vitronectin. One of its ligand, FN, was found on the surface of several target cells, and LAK cell cytotoxicity against them was blocked by anti-FN antibody at the target site. Similarly, cytotoxicity of a H-2d-specific CTL clone was inhibited by RMV-7 and anti-FN antibody as well. These results indicate that a unique very late activation Ag-like extracellular matrix receptor on murine CTL and LAK cells contributes to target cell binding and cytotoxicity.     

Human lymphokine-activated killer (LAK) cells: III. Effect ofl-phenylalanine methyl ester on LAK cell activation from human peripheral blood mononuclear cells: Possible protease involvement of monocytes,natural killer cells and LAK cells

Summary We have shown that depletion of monocytes from human peripheral blood mononuclear cells (PBMC) byl-phenylalanine methyl ester (PheOMe) enhanced lymphokine-activated killer cell (LAK) generation by recombinant interleukin-2 (rIL-2) at high cell density. In this study, we have investigated the mechanism of action of PheOMe on LAK activation by using trypsin, chymotrypsin, tosylphenylalaninechloromethanol (TPCK, a chymotrypsin inhibitor), tosyl-l-lysinechloromethane (TLCK, a trypsin inhibitor), phenylalaninol (PheOH), and benzamidine. PBMC were treated with 1–5 mM PheOMe for 40 min at room temperature in combination with the various agents, washed and assessed for their effects on natural killer (NK) activity against K562 cells and monocyte depletion. The treated cells were then cultured with or without rIL-2 for 3 days. LAK cytotoxicity was assayed against⁵¹Cr-labeled K562 and Raji tumor target cells. TPCK at 10 µg/ml partially inhibited depletion of monocytes by PheOMe. TLCK did not prevent depletion of monocytes nor inhibition of NK activity induced by PheOMe. TPCK and TLCK inhibited NK activity by themselves. TPCK but not TLCK inhibited rIL-2 induction of LAK cells. On the other hand, PheOH and benzamidine (analogs of PheOMe) lacked any effect on monocyte depletion but abrogated the inhibitory effect of PheOMe on NK activity. They had no effect on rIL-2 activation of LAK activity enhanced by PheOMe. Trypsin potentiated the inhibitory effect of PheOMe on NK activity and monocyte depletion. Trypsin partially inhibited IL-2 activation of LAK activity enhanced by PheOMe. Chymotrypsin had little effect on NK activity but prevented the inhibitory effect of PheOMe on NK activity. It had little effect on monocyte depletion induced by PheOMe. PheOMe was hydrolysed by monocytes and chymotrypsin to Phe and methanol as determined by HPLC. TPCK inhibited hydrolysis of PheOMe by monocytes. Our data suggest that the effects of PheOMe on monocytes, NK cells and LAK activation involve protease activities of monocytes.