distinction of partially purified human natural killer ... · pbl- or idi conditioned media...

[CANCER RESEARCH 48, 891-898. February 15. 1988]

Distinction of Partially Purified Human Natural Killer Cytotoxic Factor from

Recombinant Human Tumor Necrosis Factor and RecombinantHuman Lymphotoxin1

Tammo Bialas,2 Jonathan Kolitz, Ester Levi, Andrei Polivka, Sadik Oez, Glenn Miller, and Karl Weite

Laboratory ofCylokine Biology, Memorial Sloan-Kettering Cancer Center, New York, New York 10021

ABSTRACT

Natural killer cell cytotoxic factor (NKCF), a cytotoxic factor contributing to human natural killer cell-mediated cytotoxicity, was generatedfrom lymphocyte-conditioned medium using various stimuli. Crude NKCFactivity was concentrated, and partially purified by ammonium sulfateprecipitation and gel filtration. NKCF activities eluted as two molecularweight peaks, corresponding to M, 33,000-43,000 (pool I) and approximately M, 5,000 (pool II). The cytotoxic activity and target specificity ofthe partially purified NKCFs were found to be different from bothrecombinant human TNF and recombinant human lymphotoxin. In theNKCF assay, up to 10* units/ml of TNF and lymphotoxin had virtuallyno effect, whereas both NKCFs lysed 22% (range 17-33%) of the NK-sensitive target K562. In contrast, TNF and lymphotoxin were active ina standard assay against the sensitive murine L929 fibroblast cell line inall concentrations tested (10 ' -10'' units/ml). In addition, the effect of

these cytotoxic factors in a short-term (4-h) chromium-release assayusing peripheral blood mononuclear cells as effector cells was tested:only NKCF (pool I), but not TNF, lymphotoxin, or low molecular weightNKCF (pool II), enhanced NK and lymphokine-activated killer cellcytolysis, both against the NK-sensitive target K562 and the NK-resistantmelanoma cell line SK-MEL 30. Results were not affected in the presenceof neutralizing antibodies against TNF. NKCF could, therefore, be distinguished from TNF and lymphotoxin with respect to their biologicalactivities.

INTRODUCTION

Several molecules may contribute to the lytic process ofCMC,3 some of which have been purified, molecularly clonedand biologically characterized (1-5). Besides TNF and lymphotoxin, other such molecules include macrophage cytotoxins (6-8), the pore-forming proteins (9, 10), perform 1 and 2(11-13),cytolysin (14-16), leukoregulin (17, 18), defensin (19), certainserine esterases (20) and certain proteoglycans (21, 22). Someof these factors are involved in the specific lytic mechanism ofMHC-restricted T-cells, whereas others are related to immunecells which kill a broad spectrum of tumor cells without apparent specificity (non-MHC restricted T-cells, NK-cells, LAK-cells, monocytes/macrophages) (15, 23-27). Interestingly, oneof them closely resembles the ninth component of the comple-

Received 3/23/87; revised 7/27/87, 11/11/87; accepted 11/16/87.The costs of publication of this article were defrayed in part by the payment

of page charges. This article must therefore be hereby marked advertisement inaccordance with 18 U.S.C. Section 1734 solely to indicate this fact.

1This work was supported in part by Grants CA 23766 and CA 33484 fromthe National Institutes of Health and Grant Bi 317/1-1 from the DeutscheForschungsgemeinschaft.

2To whom requests for reprints should be addressed, at Laboratory of CytokineBiology, Memorial Sloan-Kettering Cancer Center, 1275 York Avenue, NewYork, NY 10021.

3The abbreviations used are: CMC, cell-mediated cytotoxicity; TNF, tumornecrosis factor; MHC, major histocompatibility complex; NK, natural killer,LAK, lymphocyte-activated killer; PBMC, peripheral blood mononuclear cell;NKCF, natural killer cytotoxic factor, rh-INF, recombinant human interferon;CM, complete medium; FCS, fetal calf serum; PBS, phosphate buffered saline;LGL, large granular lymphocytes; E/S ratio, effector/stimulator ratio; E/T ratio,effector/target ratio; PHA, phytohemagglutinin; ConA, concanavalin A; PMA,phorbol myristate acetate; AcA, acrylamide-agarose gel; rh-IL-2, recombinanthuman interleukin-2; EIA, enzyme immunoassay; MTT, 3-(4,5-dimethylthiazol-2-yl)-2,5-diphenyltetrazolium bromide, thiazolyl blue (Sigma, no. M2128); OD,absorption data; LUÂ»,lytic unit.

ment system (11, 13), thus linking mediators of cellular killingto the humoral system of immune defense.

Studies characterizing the cytotoxicity mediated by non-MHC-restricted lymphoid effector cells have identified a cytotoxic activity in the conditioned medium of nonadherent PBMC(28). This factor has been termed NKCF (29). While information is available about this putative factor and its role in NK-CMC (30-39), it has not been fully biochemically characterized,and its biological significance remains unclear.

In this study we report the partial purification of NKCF fromhuman lymphocyte-conditioned medium and compare its biological activities to rh-TNF and recombinant human lymphotoxin.

MATERIALS AND METHODS

Blood Cell Separation and Effector Cell Preparation

Human peripheral blood or leukocyte-enriched buffy-coat preparations from normal donors were obtained from the blood bank ofMemorial Hospital. Cells were processed within 1 to 2 h after collection. The mononuclear cell fraction (PBMC) was isolated by Ficoll-Hypaque density separation as previously described (40), washed, andresuspended in CM containing RPM1 1640 supplemented with 10%heat-inactivated FCS, 20 mmol glutamine, 15 mmol 4-(2-hydroxy-ethyl)-l-piperazineethanesulfonic acid buffer, 1% penicillin, and 1%streptomycin (GIBCO, Grand Island, NY). PBMC were then plated ata concentration of 3 x 106/ml on 100-ml tissue culture plates (Falconno. 3003; Becton Dickinson, Oxnard, CA) for 1 to 2 h at 37Â°Cto allow

adherence of monocytes/macrophages. Nonadherent cells were washedoff with cold PBS and plates were rinsed vigorously five to 10 times.After collection, all nonadherent lymphoid cells (referred to as PBL)were resuspended in CM for further use.

Enrichment for LGL Cells

Two variations of a Percoli density separation method were used:the original seven-layer discontinuous gradient centrifugation (41) anda modification using a two-step double-layer gradient. Briefly, in theoriginal procedure, 2 ml of seven Percoli concentrations (42.5-57.5%in 2.5% steps; Pharmacia, Piscataway, NJ) in PBS were layered in a15-ml conical centrifuge tube (Falcon, Becton-Dickinson, Mountain-view, CA, or Corning, Corning, NY). Two ml of PBL at a density of3-5 x 107/ml were layered on this gradient. After centrifugation at 550x g for 30 min at 20Â°C,cells from each layer were collected, washed

2x, and resuspended in CM. Examination of May-Gruenwald-Giemsastained cells of cytospin preparations from each layer demonstratedthat large granular lymphocytes (LGL) were enriched in interfaces 2and 3 and, partially, in 4 (corresponding to 45 to 50% Percoli). Ten %of the total load was present in these layers, each of which contained30-50% LGL. Viability using Trypan blue exclusion was always >95%.The second method utilized 50-ml tubes with 52 and 45% Percolihilavor. loaded with 10 ml of PBL suspension containing 1-2 x 10"

cells/ml. The yield with this method was ~25%, with enrichment of>50% LGL.

Cell Lines and Target Cell Preparation

The human erythroleukemia cell line K562, the human melanomacell line SK-MEL 30, the TNF-sensitive murine fibrosarcoma cell line

891

on April 3, 2017. © 1988 American Association for Cancer Research. cancerres.aacrjournals.org Downloaded from

http://cancerres.aacrjournals.org/

DISTINCTION OF NKCF FROM TNF AND LYMPHOTOXIN

WEHI 164, and the TNF-sensitive and -resistant subclones of thetransformed murine fibroblastic cell line L929 were grown in CM.L929r was obtained from Ms. B. Williamson (Sloan-Kettering Institute,Rye, NY). The human diploid fibroblastic cell line I S 4 was maintainedin minimum essential medium with 10% FCS. All cell lines were testedfor absence of mycoplasma and washed with PBS prior to each assay.

NKCF Generation

NKCF was generated by a procedure originally described by Wrightand Bonavida (29, 37, 38) with only minor modifications. Briefly, PBLor enriched LGL, were incubated with KS62 cells at an E/S ratio of100:1 (standard ratio; other ratios tested: 25:1, 50:1, and 200:1) for 24h under serum-free conditions (RPMI 1640, supplemented with 15 IHM4-(2-hydroxyethyl)-l-piperazineethanesulfonic acid buffer, 1% nones-sential amino acids, 1% glutamine, 1% vitamins, 1% penicillin, 1%streptomycin) at 37Â°Cand 5% CO2. Concentration of effector (producer) cells was 5 x 106/ml and cultures were started either in tissueculture flasks (Corning, 25 cm2 and 75 cm2; Corning, NY) or in six-

well tissue culture plates (Nunc, Interlab; Thousand Oaks, CA). Thesame effector cell preparations were also stimulated with either 1%PHA (GIBCO), 10 Mg/ml ConA (Sigma, St. Louis, MO), or 20 ng/mlPMA (Sigma), for 24 h under identical conditions in the presence orabsence of rIL-2 (1000 U/ml). Controls consisted of producer cells(PBL or LGL) alone in medium (spontaneous generation) and targetcells (K562) alone. PHA, ConA, and PMA were not cytotoxic againstK562 in 20-h 51-Cr release assays (data not shown). Supernatants fromeach stimulation were harvested, filtered through 0.45-jim Nitex filters,supplemented with 0.02% sodium azide, stored at 4Â°C,and assayed

within 1 month.

NKCF Purification

PBL- or Idi conditioned media containing NKCF activity werepooled from nine donors, and proteins were concentrated by ammonium sulfate precipitation (80% saturation). Precipitates were resus-pended in PBS, dialyzed extensively against PBS (50x volumes withrepeated changes), supplemented with antibiotics and sodium azide(0.02% v/v; Sigma), and assayed for NKCF activity. Active batcheswere pooled and loaded on an acrylamide-agarose gel filtration column(AcA-54, 500 ml, 2,6 x 90 cm; LKB, Gaithersburg, MD). Singlefractions (total volume, 7 ml) were tested for biological activity andtotal protein content, and active fractions were pooled into a highmolecular weight pool (M, 33,000-43,000; pool I) and a low molecularweight pool (M, approximately 5,000; pool II).

Cytokines

rh-IL-2 was kindly provided by Cetus (Cetus Corporation, Emeryville, CA). rh TNF or rh TNF-a and rh lymphotoxin or rh TNF-/3 werea gift of Drs. Aggarwal and Shepard (Genentech, South San Francisco,CA). rh IFN-a2B was obtained from Schering-Plough (Kenilworth,NJ).

Cellular Cytotoxicity Assays

NK Assay. Effector cells were plated at a concentration of 5 x IO6/ml in logi dilutions in 96-well U-bottom microtiter plates (Costar No.3799, Cambridge, MA) using the Cetus Propelle apparatus. Experi-menls were plaled using eighl E/Tratios (100:1 lo 0.78:1) in triplicati.-.Target cells (106-107) were incubated with 100 nCi 51-chromium(5'Na2CrO4; specific aclivily, 1 //Ci/ml; New England Nuclear, DuPonl,Boslon, MA) for 60-90 min al 37Â°C.After washing Ihree limes, cellswere adjusted lo a plaling concenlration of 5 x 104/rnl to yield 5000larget cells per well. Plates were incubaled for 4 h al 37Â°Cand 5%

humidified CO2, harvesled wilh a Skalron harvesting system (Skatron,Slerling, VA) and counled wilh a gamma counter (Packard 5500;Packard, Slerling, VA). CM was used as negalive control (spontaneousrelease), and 2% sodium dodecyl sulfate in medium was used as posilivecontrol (maximum release). The release index (sponlaneous release:maximum release x 100) was usually -5% and always <10%.

LAK Assay. Cullures coniaining 1 or 2 x 106/ml effeclor cells (PBLor LGL) in 25-cnr lissue cullure flasks (Corning) were incubaled wilh

1000 U/ml of rh-IL-2. Following 6 lo 7 days of incubalion at 37Â°C,

5% CO2, which is optimal for the generation of LAK cytotoxicily (dalanoi shown), cells were washed Iwo limes, and replaled as describedabove for ine NK assay.

TNF and Lympholoxin Assays

TNF and lympholoxin cyloloxicily were assayed in a semiaulomateddye-coupled MTT-EIA (42) as follows: one hundred n\ of NKCF-conlaining supernalanls, TNF or lymphotoxin standards were platedout serially in logio dilutions using eight concentrations in triplicate in96-well U-bottom microtiter piales. Sensilive or resistant L929 largelcells, prelrealed with 5 Mg/ml mitomycin C (Sigma) for 2 h, were addedal a conceniralion of 2.5 x 10" cells per well in 100 M'-Assays were

incubaled for 44 h. 50 M!of MTT added per well (5 mg/ml), and pialeswere incubated for another 4 h. Piales were Ihen harvested, 100 til acid-isopropanol (0.04 N HC1 in 2-propanol; Sigma) was added to each welland the optical densily was measured in an enzyme-linked immunosor-benl assay auloreader (Bio-Tek, model EL 309; Bio-tek Instruments.Burlington, VT) in dual-wavelength mode at 570 and 630 nm. ODvalues were compuled and expressed as described below.

NKCF Assay

A slandard 20-h 51-chromium release assay was employed (33, 38).NKCF-conlaining samples (PBMC supernalanls, crude or partiallypurified NKCF) were tesled al four differenl dilutions (Iog2 dilutions)in 96-well U-bollom microliter plates. Preparation and harvest of targetcells and calculation of Ihe release index were performed as in ihe NKassay. The release index was always below 20%. In some experiments,the assay was extended lo 48 h, and a second sel of samples weresimullaneously assayed in an MTT-EIA for 48 h.

Interferon Assay

In order lo exclude ihe presence of inlerferon aclivity in NKCFsamples, a standard neutralization assay measuring the inhibition ofthe cylopalhic effecl of vesicular stomatilis virus on human cells wasemployed. In brief, aliquols of lesi samples were serially log dilutedand plaled in 50-^1 volumes in flal-bollom 96-well microtiter piales.CM and 150 U/ml inierferon-a2B were used as negalive and posiliveconlrols. Cells of ihe human fibroblaslic cell line FS4 were added at 4x 104/ml in 100-Mlvolume and plates were incubaled for 24 h. After

conlrol for viable and oplimal monolayer formation. 50-^1 volumescontaining 10s plaque-forming units of vesicular stomatitis virus wereadded per well, followed by another 24-h incubation period. Each wellthen was scored for plaque formalion or cytopathic effects and resultsare expressed as the tiler equivalenl lo ihe lowesl posilive dilution.

Anlibody Neutralization Assays

Monoclonal antibody against rh-TNF was obtained from Genenlech(Soulh San Francisco, CA) (TNF-E, loi no. 3314-16; 1.4 mg/ml;neutralization activily: 6000 NU/M!). All neutralization assays withanti-TNF anlibodies were performed in 96-well microliter plaies, asdescribed for NK, NKCF, and TNF assays, with 6000 NU of antibodyper well being added in 10 M'volume. Control wells received 10 M'ofmedium.

Data Expression and Analysis

Data from ihe cylotoxicity assays (counts per minute and OD) werefirsl averaged from triplicate assays and Ihen converted to percentageaccording lo Ihe formula

% Specific lysis =experimental rei. - sponlaneous rei.

maximum rei. - sponlaneous rei. x 100

For LU calculations, ihe percentage of specific lysis corresponding toevery dilution or concentralion was plotted and the best-fiiting regres-sional analysis was utilized, generally linear rÃ©gressionor ihe exponential fil equation (43). Results of cellular killing (NK and LAK) areexpressed eilher in percentage of specific lysis at a particular E/T ratio

892




or in LU20/106 cells, with one LU being an index of the effector cell

number required to kill 1000 target cells. Factor cytotoxicity is expressed either in percentage of specific lysis at a given concentration ordilution or as LU50, where half the number of targets is defined to havebeen killed.

RESULTS

NKCF Generation

The median cytotoxicity of crude PBMC supernatant conditioned by the stimuli detailed in "Methods" was 10% (range,

0-16%; SD, Â±1.5%)in the standard NKCF assay against K562.E/S ratios of 25:1, 50:1, 100:1, and 200:1, 24- and 48-h stimulation times, 1%, 5%, and 10% PCS and serum-free mediumwere not associated with major differences with respect toNKCF generation (data not shown). We elected to use a 24-hstimulation period at a 100:1 E/S ratio under serum-free conditions in subsequent experiments.

Table 1 shows the cytotoxic effect of conditioned mediagenerated under different conditions using various stimuli onK562 cells. Results are shown as mean and range of experimentswith ten different donors. In approximately one third of allexperiments we observed low, but measurable "spontaneous"

cytotoxicity on the part of unstimulated cell supernatants aswell as no cytotoxic effect mediated by fully stimulated producercells. The lytic activity of the supernatant generated from theLGL-enriched effector cells was slightly higher than the PBMCgenerated supernatant (Table 1).

NKCF Purification

Supernatants from nine human donors were concentrated byammonium sulfate precipitation (80% saturation), dialyzedagainst PBS and tested for cytolysis (Table 2). The concentratewas then further concentrated by dialyzing against PBS containing 50% (polyethylene glycol 6000-8000; Sigma, St. Louis,MO). This material, approximately 20-fold concentrated, wasfractionated by gel filtration on an AcA-54 column. The bioac-tivity, as assayed in the NKCF assay, was mainly recovered asa high molecular weight peak (pool I, fractions 34-40) corresponding to a molecular weight of 33,000-43,000, and as a lowmolecular weight peak (pool II, fractions 66-70) correspondingto a molecular weight of approximately 5000 (Fig. 1). Fractionsof both molecular weight ranges were pooled and tested forbiological activity. Results are shown in Table 2. Both molecularweight pools remained active for at least 3 months.

Biological Characterization of NKCF

Activity against Different Target Cell Lines (K562, SK-MEL30, and L929) and Comparison to TNF and Lymphotoxin. Fig.

Table 1 Cytotoxicity of crude NKCF supernatants generated by variousstimulatory agents

Effector cells were stimulated at 5 x 10'/rnl for 24 h with stimuli shown.

Supernatants were assayed in standard NKCF assay against KS62 target cells.Results reflect experiments of different normal donors (n = 13), indicating meanvalues and ranges. Only preparations with activity >2% specific lysis wereconsidered. Standard deviation of specific lysis was Â±1%.

PBL LGL

StimulantControl

K562 (100:1)PHA(1%)ConA(10iig/ml)IL-2 (1000 U/ml)PMA (20 ng/ml)K562 + IL-2n10

10137

558Specific

lysis(%)3.7

(3-6)6.0 (4-10)9.5(4-15)5.2 (3-9)7.5(5-10)6.0(2-10)5.5(3-10)n3

333

3Specific

lysis(%)4.0

(3-6)7.0(3-11)

13.0(9-15)10.6(9-13)

11.0(5-15)

Table 2 Comparison of crude and partially purified NKCF cytotoxicity againstKS62 target cells

All NKCF preparations were assayed in standard NKCF assay (20-h "Crrelease) against K562. Final concentration of NKCF was 50% (v/v) in 200 pi.Standard deviation of specific lysis was Â±1.5%.Approximate molecular weightswere determined by molecular sieving through an AcA-54 gel nitration column.

NKCF preparation Specific lysis (%) Molecular weight (AD)

Crude supernatant (LCM) 5-16Ammonium-sulfate precipitate 16-48Gel filtration:Pool I 17-19Peak fraction 39 of Pool I 23Pool II 8Peak fraction 68 of Pool II 33

33-4334~5~5

X specific lysis50 r

68 k.D 42 kO 25 kO

Fig. 1. NKCF bioactivity after AcA-54 gel filtration. 50 ml of approximately20-fold concentrated LCM supernatant was fractionated on an AcA-54 gelfiltration column. Three molecular weight standards (albumin, ovalbumin, andchymotrypsinogen) with molecular weights of 68,000. 42,000, and 25,000, respectively, elute at peak fractions (top arrows). Single fractions were assayed in astandard NKCF assay and active peak fractions were pooled in two pools (bottomarrows): pool l, M, 33,000-43,000; pool II, ~M, 5,000.

2 describes the assay systems using TNF and lymphotoxin. Fig.2, A and C show the activity of TNF and lymphotoxin in an L-cell assay against sensitive and resistant murine L929 cells,tested in a MTT-EIA assay. The effects of TNF and lymphotoxin were also tested in this assay using the NK-sensitive K562cells and the NK-resistant SK-MEL 30 cells as targets (Fig. 2,A and C). Both molecules demonstrated a dose-dependentcytotoxicity against the TNF-sensitive L929 cells, but virtuallyno effect on K562 or the TNF-resistant L929 subclone (minimalcytotoxicity above IO5 U/ml). Similar dose-response kineticswere obtained when cytotoxicity was measured in 48-h 51Cr-

release assays (Fig. 2, B and D). A relative discrepancy betweenresults obtained with the 51Cr-release and MTT-EIA assays wasobserved only with the SK-MEL 30 cell line.

Fig. 3 compares crude NKCF activities with TNF and lymphotoxin. In a 48-h MTT-EIA assay, crude NKCF (concentration, 50% v/v) demonstrated high K562 killing (51%), andmoderate cytotoxicity against SK-MEL 30 (32%) and L929s(31 %) (Fig. 3, top). L929r was not killed by TNF, lymphotoxin,or crude NKCF. The 48-h 51Cr-release assays yielded similar

results (Fig. 3, middle). In the NKCF assay, only NKCF butnot TNF or lymphotoxin, had activity against K562 or SK-MEL 30 (Fig. 3, bottom).

Using partially purified NKCF (pool I or low molecularweight pool II) in these assays, higher cytotoxicity as comparedwith crude NKCF was obtained; 17-33% in the standard NKCFassay (Table 2) and 76-97% in the MTT-EIA, both against

893




10Â«TNF conccntrotion [Units/ml]

IO5 104 103 )0< IO1

TNF concentrotion [UnÃ¬t3/ml]

10Lymphotoxin concentrotion [Units/ml]

IO"1 lÃ²6 IO5 io-'Lymphotoxin concentration [Units/ml]

Fig. 2. Dose-response kinetics of TNF (A and B) and lymphotoxin (C and l>) cytotoxicity were determined over an eight-log concentration range against anidentical target cell panel in different assay systems: I and C, in an MTT-EIA assay; B and />, in a "Cr-release assay; both over 48 h. Both cytokines were testedagainst four targets, K562 (D). SK-MEL 30 (A), L929r (â€¢)and L929s (O).

K562. Fig. 4 shows the lytic activity of pool I and pool IIpartially purified NKCF and TNF measured using both 51Cr-release and MTT-EIA assays against K562 and a panel of TNF-sensitive and -resistant cell lines. Both NKCF pools I and IIare able to lyse TNF-sensitive targets (L929s and WEHI 164)as well as TNF-resistant targets (L929r and K562), whereasTNF is active on the TNF-sensitive targets only.

While both partially purified NKCFs have the same or verysimilar target specificity, they differ, besides their molecularweight, in the lytic potency and kinetics. Pool II does not exertcytotoxic effects within 20 h, and, even after 48 h, is less potentthan pool I (Fig. 4).

Enhancement of NK-Cell Activity of PBMC by NKCF. Fourh MCr-release NK assays using PBMC or LGL as effector cells

(E/T ratio, 50:1) in the presence of crude NKCF (25%; v/v)resulted in enhancement of cytotoxicity against K562 (59 versus30%) and SK-MEL 30 (42 versus 6%) (Fig. 5). This effect wasonly seen with NKCF, but not with TNF (1000 U/ml) orlymphotoxin (1000 U/ml). Higher concentrations of TNF (upto IO5 U/ml) were also without enhancing effects (data not

shown).Similar results were seen using partially purified NKCF (pool

I). Fig. 6 shows the increase of specific NK lysis with additionof NKCF (pool I; concentration, 25% v/v) from a representativeexperiment. The median increase in cytotoxicity was 157%(range, 44-310%), corresponding to 1.4- to 4.1-fold enhancement (E/Tratio 50:1). Data were derived from five independentexperiments with three different donors and tested against twotargets (Table 3). Only NKCF (pool I) was active against both

targets; no effect of NKCF (pool II) or controls was observed,suggesting that the two activities correspond to biologicallydifferent molecules. Controls included NK effector cells aloneand effector cells stimulated with other cytokines (TNF, lymphotoxin). Medium controls (spontaneous release) were always<10%. NKCF (pool I) alone, without effector cells, was alwaysnegative in these 4-h assays.

Enhancement of LAK-Cell Cytotoxicity by NKCF. LAK-cellswere used in standard LAK assays against K562 and SK-MEL30 targets. When LAK assays were performed in the presenceof crude NKCF (25% v/v), enhancement of cytotoxicity similarto that obtained with NK-cells was seen (Table 4). Crude NKCFenhanced LAK-cell lysis 2.5-fold against K562 and 4.7-foldagainst SK-MEL 30. Neither TNF nor lymphotoxin showedenhancement of LAK-cell killing above baseline in this system

(Table 4). A marked increase in cytolysis compared with eithercontrols or LAK-cells stimulated with other cytokines could beobserved using partially purified NKCF (pool I): K562 cytotoxicity was increased 2-fold and SK-MEL 30 cytotoxicity 4.3-fold

(Table 4). No enhancing effect of low molecular weight NKCF(pool II) was detected.

Exclusion of Interferon Activity in Partially Purified NKCF.Since interferons mediate an augmentation of NK cytotoxicity(44-46), we assayed for IFN activity in the partially purified

and biologically active NKCF preparations: both active fractions, pool I and pool II, are completely devoid of interferonactivity, as determined in a standard interferon bioassay (Table5).

894




Â«SPECIFIC LYSIS X SPECIFIC LYSIS100

l Mi SU-XL 30 L MS r L SÂ» l

I SPECIFIC LYSIS

ftÂ« Ji â€”aJIK 562 SK-NCL L 9M r . 9?9 *

Fig. 3. Lytic activity of TNF, lymphotoxin, and NKCF tested against thesame target cell panel in three different assays. Top, MTT-EIA over 48 h; middle,"Cr-release assay over 48 h; bottom, "Cr-release assay for 20 h. Shown are

maximal level of specific lysis for each cytokine target combination, with TNFand lymphotoxin concentration up to 10*/ml and NKCF concentration 50% (v/v).m. TNF:SI. lymphotoxin: â€¢NKCF.

Antibody Neutralization Studies with Monoclonal Anti-TNFAntibody. When partially purified NKCFs (pool I or pool II)were tested against K562 or sensitive L929 cells in the presenceof anti-TNF moAb (6000 neutralizing units/well), no inhibitionof NKCF-mediated lysis could be detected. Assayed in the 48-h MTT-EIA, pool I (1:4 dilution) lysed 99% L929s cells and97% K562 cells, and 86 and 90% of those targets, respectively,at 32-fold dilution. This level of cytotoxicity was identical inthe presence and absence of anti-TNF moAb (Fig. 7). Similarly,cytotoxic activity of pool II was not inhibited. At 1:2 dilution,31% of L929s cells and 30% of K562 cells were killed. Incontrast, TNF lysis of L929 cells (at 1000 U/ml) was completelyinhibited by anti-TNF under the same assay conditions. Cytotoxic effects of pool I, pool II, and TNF on TNF-sensitive L929fibroblasts in the presence of monoclonal anti-TNF antibodyare shown in Fig. 7.

In addition, no neutralization effect of anti-TNF moAb wasobserved when the effect of pool I partially purified NKCF onPBMC effector cells was tested in a NK assay against K562target cells. Lytic activity with and without 6000 NU/well ofanti-TNF antibody remained identical (data not shown).

DISCUSSION

This study describes and biologically characterizes two biochemically distinct cytotoxic activities produced by stimulatedhuman NK-cells. There is evidence that NK-cells produce one,

L 929 s L 929 r HEHI 164

X SPECIFIC LYSIS

100

40

20

L 929 s L 929 r MEHI 164

Fig. 4. High molecular pool I and low molecular pool II partially purifiedNKCF, as well as TNF. were tested for lytic activity against K562 and a TNFtarget cell panel, consisting of L929s, L929r, and WEHI 164. Top, MTT-EIAover 48 h; bottom in 5lCr-release assay for 20 h. Shown are maximal level of

specific lysis. The TNF (â€¢)concentration is 10V U/ml. pool I O) and pool IIm NKCF are onefold diluted (50% v/v).

specific lysis100 r

Fig. 5. Enhancement of NK-cell cyloioxicity by NKCF. Fresh isolated PBMC(50 ti\ at 5 x I0'/ml) as effector cells and 50 â€žIof crude NKCF (25% v/v), TNF

(1000 U/ml). or medium (control) were mixed and tested against K562 (â€¢)andSK-MEL (if) 30 target cells in a standard NK assay. The Â£y7"ratio shown is 50:1.

and possibly more, humoral factors mediating target cell killing(28, 29, 32, 34). A model for NK-cell cytotoxicity has beenproposed, in which NKCF mediates the lethal hit to target cells(47). Crude preparations of this factor have been usually testedin long-term radioisotope-release assays (5lCr, '"In) against

K562 (30, 33, 36, 38). Results have yielded a relatively lowlevel of killing, with specific lysis generally in the range of 10-20% (39), and seldom exceeding 25% (30, 34, 47). This hasbeen attributed to impurity and/or dilution of this soluble

895




X Specific LysisSO

20

Control + Buffer * PoolI + PoolII t TNF

Fig. 6. Enhancement of NK-cell cytotoxicity by partially purified NKCF.Fresh isolated PBMC (50 Â¡Aal 5 x I0*/ml) as effector cells were mixed withdifferent gel filtration pooled fractions (buffer control, pool I. pool II) at 25% v/v and controls (medium and TNF at 1000 U/ml), and were tested in a NK assay(E/T ratio, 50:1). Shown is one representative out of five experiments (Table 3).The increase in NKCF (pool 1)enhanced NK lysis is here 107% on K562 (â€¢)and266% on SK-MEL 30 (D).

Table 3 Enhancement of NK-cell-mediated cytotoxicity bypartially purified NKCF

PBMC effector cells (5 x 106/ml) in 50-pl volume were plated in microtiterplates and supplemented with 50 ill of either NKCF (pool I; 25% v/v), cytokinecontrol (TNF or lymphotoxin; 1000 U/ml), or medium (NK control). "Cr-labeledK562 or SK-MEL 30 cells were added as targets, resulting in an E:T ratio of50:1. Standard deviation of specific lysis was <Â±1%. An enhancing index of +2.6(range. 1.4-4.1). corresponding to an increase in specific lysis of 157%, wascalculated from five experiments with three donors as shown above. (K562 index,+2.6; SK-MEL 30, +2.5).

Specific lysis

ExperimentDonor

1

Donor 2

Donor 3TargetK562

SK-MEL 30K562SK-MEL 30K562SK-MEL 30NKcontrol5

NDÂ°

3918146NK

+ NKCF(poolI)20.5

ND622629IIEnhancing

index(fold)+4.1

ND+ 1.6+ 1.4+2.1+3.6

' ND, not determined.

Table 4 Enhancement of LAK-cell cytotoxicity by crude andpartially purified NKCF

LAK effector cells (7-day IL-2-stimulated PBMC) were supplemented withcrude or partially purified NKCF (25% v/v), TNF (1000 U/ml). lympholoxin(1000 U/ml) (cytokine control), or medium (LAK control), and serially Iog2diluted from 5 x 10*cells/ml. "Cr-labeled K562 or SK-MEL 30 cells were addedas targets. Lytic unit values were calculated from four E/T ratios (50:1 to 6.2:1)and reflect the mean of three experiments with different donors. Standarddeviation of specific lysis was <Â±1%. An enhancing index, corresponding to anincrease in specific lysis caused by NKCF (pool I), was +2.0 for K562 and +4.3for SK-MEL 30.

EffectorcellLAKLAKLAKLAKLAKLAKCytokineControl+TNF+Lymphotoxin+NKCF

(crude)Control+NKCF(poolI)K56210.18.810.225.17.117.7LU:oSK-MEL

3014.010.711.155.83.55.1

factor, leading to slower, and perhaps impaired, kinetics oftarget recognition and binding as compared with that obtainablewith intact factor-secreting NK-cells (30).

Biochemical characterization has provided evidence for theglycoprotein nature of the molecule (30), which was shown tobe both sensitive to Trypsin and papain (38,47), and inactivatedby heating to >63Â°C,pH 2, or 2.8 M urea treatment (31, 38,

Table 5 Determination of Interferon activity in viral inhibition assayA standard neutralization assay measuring the inhibition of the cytopathic

effect of vesicular stomatitis virus on human cells was employed: aliquots of testsamples were serially log-diluted and plated in 50-pl volumes in flat-bottom 96-well plates. Cells of the human fibroblastic cell line FS4 and 10* plaque-formingunits were added. After 48-h incubation each well was scored for plaque formationor cytopathic effects.

Cytokine

Cytopathic effect

Concentration (dilution) Pool 1 Pool II

NKCFIFN-tt2aCM1:1 ++1:10++1:100++1:1000++150

U/ml++

TNF

Pool I

Pool II

20 40 60

X SPECIFIC LYSIS

80 100

Fig. 7. Anti-TNF antibody neutralization assay. Partially purified NKCF,pool I and pool II, and TNF were tested for cytotoxic activity on L929s cells (48hMTT-EIA) in the presence of 30 NU/ml of anti-TNF monoclonal antibody (â€¢).The neutralizing capacity of the moAb was 6000 NU/well, TNF concentrationwas 1000 U/ml, and NKCF samples were, after serial log} dilution, tested atdilution indicated. Controls (S) included experiments without antibody as well asdemonstration of neutralization of TNF activity with antibody.

47). Several groups reported inhibition of the lytic activity afteraddition of certain 6'-phosphorylated sugars (monosaccha-

rides), suggesting presence of functional carbohydrate groupsor side chains (30, 38). Preliminary attempts to purify the factorindicated a broad molecular weight range with molecularweights of 15,000-40,000 (47), M, 18,000-22,000 (23), M,20,000-40,000 (30), and M, 20,000-25,000, 18,000, and37,000, and 40,000-45,000 (31).

Our results, in concordance with earlier findings, show amodest level of direct K562 lysis by crude lymphocyte-conditioned medium (Table 1), but a substantially increased specificlysis after partial purification. The recovered bioactivity of alleluted fractions from the AcA-54 gel filtration column showedtwo clear peaks corresponding to molecular weights of approximately 33,000-43,000 and approximately M, 5,000.

Ransom el al. reported the differences between NKCF andLeukoregulin, which is generated in a fashion similar to NKCF(17). Purified LRG has a molecular weight of 135,000, anddissociates in M, 32,000 units under reducing conditions. It isprobably produced by both T-cells and LGL, but it is distinctfrom NKCF in that it is cytostatic, rather than cytolytic, has aslower kinetic and a different target cell spectrum (17, 18).Ortaldo et al. were the first to compare crude NKCF with rh-TNF and lymphotoxin (48). They demonstrated clear differences between these factors in biological assays as well as inantibody neutralization studies (48).

We here confirm their findings and extend them by using896




partially purified NKCF and other target cell lines, such as SK-MEL 30, and other assay systems, including the LAK assay.Our results demonstrate that rh-TNF and lymphotoxin do notlyse K562 cells or SK-MEL 30 cells, even in very high concentrations. Even though it cannot be ruled out that small amountsof TNF or lymphotoxin are present within the NKCF preparation, these factors do not contribute to NK-cell-mediatedcytotoxicity. This, too, is in concordance with recent reportsfrom others (49). On the other hand, Wright and Bonavidareported that TNF, but not lymphotoxin, might contribute toNK-cell killing. They found a similar target specificity of NKCFand TNF and obtained a partial inhibition of NKCF activity byanti-TNF antibodies (50). However, this was observed in onlyhalf of the cases studied and also tested with crude NKCFsupernatants. It is known that crude cytotoxin preparationsmay contain TNF, lymphotoxin, NKCF, IL-1, and interferons(51), depending on individual (donor) cell types and inducingagents. While both crude and partially purified NKCF (pools Iand II) are able to lyse L929 cells, the L-cell assay is not specificfor TNF or lymphotoxin. Furthermore, L929 cells are alsosusceptible to NK-cell lysis (data not shown).

The observation that the TNF effect on SK-MEL 30 cellsfrequently did not correlate in the two assay systems, chromium-release assay and MTT immunoassay, employed in ourstudies (Fig. 2, A and B), can be explained by the two differentendpoints of these assays. The chromium-release assay is indicative of cell lysis, whereas the cleavage of MTT dye, and theconsequent color change reaction, depends on functional mitochondria! dehydrogenases in intact cells. Thus, this assaydoes not necessarily indicate cell death or lysis. As a cytopathicassay, the MTT assay is extremely valuable and it often correlates with cytostasis and, occasionally, cytolysis. This does notapply, however, to every cell type or tumor target cell, with SK-MEL 30 being one example of absence of such a correlation.This has also been observed by Ruggiero and coworkers, whoreported the induction of a protective molecular mechanism inmelanoma cells after exposure to rTNF, but not in other tumorcell lines tested (5).

Additional support for the distinction of NKCF from TNFand lymphotoxin comes from our studies reported here on theeffect of NKCF on CMC. Experiments with crude NKCFrevealed both enhancement of NK-cell-mediated lysis (Fig. 5)and LAK-cell-mediated lysis (Table 4). In summary, our resultsindicate that only NKCF (pool I), but none of the other factorstested, enhanced NK- (Table 3; Fig. 6) and LAK-cell-mediatedlysis (Table 4). The lower molecular weight pool NKCF (poolII) was always ineffective, showing identical lytic values aseffector cells alone (NK or LAK control). Pool II was active inlong-term factor killing assays (20 or 48 h), suggesting that thetwo pools contain factors with distinct biological characteristics. The enhancement of NK-CMC by NKCF (pool I) againstK562 was seen in every donor tested, regardless of donor PBMCbaseline NK activity. Moreover, SK-MEL 30, a relatively resistant target to fresh NK-lysis, was also almost always killedto an extent not obtainable in the absence of NKCF (pool I).

Since other groups used crude NKCF preparations, this isthe first report describing the separation of two distinct cytolyticactivities of different molecular weights and biological effects.This is an initial description of enhancement of NKCF onCMC, with augmentation of both NK- and LAK-cell lysis. Thefinding that neither pool I nor pool II of partially purifiedNKCF could be neutralized with anti-TNF monoclonal antibody and, in addition, that the enhancing activity of NKCF onNK-CMC was not blocked by anti-TNF moAbs, further sug

gests that NKCF is a distinct cytotoxic factor, acting differentfrom TNF and lymphotoxin.

These results provide new support for the involvement ofNKCF in NK-CMC and for the distinction of these activitiesfrom other cytotoxic molecules. It can be concluded that TNF,lymphotoxin, and NKCF are different cytokines and that TNFand lymphotoxin do not play a major role in NK cytoxicity.Final elucidation of the functional roles and the mechanismsof action of both NKCF activities should allow from effortsdirected at their purification to homogeneity and molecularcloning.

ACKNOWLEDGMENTS

We would like to thank Dr. Alan Houghton for providing themelanoma cell line SK-MEL 30, Barbara Williamson for providing theTNF-resistant L929 subclone, Dr. Lilian Reich and the staff of theMemorial Hospital Blood Donor Room for providing the buffy-coatpreparations, Dr. SolÃanGupta and Dr. Gheeta Sharma for the cooperation with the interferon assay, and Loma Barnett and Azita Mair-zadeh for their excellent technical assistance.

REFERENCES

1. Aggarwal, B., Moffat. B.. and Harkins, R. N. Human lymphotoxin. Production by a lymphoblastoid cell line, purification, and initial characterization.J. Biol. Chem.. 259: 686-691, 1984.

2. Pennica, D.. Nedwin, G. E., Hayflick, J. S., Seeburg, P. S., Derynck, R.,Palladino, M. A., Kohr. W. J., Aggarwal, B. B., and Goeddel, D. V. Humantumor necrosis factor: precursor structure, expression and homology tolymphotoxin. Nature (Lond.), 312: 724-729, 1984.

3. Rubin, B. Y., Anderson. S. L., Sullivan, S. A.. Williamson, B. D., Carswell,E. A., and Old, L. J. Purification and characterization of a human tumornecrosis factor from the LuKII cell line. Proc. Nati. Acad. Sci. USA, 82:6637-6641, 1985.

4. Old, L. J. Tumor necrosis factor. Science (Wash. DC), 230: 630-632, 1985.5. Ruggiero. V.. Latham. K.. and Baglioni, C. Cytostatic and cytotoxic activity

of tumor necrosis factor on human cancer cells. J. Immunol.. 138: 2711-2717. 1987.

6. Uchida, A., and Klein, A. Natural cytotoxicity of human blood monocytesand natural killer cells and their cytotoxic factors: discriminating effects ofactinomycin D. Int. J. Cancer, 35: 691-699. 1985.

7. Uchida, A., Vanky. F., and Klein, E. Natural cytotoxicity of human bloodlymphocytes and monocytes and their cytotoxic factors: effect of interferonon target cell susceptibility. J. Nail. Cancer Inst., 75: 849-857, 1985.

8. Granger, G. A., Orr. S. L.. and Yamamoto. R. S. Lymphotoxins, macrophagecytotoxins, and tumor necrosis factor: an interrelated family of antitumoreffector molecules. J. Clin. Immunol.. 5: 217-219, 1985.

9. Podack, E. R., and Dennert. G. Assembly of two types of tubules withputative cytolytic function by cloned natural killer cells. Nature (Lond.), 302:442-445, 1983.

10. Young. J. D., Hengartner. H., Podack, E. R., and Cohn, Z. V. Purificationand characterization of a cytolytic pore-forming protein from granules ofcloned lymphocytes with natural killer activity. Cell, 44: 849-859, 1986.

11. Tschopp, J., Masson, D., and Stanley, K. K. Structural/functional similaritybetween proteins involved in complement- and cytotoxic T-lymphocyte-mediated cytolysis. Nature (Lond.), 322: 831-834, 1986.

12. Tschopp, J.. Masson. D.. and Schaefer. S. Inhibition of the lytic activity ofperforin by lipoproteins. J. Immunol.. 137: 1950-1953. 1986.

13. Young, J. D., Cohn, Z. A., and Podack, E. The ninth component of complement and the pore-forming protein (Perforin 1) from cytotoxic T cells:structural, immunological. and functional similarities. Science (Wash. DC),233: 184-190. 1986.

14. Blumenthal, R., Millard, P. J., Henkart, M. P., Reynolds, C. W., andHenkart, P. A. Liposomes as targets for granule cytolysin from cytotoxiclarge granular lymphocyte tumors. Proc. Nati. Acad. Sci. USA. */: 5551-5555, 1984.

15. Henkart. P. A. Mechanism of lymphocyte-mediated cytotoxicity. Ann. Rev.Immunol., Â¿.-31-58, 1985.

16. Henkart. P. A., Yue, C. C., Yuang, J., and Rosenberg. S. A. Cytolytic andbiochemical properties of cytoplasmatic granules of murine lymphokine-activated killer cells. J. Immunol.. 137: 2611-2617, 1986.

17. Ransom, J. H., Evans, C. H., McCabe, R. P., Pomato, N., Heinbaugh, J. A.,Chin, M., and Hanna, M. G. Leukoregulin, a direct-acting anticancer immunological hormone that is distinct from lymphotoxin and interferon.Cancer Res., 45: 851-862, 1985.

18. Sayers, T. J., Ransom. J. H., Denn HI, A. C., Herbermann, R. B., andOrtaldo. J. R. Analysis of a cytostatic lymphokine produced by incubationof lymphocytes with tumor cells: relationship to leukoregulin and distinctionfrom recombinant lymphotoxin, recombinant tumor necrosis factor, and

897




natural killer cylotoxic factor. J. luminimi.. 137: 385-390, 1986.19. Lichtenstein, A., Ganz, T., Selsted, M. E., and Lehrer, R. I. In vitro tumor

cell cytolysis mediated by peptide defensins of human and rabbit granulocytes.Blood, 68: 1407-1410. 1986.

20. Pasternack, M. S., Verrei, C. R., Liu, M. A., and Eisen. H. N. Serine esterasein cytolytic T lymphocytes. Nature (Lond.), 322: 740-743, 1986.

21. Schmidt, R. E., MacDermott, R. P., Hartley, G., Bertovich, M., Amato, D.A., Austen, K. F., Schlossmann, S. F., Stevens, R. L., and Ritz, J. Specificrelease of proteoglycans from human natural killer cells during target lysis.Nature (Lond.), 318: 289-291, 1985.

22. Tschopp, J., and Conzelmann, A. Proteoglycans in secretory granules of NKcells. Immunol. Today, 167: 135-136, 1986.

23. Ortaldo, J. P.., and Reynolds, C. Summary of the second internationalworkshop on mechanisms of cell-mediated cytotoxicity. Nat. Immun. CellGrowth Regul., 4: 2-6. 1985.

24. Ortaldo. J. R., Mason, L. H., Mathieson, B. J., Liang, S-, Flick. D. A., andHerbermann, R. B. Mediation of mouse natural cytotoxic activity by tumornecrosis factor. Nature (Lond.). 321: 700-702, 1986.

25. Marx, J. L. How killer cells kill their targets. Science (Wash. DC), 231:1367-1369, 1986.

26. Podack, E. R. The molecular mechanism of lymphocyte-mediated tumor celllysis. Immunol. Today, 6: 21-27. 1985.

27. Feinman. R., Henrikson-DeStefano. D., Tsujimoto. M., and Vilcek, J. Tumornecrosis factor is an important mediator of tumor cell killing by humanmonocytes. J. Immunol., 138:635-640. 1987.

28. Wright. S., and Bonavida. B. Selective lysis of NK-sensitive target cells by asoluble mediator released from murine spleen cells and human peripheralblood lymphocytes. J. Immunol., 126: 1516-1523, 1981.

29. Wright. S. C.. and Bonavida, B. Studies on the mechanism of natural killer(NK) cell-mediated cytotoxicity (CMC). I. Release of cytotoxic factors specific for NK-sensitive target cells (NKCF) during co-culture of NK effectorcells with NK target cells. J. Immunol., 129: 433-441, 1982.

30. Bianca, L, Herbermann, R. B.. and Ortaldo. J. R. Human natural killercytotoxic factor. Nat. Immun. Cell Growth Regul.. 4: 48-59, 1985.

31. Callewaert, D. M. Natural cytotoxicity: recent progress and continuingcontroversy. Nat. Immun. Cell Growth Regul., 4: 61-77, 1985.

32. Deem, R. L., and Targan, S. R. Sequential substages of natural killer cell-derived cytolytic factor (NKCF)-mediated cytolysis as defined by glutaralde-hyde modulation of the target cell. J. Immunol.. 133: 1836-1840. 1984.

33. Degliantoni. G.. Murphy. M., Kobayashi, M., Francis, M. K., Perussia, B.,and Trinchieri. G. Natural killer (NK) cell-derived hematopoietic colony-inhibiting activity and NK cytotoxic factor. J. Exp. Med., 162: 1512-1530,1985.

34. Farram, E., and Targan, S. R. Identification of human killer soluble cytotoxicfactor(s) (NKCF) derived from NK-enriched lymphocyte populations: specificity of generation and killing. J. Immunol., 130: 1252-1256, 1983.

35. Steinhauer, E. H.. Doyle. A. T.. and Kadish, A. S. Human natural killercytotoxic factor (NKCF): role of IFN-a. J. Immunol., 135: 294-299, 1985.

36. Steinhauer. E. H., Doyle, A. T., Reed, J., and Kadish. A. S. Natural killercytotoxic factor induction by K562 cells in patients with advanced cancer:correlation with production of interferon. J. Nati. Cancer Inst., 75: 1017-1023. 1985.

37. Wright, S. C., and Bonavida, B. Studies on the mechanism of natural killercell-mediated cytotoxicity. V. Lack of NK specificity at the level of inductionof natural killer cytotoxic factors in cultures of human, murine, or rat effectorcells stimulated with mycoplasma-free cell lines. J. Immunol., 133: 3415-3423, 1984.

38. Wright, S. C., and Bonavida. B. Studies on the mechanism of natural killercell-mediated cytotoxicity. VI. Characterization of human, rat. and murinenatural killer cytotoxic factors. Nat. Immun. Cell Growth Regul., 4: 202-220, 1985.

39. Graves, S. S., Bramhall, J., and Bonavida, B. Studies on the lethal hit stageof natural killer cell-mediated cytotoxicity. I. Both phorbol ester and iono-phore are required for release of natural killer cytotoxic factors (NKCF),suggesting a role for protein kinase C activity. J. Immunol., 137:1977-1984,1986.

40. Boyum. A. Separation of leukocytes from blood and bone marrow. Scand. J.Clin. Lab. Invest.. 2/(Suppl. 197): 1-13, 1968.

41. Timonen, T., Ortaldo, J. R., and Herbermann, R. B. Characteristics of humanlarge granular lymphocytes and relationship to natural killer and K cells. J.Exp. Med., 153: 569-582. 1981.

42. Mosmann, T. Rapid colorimetrie assay for cellular growth and survival:application for proliferation and cytotoxicity assays. J. Immunol. Methods,Â«5:55-63, 1983.

43. Pross, H. F., and Maroun, J. A. The standardization of NK cell assays foruse in studies of biological response modifiers. J. Immunol. Methods., 68:235-249, 1984.

44. Herbermann. R. B., Djeu. J. Y., Kay, D. H., Ortaldo, J. R., Riccardi, C.,Bonnard, G. D., Holden, H. T., Fagnani, R., Santoni, A., and Puccetti, P.Natural killer cells: characteristics and regulation of activity. Immunol. Rev.,44: 43-70, 1979.

45. Burns, G. F., Begley, C. G., Mackay. I. R., Triglia, T., and Werkmeister, J.A. Supernatural killer cells. Immunol. Today, 6: 370-373, 1985.

46. Shau, H.. and Golub. S. H. Signals for activation of natural killer and naturalkiller-like activity. Nat. Immun. Cell Growth Regul., 4: 113-119, 1985.

47. Bonavida, B., and Wright, S. Role of natural killer cytotoxic factors in themechanism of target-cell killing by natural killer cells. J. Clin. Immunol., 6:1-8. 1986.

48. Ortaldo, J. R., Ransom, J. R., Sayers, T. J., and Herbermann, B. Analysis ofcytostatic/cytotoxic lymphokines: relationship of natural killer cytotoxicfactor to recombinant lymphotoxin, recombinant tumor necrosis factor, andleukoregulin. J. Immunol.. 137: 2857-2863, 1986.

49. Peters, P. P., Ortaldo, J. R., Shalaby, M. R., Svedersky, L. P., Nedwin, G.E.. Bringman, T. S., Hass, P. E., Aggarwal. B. B.. Herbermann. R. B..Goeddel. D. V., and Palladino. M. A. Natural killer-sensitive targets stimulate production of TNF-u but not TNF-rf (lymphotoxin) by highly purifiedhuman peripheral blood large granular lymphocytes. J. Immunol., 137:2592-2598, 1986.

50. Wright, S. C., and Bonavida, B. Studies on the mechanism of natural killercell-mediated cytotoxicity. VII. I mimmi,ij comparison of human naturalkiller cytotoxic factors with recombinant lymphotoxin and tumor necrosisfactor. J. Immunol.. 138: 1791-1798, 1987.

51. Wallach, D. Cytotoxins (tumour necrosis factor, lymphotoxin and others):molecular and functional characteristics and interactions with interferons.in: I. Gresser (ed.), Interferon 7. London: Academic Press, 1986.

898



1988;48:891-898. Cancer Res Tammo Bialas, Jonathan Kolitz, Ester Levi, et al. Recombinant Human LymphotoxinFactor from Recombinant Human Tumor Necrosis Factor and Distinction of Partially Purified Human Natural Killer Cytotoxic

Updated version

http://cancerres.aacrjournals.org/content/48/4/891

Access the most recent version of this article at:

E-mail alerts related to this article or journal.Sign up to receive free email-alerts

Subscriptions

Reprints and

[email protected] at

To order reprints of this article or to subscribe to the journal, contact the AACR Publications

Permissions

[email protected] at

To request permission to re-use all or part of this article, contact the AACR Publications


http://cancerres.aacrjournals.org/content/48/4/891

http://cancerres.aacrjournals.org/cgi/alerts

mailto:[email protected]

mailto:[email protected]


distinction of partially purified human natural killer ... · pbl- or idi conditioned media...

Documents