AIDS RESEARCH AND HUMAN RETROVIRUSESVolume 8, Number 5, 1992Mary Ann Liebert, Inc., Publishers

Functional Activity of an HIV-1 Neutralizing IgG HumanMonoclonal Antibody: ADCC and Complement-Mediated Lysis

MARSHALL R. POSNER, HILLARY S. ELBOIM, THOMAS CANNON, LISA CAVACINI, andTERU HIDESHIMA

ABSTRACT

The IgG,K, human monoclonal antibody (HMAb), F105, was studied for functional activity in antibody-dependent cellular cytotoxicity (ADCC) and complement-dependent cytotoxicity (CDC). F105 reacts with a

discontinuous epitope on the CD4 binding site of the HIV-1 envelope glycoprotein, gpl20, expressed on thesurfaces of infected cells and neutralizes diverse viral strains at antibody concentrations readily achievable inhumans. Neither F105 nor serum (diluted 1:50) from HIV seropositive donors mediate CDC against an

SF2-infected cell line with rabbit or human sera as a source of complement. F105 and HIV-1 sera mediate ADCCagainst the SF2 strain. Normal human serum reduced spontaneous lysis of SF2 by peripheral blood monocytes(PBM). Although mixing of F105 with normal human serum reduced the lysis observed (36 ± 8 vs. 42 ± 8%),this still was significantly greater than lysis in media (30 ± 5%) or normal human serum (23 ± 6%) (p < .05).A murine antibody to CD16 significantly reduced spontaneous lysis observed with media (30 ± 5 vs. 18 ± 3%)while normal mouse serum had no effect (31 ± 7%). ADCC mediated by F105 is completely abrogated by theanti-CD16 antibody (42 ± 8 vs. 22 ± 4%), while only a fraction of ADCC mediated by HIV sera is inhibited byanti-CD16 (60 ± 9 vs. 46 ± 6%), suggesting that several populations of effector cells function in ADCC mediatedby the polyclonal sera. Thus, F105, as opposed to polyclonal sera, mediates ADCC through a CD16+ PBMpopulation.

INTRODUCTION

AFTER INFECTION, HUMAN IMMUNODEFICIENCY VIRUS, Type1 (HIV-1) employs multiple mechanisms to evade elimina-

tion and/or control by the human immune system.'2 Humoraland cell-mediated responses to the virus fail to prevent establish-ment and eventual progression of the disease in the vast majorityof cases. The human antibody response to HIV-1 can neutralizefree virus and eradicate infected cells in vitro; however, therelevance of antibodies in promoting these processes in vivo hasbeen controversial. ' 6 Both neutralizing and functionally activeantibodies are present in seropositives in whom the diseaseeventually progresses, suggesting a marginal role for antibody incontrolling disease. In support of a role for antibodies, function-ally effective antibodies are associated with improved survival;

and the loss of neutralizing and functional antibodies occurs withdisease progression.5~K Furthermore, protection from verticaltransmission in maternal-fetal pairs may be associated withfunctionally effective antibodies.9

Study of the functional activity of human antibodies has beenlimited because of the polyclonal nature of the responses;because of the diverse nature of the target epitopes; and becausexenogenic sera and xenogenic monoclonal antibodies do notadequately model human antibody responses in this dis-ease.10"13 Differences exist between the functional activity ofthe human responses to this primate-restricted virus and thoseelicited in other species. For example, xenogenic sera, nonhu-man primate sera, and xenogenic monoclonal antibodies to

gp 120 can mediate complement-dependent cellular cytotoxicity(CDC) while human serum antibodies fail to mediate this

Human Monoclonal Antibody Laboratory; the Department of Medicine; and Division of Hematology/Oncology. New England DeaconessHospital, and the Harvard Medical School. Boston, MA 02215.

553

554 POSNER ET AL.

function.'213 Evaluation of the relevant human, humoral im-mune response will be facilitated by the study of humanmonoclonal antibodies (HMAb) reacting with diverse targets on

the viral envelope.1""In this study we evaluated the functional activity of an HMAb,

F105. F105 reacts with a discontinuous epitope on the HIV-1envelope glycoprotein, gpl20. The F105 epitope includes por-tions of the CD4 binding site and F105 inhibits CD4/gpl20binding.1"14 F105 neutralizes diverse viral strains at antibodyconcentrations readily achievable in humans. The F105 epitopeis highly conserved among HIV-1 strains tested and is expressedby different strains in varying densities on the surfaces ofinfected cells where it can serve as a target for immune attack.Because FI05 is an IgG,K antibody it should mediate antibody-dependent cellular cytotoxicity (ADCC) and CDC.

MATERIALS AND METHODS

Cell culture, cell lines, reagents, andmonoclonal antibodies

Cell lines and established hybridomas were grown as previ-ously described.1" The origins of the FI05 IgG HMAb havebeen described. '"F105 was purified by affinity chromatographyon a protein G Column (Pharmacia, Upsala. Sweden) andquantified. P6A3 is a murine monoclonal antibody (MMAb)which reacts with all nucleated human cells tested and serves as

a complement-positive control (Posner. personal communica-tion). Anti-CD16 MMAb (CDI6) was obtained from spentsupernatant from the B73.1 hybridoma, kindly provided byGiorgio Trinchieri (Wistar Institute, Philadelphia. PA).15 Anti-human IgM MMAb was obtained from spent supernatant of theDA4-4 hybridoma (HB57, ATCC, Rockville. MD).

HUT-78/HIV-ls[,2 (SF2) from Dr. Jay Levy, H9/HTLV-IIIRFNIH1983 (RF) from Dr. Robert Gallo, and H9/HTLV-IIIMN(MN) NIH 1984 from Dr. Robert Gallo were provided by theAIDS Research and Reference Reagent Program. AIDS Pro-gram, NIAID, NIH (ERC Bioservices Corp. Rockville, MD).HIVMIlJ infected HT-H9 cells were obtained from the ATCC(ATCC. Rockville, MD). CC, a Haitian strain of HIV-1. was

kindly provided by Dr. Peter Nara (NIH, Bethesda, MD). '"'6

Peripheral blood mononuclear (PBM) effectors were obtainedfrom normal, séronégative, volunteers by density gradient cen-

trifugation. Normal sera and sera from HIV-1-infected patientswith acquired immunodeficiency syndrome (AIDS) or AIDS-related complex (ARC) were obtained voluntarily. Sera was

stored at -80°C.

Indirect immunofluorescence of cell surface antigensIndirect immunofluorescence (IF) for the detection of HMAb

reactive with cell surface antigens was performed as reportedpreviously.'" In brief, cell lines were washed with phosphatebuffered saline (PBS) and 1 x 106 cells in 100 pJ were

transferred to test tubes. Aliquots of 100 pJ of appropriatelydiluted serum or monoclonal antibody were added to the cellsand incubated at 4°C for 30 minutes. The cells were then washed

twice with PBS. followed by the addition of 100 pJ of a 1:50dilution of fluorescein-conjugated F(ab')2 goat anti-human IgG(Tago, Inc.. Burlingame, CA) in Puck's Saline G (PSG) with2.5% fetal bovine serum. The sample was incubated for 30minutes at 4°C and then washed once in PBS. Following this, thepellet of live cells was fixed by resuspension in 500 pJ of 1%(v/v) formaldehyde. The fixed cells were resuspended andstored up to 5 days at 4°C until analysis. Analysis was performedon a FACScan (Becton-Dickinson. Lincoln Park, NJ). Negativecontrol sera diluted 1:100, from normal volunteers were used toestablish a background fluorescence of 5-7% positive cells, andsimilarly diluted sera from HIV patients served as positivecontrols.

Antibody-directed cellular cytotoxicity andcomplement-mediated cellular cytotoxicity

Target cells were prepared for ADCC and CDC by labelling1-10 x 10" cells in 1 ml of PSG with 250 p.Ci of Crsl as

chromate (ICN, Costa Mesa. CA) for 1 h at 37°C and washingthree times in cold PSG. Cells were resuspended in regulargrowth media at 1 x lO'Vml and 50 pJ aliquoted per test well in96-well, V-bottom microtiter wells. Next, 50 pd of test antibodyor heat-inactivated sera (final dilution 1:50) from normal con-

trols or HIV-1 séropositive volunteers was added to the wells.All data points were performed in triplicate. A negative controlwell received media without added antibodies.

For CDC. 50 pd of newborn rabbit complement (Pel Freez,Brown Deer, WI) diluted to obtain a final concentration of 1:15in the wells was added. The wells were incubated at 37°C for 1 h,centrifuged at 80 g for 5 minutes, and 75 u.1 of supernatant was

transferred to plastic tubes and counted. For some experimentsplates were kept at 4°C until complement was added thenallowed to reach 37°C during 1 h of incubation. Spontaneouslysis was measured in test wells containing cells, media, andcomplement. Total chromium release was determined by substi-tuting 50 p.1 of 1% Triton-X for complement in one set of testwells. Specific release was calculated by the standard formula:({Experimental Release

—

Spontaneous}/{Total Release—Spontaneous Release}] x 100%. A positive control consisted of

a MMAb, P6A3, which fixes complement and reacts with allnucleated human cells tested.

For ADCC, PBM from normal healthy donors were resus-

pended at an appropriate concentration and added to the wells in50 u.1 of complete media such that the effector to target ratio (E:Tratios) was 1:10 or 1:20. The plates were centrifuged at room

temperature and 80 g for 5 minutes and incubated for 4 h at 37°C.After incubation the plates were centrifuged and aliquots re-

moved and counted as above. Specific release was calculated bythe same formula; however, spontaneous release was defined as

release by the labelled target cells without the addition ofeffectors.

For certain experiments PBM were premixed with spentmedia containing MMAb or nonspecific normal mouse serum

before being added, without washing, to wells for ADCC.Statistical analysis of data was performed using the Wilcoxonsigned-ranks test.

ADCC AND CDC WITH A HUMAN MONOCLONAL ANTIBODY TO HIV-l/gpl20 555

RESULTS

F105 reacts with gpl20 expressed on the surfaces of cellsinfected with diverse isolates of HIV-1. '"As shown in Figure 1,the most uniform F105/gpl20 reactivity, as determined by flowcytofluorometry, is seen on the SF, strain. The entire SF2population is reactive in an almost normal distribution whencompared with other strains tested. Among these strains, reac-

tivity is more variable within the population of infected cells andfluctuates with time and culture conditions, while SF2 expres-sion remains relatively uniform and stable (data not shown). Forthese reasons SF2 was chosen as the target for subsequent studiesof ADCC and CDC, except as noted below. The concentration-dependent cell surface reactivity of F105 with SF2 is shown inFigure 2. As can been seen, maximal F105 reactivity occurs atconcentrations 3= 10 p,g/ml. F105 was used in subsequent exper-iments at a concentration of 25 p.g/ml, except where indicated.

FI05-mediated CDC against SF2 was examined using rabbitserum as a source of complement. As can be seen in Table 1,neither FI05 nor serum (diluted 1:50) from several HIV sero-

positive donors mediated CDC against the SF2 cell line at 37CC.As a control for complement function P6A3 effectively medi-ated CDC of these cells. Increasing the concentration of F105 to100 p,g/ml or substituting fresh normal human serum (1:5-1:25)for rabbit serum as a source of complement had no effect (datanot shown). Similar results were obtained with HIVI1IB as a

FIG. 1. Histograms of the reactivity of F105 and normalhuman serum with cell lines infected with HIV-1 strains mea-

sured by indirect immunofluorescence with flow cytometry. TheX axis is log green fluorescence of the cells and the y axis is thenumber of positive cells. The strain tested in each histogram isgiven in the upper right-hand corner of the graph.

0.025 0.100 1.000 10.000F105 CONCENTRATION (ugm/ml)

FIG. 2. The dose-dependent reactivity of F105 with SF2 cellsby flow cytometry. The X axis is F105 concentration and the Yaxis is percent of cells with fluorescence.

target. Allowing the lytic reaction to take place over a tempera-ture range from 4-27°C also did not result in lysis of cells.

In contrast to the results seen with CDC. F105 mediatesADCC against the SF2 strain (Fig. 3). Compared with poly-clonal HIV sera, F105 induced a smaller but significant increasein lysis over that observed with media (p < .05). ADCC withF105 and HIV sera increases with increasing effector to targetratios (E:T). An E:T of 20:1 was chosen for subsequent experi-ments because of the consistent reproducibility and separation ofbackground in experimental lysis at this E:T. Of importance,normal human serum (NHS) reduces spontaneous lysis of SF2 byPBM. Although not significant at low E:T, NHS inhibitionbecomes significant at higher E:T where separation from back-ground lysis is more readily observed (p < .05).

As would be expected, F105-mediated ADCC is concentra-tion dependent (Fig. 4). The relative amount of lysis mediatedby F105 correlates with the dose-dependent saturation of SF2binding sites by F105 shown in Figure 2. Half maximal lysisappears at approximately 5 p,g/ml, which is similar to theconcentration for half maximal cell surface reactivity of F105with SF2.

Because NHS appears to inhibit spontaneous lysis, the effectof NHS on Fl()5-mediated ADCC was tested by comparingADCC with media, F105, NHS, and F105 plus NHS. As seen inFigure 5, although mixing of F105 with NHS reduces theabsolute lysis observed (36 ± 8 vs. 42 ± 8%), F105-mediated

Table 1. Complement-Mediated Lysis of SF, with F105

Conditions Lysis (%)MediaP6A3 MMAbNHSF105 HMAbHIV sera

3.4 ± I59.1 ±3

5.0 ± 28.6 ± 35.0 ± 2

The results of four experiments with rabbit complement-mediated cellular cytotoxicity with media, normal serum. FI05antibody, and sera from HIV-1 seropositives expressed as themean percent lysis ± SEM.

556 POSNER ET AL.

in>-

oenUJ

<>X

FIG. 3. ADCC by PBM from normal volunteers against SF2cells treated with media, normal serum diluted 1:50 (NS), F105at a concentration of 25 p-g/ml, and serum from HIV-1-infectedvolunteers diluted to l:50(HIV.S). Two effector to target ratios,10:1 and 20:1 are presented and each bar represents the mean

lysis ± the SEM from 5 (E:T 10:1) or 7 (E;T 20:1) separateexperiments. * Indicates a significant difference (p < .05)between the experimental condition and the results obtainedwith media.

in

MEDIA

FIG. 5. F105 mediates ADCC in the presence of normalserum. This was tested with media, normal serum (NS) 1:50,F105 25 p-g/ml, and F105 and normal serum mixed to a finalconcentration of 25 u.g/ml and 1:50, respectively. Each barrepresents the mean lysis ± the SEM from four separateexperiments. * Indicates a significant difference (p < .05)between the experimental condition and the results obtainedwith media.

lysis was still significantly greater than that occurring withmedia (30 ± 5%) or in NHS alone (23 ± 6%) (p < .05). Thus,F105 mediates ADCC, and does so to the same relative degree inthe presence or absence of NHS. F105 had no effect on ADCCmediated by HIV serum (data not shown).

To determine the cells responsible for ADCC with the F105antibody and HIV serum, we employed the anti-CDI6 antibodyB73.1 (anti-CD16) to block the function of NK cells.15 Anti-CD 16 significantly reduced the spontaneous lysis observed withmedia alone (30 ± 5 vs. 18 ± 3%) (Fig. 6). An anti-IgMantibody that reacts with B cells (DA4-4) also reduced sponta-neous lysis slightly (24 ± 4%), while normal mouse serum

(NMS) had no effect (31 ± 7%). DA4-4 was chosen as a coldtarget control antibody for CD 16 because it is reactive with a

small population of cold targets (IgM+ B cells) in PBM. The

100

en

3

size of the IgM + population in the PBM of normals is relativelyequivalent to that of the CD16+ population (data not shown).Furthermore, the inhibition of lysis observed with DA4-4 is theresult of cold target competition for effectors, because a normal,nonspecific murine serum serving as a second negative controldoes not inhibit lysis. This result also reveals that murineimmunoglobulins are ineffective in reducing spontaneous lysisby human effectors, which contrasts with NHS. Spontaneouslysis in NHS follows the same pattern of effect as media with theadded murine antibodies. Of importance, ADCC mediated byF105 is abrogated completely by the anti-CD 16 antibody (42 ±

0.100 1.000 10.000F105 CONCENTRATION (ugm/ml)

100.000

FIG. 4. Dose-dependent specific lysis of SF2 cells in ADCCmediated by F105. The X axis is F105 concentration and the Yaxis is percent of maximal specific lysis.

c/i>-

oer

MEDIA NORMALSERA

F105 HIVSERA

FIG. 6. ADCC by F105 is mediated by a CD16+ populationof cells. Effector cells were premixed with media, anti-CD16antibody (CD 16), anti-IgM antibody (DA4-4), or normal mouseserum (1:50) and then added, without washing to the targets.Each bar represents the mean lysis ± the SEM froi 1 separateexperiments. Significant differences (p < .05) between theexperimental conditions are described in the text.

ADCC AND CDC WITH A HUMAN MONOCLONAL ANTIBODY TO HIV-l/gpl20 557

8 vs. 22 ± 4%). While F105 mediates ADCC through a CD16 +

PBM population, only a portion of ADCC mediated by HIV sera

is inhibited by anti-CD16 (60 ± 9 vs. 46 ± 6%).

DISCUSSION

We have described studies of the in vitro functional activity ofF105, a neutralizing human monoclonal antibody reactive with a

discontinuous epitope on the CD4 binding site of HIV-l/gpl20.The derivation and reactivity of F105 with HIV-1/gp 120 havebeen described.'" As an IgG,K human antibody, F105 shouldefficiently mediate CDC and ADCC. Although SF2-infectedcells were lysed in CDC with a cell surface-reactive MMAb,neither the HMAb F105, nor HIV-1 + sera, were capable ofmediating CDC under warm or cold conditions. These resultsare consistent with those reported by Nara et al., who demon-strated a lack of CDC for infected cells by positive human sera incomparison to effective antibodies produced by infected or

immunized nonhuman primates or immunized rabbits andgoats. '2 Others also have shown a failure of CDC with humansera on gp 120-coated CD4+ cells. '3 The basis for this differencebetween human and xenogenic sera remains speculative. Warmand cold reactive lymphocytotoxins and HMAb fix xenogeniccomplement readily in other antigen systems; it is unlikely,therefore, that the failure in HIV-1-infected cells to obtain CDCwith human anti-HIV antibodies results from poor interaction ofhuman antibodies with xenogenic complement components.17It is possible that HIV-1 envelope proteins possess comple-ment regulatory functions, or, when exposed on the surfacesof cells, interact with cellular components and human anti-bodies to modify specifically complement activity directed at

gpI20 and gp41.'K The failure of human antibodies to me-

diate CDC in animal serum is a distinct difference betweenhuman and xenogenic antibody interactions with HIV-1 infectedcells.

Antibodies from the sera of HIV-1-infected patients can

mediate substantial amounts of ADCC. s'"x It has been suggestedthat ADCC, independent of total antibody reactivity, correlateswith disease stage in seropositive patients, although this remainscontroversial.5-7 Antibody-mediating ADCC activity occasion-ally remains present at high levels despite progression of thedisease, and loss of ADCC activity may be more a reflection ofthe further dysregulation of the immune system rather than a

contributing cause of progression.6 Nonetheless, evidence con-

tinues to support the notion that antibodies that mediate ADCCeffect the course and transmission of HIV-1 infection.9 SerumADCC activity appears to be mediated through antienvelopeIgG, antibodies, the predominant isotype in the immune re-

sponse to HIV-1 envelope proteins. I9~2' Previous studies of thefunctional activity of HMAb in ADCC have used IgG2 antibod-ies to gp41 rather than IgG, antibodies or antibodies reactivewith gpl20." ADCC in these studies and the present studyquantitatively correlate with cell surface reactivity of the anti-bodies. Of importance in future studies of HMAb, studies withMMAb have shown that antibodies of identical isotype, with cellsurface reactivity to related linear epitopes on HIV- 1/gpl 20, andwith equivalent affinity can differ in the functional mediation ofADCC such that one antibody is highly effective and the second

completely ineffective.22 Although the reasons for such dispar-ate results are speculative, isotype and substantial cell surfacereactivity do not assure functional activity of a HMAb in ADCC.

Although ADCC is readily apparent in these assays, naturalkiller cell (NK) and nonspecific lysis also were present. Al-though this activity limited the E:T ratios to 10:1 and 20:1 in thissystem compared with those E:T ratios used with CEM.NKRtargets, the higher E:T ratios were not necessary to observespecific lysis. ' ' An important previously unreported note is thatnonspecific lysis by PBM of infected cells was inhibited by thepresence of NHS at low concentration (2%). This reduction isspecific to NHS, as NMS had no effect. A small residualcomponent of lysis in NHS and media controls resulting fromNK activity could be demonstrated and further reduced byblocking the function of the CD16/Fc receptor, FcRIII, with theMMAb B73.1, an antibody reactive with CD16, expressedpredominantly on NK cells.15

Inhibition of nonspecific lysis with NHS did not eliminateFI05-mediated ADCC. The difference between lysis in thepresence of NHS and in F105/NHS was significant, specific,and equivalent to the difference between media and F105. SinceF105-mediated ADCC was not inhibited by NHS, the HMAbpossibly may mediate this effect in vivo.

F105-mediated ADCC was inhibited completely and specifi-cally by treatment of the PBM with anti-CD 16 MMAb, whileADCC mediated by HIV sera was only modestly inhibited. Thiseffect on F105-mediated ADCC was not the result of cold targetinhibition because a cell surface reactive MMAb of the same

isotype only minimally reduced ADCC. Thus F105, mediatesADCC through the CD16+ population of PBM. FcRIII (CD 16)is a low affinity receptor for IgG and is found on NK cells.,5,23Others have suggested that CD16+ NK cells are the principaleffectors responsible for ADCC mediated by HIV-1 + sera.23'24Polyclonal serum antibodies from HIV-1 f volunteers in thesystem described here appear to mediate ADCC through a

variety of different cell populations and receptors, in compari-son to ADCC mediated by the HMAb F105. These results withsera contrast with those described by others; however, méthod-ologie differences may account for differing results. For exam-

ple, Koup et al. used a lytic vaccinia construct to infectEpstein-Barr virus- (EBV) transformed lymphocytes and ex-

press envelope proteins to evaluate ADCC.6 This system differsbiologically from studies with HIV-1-infected T cells, as above,or gpl20-coated T cells. In addition, others have used antibodyand complement to eliminate the putative CD16+ NK popula-tion without testing lysis of another target population such as theIgM+ B cells.6'24 As a result, activated complement compo-nents, immune complexes, and cellular debris on this system,may have interfered with ADCC. This contrasts with the use ofa blocking antibody, B73.1, to inhibit NK function as opposedto eliminating CD16f cells.15 Thus, antibodies in HIV-1 sera

appear to mediate ADCC through CD 16 and additional recep-tors, possibly on monocytes and CD16 cells.

The in vivo importance of antibody-mediated functions in theprevention or control of HIV-1 infection remains unclear.HMAb provide tools for evaluating components of the humoralimmune response in vitro as demonstrated above. The possibleuse of HMAb as therapeutic tools, either as directed therapy or

as a guide to targets for vaccine research, lend importance to

558 POSNER ET AL.

studying the different functional activities of these antibodiesand relating these to in vivo function. '

ACKNOWLEDGMENT

This work was supported in part by a grant from theNIAID:R01-AI26926.

REFERENCES

1. Nara PL, Garrity RR, and Goudsmit J: Neutralization of HIV-1 : a

paradox of humoral proportions. FASEB 1991;5:2437-2455.2. Koff WC and Hoth DF: Development and testing of AIDS vaccines.

Science 1988;241:426-432.3. Groopman JE, Benz PM, Ferriani R, Mayer K, Allan JD, and

Weymouth LA: Characterization of serum neutralization responseto the human immunodeficiency virus (HIV). AIDS Res HumanRetroviruses 1978;3:71-85.

4. Tremblay M. Numazaki K, Xuguang L, Gornitsky M, Hiscott J,and Wainberg MA: Resistance to infection by HIV-1 of peripheralblood mononuclear cells from HIV-1 infected patients is probablymediated by neutralizing antibodies. J Immunol 1990:145:2896-2901.

5. Rook AH, Lane HC, Folks T, McCoy S. Alter H, and Fauci AS:Sera from HTLV-III/LAV antibody-positive individuals mediateantibody-dependent cellular cytotoxicity against HTLV-III/LAV-infectedT cells. J Immunol 1987;138:1064-1067.

6. Koup RA, Sullivan JL, Levine PH. Brewsler F. Mahr A, MazzaraG, McKenzie S, and Panicali D: Anfigenic specificity of antibody-dependent cell-mediated cytotoxicity directed against human im-munodeficiency virus in antibody-positive sera. J Virol1989;63:584-590.

7. Ljunggren K, Bottiger B, Biberfeld G, Karlson A, Fenyo EM, andJondal M: Antibody-dependent cellular cytotoxicity-inducing anti-bodies against human immunodeficiency virus. J Immunol1987;139(7):2263-2267.

8. Tyler DS, Stanley SD, Nastala CA, Austin AA, Bartlett JA, StineKC, Lyerly K, Bolognesi DP, and Weinhold KJ: Alterations inantibody-dependent cellular cytotoxicity during the course ofHIV-1 infection. J Immunol 1990:144:3375-3384.

9. Ljunggren K. Moschese V, Broliden PA, Giaquinto C, Quinti I,Fenyo EM. Wahren B. Rossi P. and Jondal M: Antibodies mediat-ing cellular cytotoxicity and neutralization correlate with a betterclinical stage in children born to human immunodeficiency. J InfectDis 1990;161:198-202.

10. Posner MR. Hideshima T. Cannon T, Mukherjee M. Mayer KH,and Byrn RA: An IgG human monoclonal antibody that reacts withHIV-l/gpl20, inhibits virus binding to cells, and neutralizes infec-tion. J Immunol 1991;146(12):4325-4332.

11. Tyler DS. Stanley SD, Zolla-Pazner S, Gorny MK. Shadduck PP,Langlois AJ. Matthews TJ, Bolognesi DP, Palker TJ. and WeinholdKJ: Identification of sites within gpl41 that serve as targets forantibody-dependent cellular cytotoxicity by using human mono-

clonal antibodies. J Immunol 1990:145:3276-3282.

12. Nara PL, Robey WG, Gonda MA, Carter SG, and Fischinger PJ:Absence of cytotoxic antibody to human immunodeficiency virus-infected cells in humans and its induction in animals after infectionor immunization with purified envelope glycoprotein gp!20. Im-munology 1987;84:3797-3801.

13. Lyerly HK. Matthews TJ, Langlois AJ, Bolognesi DP, and Wein-hold KJ: Human T-cell Iymphotropic virus IIIB glycoprotein(gpl20) bound to CD4 determinants on normal lymphocytes andexpressed by infected cells serves as target for immune attack. ProcNati Acad Sei (USA) 1987;84:4601-4605.

14. Thali M. Olshevsky U. Furman C. Gabuzda D, Posner M, andSodroski J: Characterization of a discontinuous HIV-1 gpl20epitope recognized by a broadly neutralizing human monoclonalantibody. J Virol 1991;65:6188-6193.

15. Perussia B, Acuto O, Terhorst C. Faust J, Lazarus R, Fanning V,and Trinchieri G: Human natural killer cells analyzed by B73.1, a

monoclonal antibody blocking Fc receptor functions. J Immunol1983:130:2142-2148.

16. Layne SP, Merges MJ, Dembo M, Spouge JL. and Nara PL: HIVrequires multiple Gpl20 molecules for CD4-mediated infection.Nature 1990;346:277-279.

17. Yamada A. Cohen PL, and Winfield JB: Subset specificity ofantilymphocyte antibodies in systemic lupus erythematosus. ArthrRheum 1985;28:262-270.

18. Cooper NR: Complement evasions strategies of microorganisms.Immunol Today 1991;12:327-331.

19. Evans LA. Thomson-Honnebier G, Steimer K, Paoletti E, PerkusME. Hollander H. and Levy JA: Antibody-dependent cellularcytotoxicity is directed against both the gpl20 and gp41 envelopeproteins of HIV. AIDS 1989;3:273-276.

20. Ljunggren K. Broliden PA, Morfeldt-Manson L, Jondal M, andWahren B: IgG subclass response to HIV in relation to antibody-dependent cellular cytotoxicity at different clinical stages. Clin ExpImmunol 1988;73:343-347.

21. Khalife J, Guy B, Capron M, Kieny MP, Ameisen JC, MontagnierL, Lecocq JP, and Capron A: Isotypic restriction of the antibodyresponse to human immunodeficiency virus. AIDS Res HumanRetroviruses 1988;4:3-9.

22. Broliden PA, Ljunggren KA. Hinkula J. Norrby E, Akerblom L,and Wahren B: A monoclonal antibody to human immunodefi-ciency virus type 1 which mediates cellular cytotoxicity andneutralization. J Virol 1990:64:936-940.

23. Gergely J and Sarmay G: The two binding-site models of humanIgG binding Fey receptors. FASEB 1990:4:3275-3283.

24. Ojo-Amaize EA, Nishanian P, Keith Jr DE, Houghton RL. HeitjanDF, Fahey JL, and Giorgi JV: Antibodies to human immunodefi-ciency virus in human sera induce cell-mediated lysis of humanimmunodeficiency virus-infected cells. J Immunol 1987; 139:2458-2463.

Address reprint requests to:Marshall R. Posner, MD

Human Monoclonal Antibody LaboratoryNew England Deaconess Hospital

185 Pilgrim RoadBoston, MA 02215