Lysis of herpes simplex virus-infected targets

CELLULAR

IMMUNOLOGY

83, 262-270 (1984)

Lysis of Herpes Simplex Virus-Infected

Targets

II. Nature of the Effector Cells’ ROBERT L. HENDRICKS’

AND JOEL SUGAR

Departments of Ophthalmology and Microbiology & Immunology, of Illinois Medical Center, Chicago, Illinois 60612

University

Received June 29, 1983; accepted September 14. I983

Peripheral blood lymphocytes (PBL) from patients with herpessimplex virus (HSV)- 1 recurrences in the cornea only (Group I) exhibited reduced lysis of HSV-l-infected targets compared to PBL from patients with oral-facial and comeal HSV recurrences(Group II). The cytotoxic lymphocytes appeared to belong to a subpopulation of natural killer (NK-HSV) cells. Monoclonal antibodies to human lymphocyte differentiation antigens were used to define the surface phenotype of the NK-HSV cells. Most ofthe NK-HSV activity was mediated by lymphocytes expressingthe surface markers Leu-7+ (HNK-I), OKT3+ (pan T), OKMl+ (myeloid and NK), Leu-2- (cytotoxic/ suppressor T cell), and Leu-8- (regulatory T cell). In contrast, lysis of K562 cells (NK-K562) was mediated by lymphocytes bearing the surface phenotype Leu-7+, OKT3-, OKMl+, Leu-2+‘-, and Leu-8-. The low level of NK-HSV activity in PBL from Group I donors appeared to be due to active suppression by suppressor T lymphocytes. Depletion of Leu-2+ cells from PBL of Group I donors resultedin significant augmentation of NK-HSV activity. Similar treatment of PBL from Group II donors either had no effect or slightly diminished the NK-HSV activity. INTRODUCTION

Most humans have undergone a primary herpes simplex virus type 1 (HSV-1) infection by adulthood as indicated by the presence of serum antibody to HSV-1. Furthermore, the virus appears to be harbored in the sensory ganglia and periodically shed into the saliva, precomeal tear film, and other mucus secretions of both symp tomatic and asymptomatic individuals (1). The factors responsible for maintaining HSV infections at a subclinical stageare poorly understood. The severedisseminated HSV infections observed in immunosuppressed individuals (2) and patients with diseasesthat compromise the cellular immune system (3, 4) suggestthe involvement of the cellular immune system in limiting HSV infections. Recent studies (5, 6) indicate that natural killer (NK) cells may play a role in limiting HSV infections. Monoclonal antibodies to lymphocyte differentiation markers ’ This work was supported in part by National Institutes of Health Grants EY-03 188-OlAl and PHS EY 07192, Grant G645 from the Fight for Sight Foundation, New York, and Post Graduate Training Grant PHS EY 07038. ’ To whom reprint requests should be addressed:University of Illinois Eye and Ear Infirmary, 1855 W. Taylor, Chicago, Ill. 606 12. 262 0008-8749184 $3.00 Gmyti&t 8 1984 by Academic Press,he. AII rIgbt.3of repmduction in any form rcszmd.

CHARACTERIZATION

OF NK-HSV EFFECI-OR CELLS

263

have been used to demonstrate that the human NK population is phenotypically heterogeneous (7-9). Functional heterogeneity in the human NK population has also been demonstrated. Data obtained with cold target competition studies ( 10) suggestedthat the NK cells that preferentially lyse HSV- 1 infected over uninfected fibroblasts (NK-HSV) recognize different antigenic structures on the target cells than those recognized by the NK cells that lyse K562 targets (NK-K562). Our interest in NK-HSV effector cells stems from our observation that peripheral blood lymphocytes (PBL) from patients whose HSV recurrences are limited to the cornea (Group I) expresslow levels of NK-HSV activity relative to patients with HSV cornea1 and skin lesions (Group II) and controls with no history of HSV lesions (submitted for publication). The cytotoxic effector cells appear to belong to the NK population because NK-HSV activity is found in the low density lymphocyte subpopulation that is enriched for cells with large granular lymphocyte (LGL) morphology. Furthermore, high levels of NK-HSV activity have been observed in the PBL of donors who are seronegative for anti-HSV-1 antibody. These studies support the concept that NK-HSV and NK-K562 effector cells belong to distinct subpopulations of NK cells. Evidence for active suppression of the NK-HSV activity of PBL from Group I patients is also presented. MATERIALS

AND METHODS

Cell lines. HEp-2, a human larynx carcinoma-derived cell line, and VERO, an African green monkey kidney cell line, were maintained in RPM1 1640 medium supplemented with antibiotics (100 units/ml of penicillin and 100 &ml of strep tomycin) and 10%(v/v) newborn calf serum (CS). A human myeloid cell line (K562) was obtained from W. D. Peterson, Jr., Child Research Center of Michigan, and maintained in a suspension culture in RPM1 1640 medium supplemented with antibiotics and 15% (v/v) CS. Virus preparation and quantitation. HSV type 1 (KOS strain) was passagedin VERO cells. Confluent monolayers were infected at a multiplicity of infection (MOI) of approximately one plaque-forming unit (PFU) per cell. The virus was then isolated from the infected cells after a majority of cells showed viral cytopathic effect. Complete virions were isolated on Percoll (11). The HSV-1 suspension was then assayed as PFUs on monolayers of VERO cells. Effector &ls. PBL were isolated from the heparinized blood of Group I and Group II patient donors and control donors. All patient donors were tested between HSV1 recurrences. Control donors were individuals who had no history of herpetic recurrences. All of the patient donors and 70% of the control donors were seropositive for antibody to HSV- 1. PBL were isolated by carbonyl iron Ficoll-Hypaque separation as previously described (12). A highly enriched lymphocyte population was obtained at the Ficoll-plasma interface. Residual monocytes were removed by adherence to plastic during a 1-hr incubation in a plastic tissue culture flask (Coming, Coming, N.Y.). The nonadherent PBL were found routinely to be 97 to 100% lymphocytes when stained with cu-naphthylacetate esteraseand counterstained with hematoxylin using an esterasestaining kit (Sigma, St, Louis, MO.). Depletion of lymphocyte subpopulations by complement-mediated lysis. In certain experiments, PBL subpopulations were sensitized with monoclonal (MC) antibodies

264

HENDRICKS

AND

SUGAR

specific for different cell surface markers and then lysed with rabbit complement (C’). Four different MC antibodies, each specific for a marker expressedon a subpopulation of human lymphocytes, were used in these studies. Anti-Leu-7 (HNK-1; BectonDickenson Monoclonal Center Inc., Mountain View, Calif.) and OKMl (Grtho Diagnostic Systems Inc., Raritan, N.J.) recognize determinants expressed on human NK cells. OKT3 (Ortho Diagnostic Systems Inc.) detects a marker expressed on human T lymphocytes. Anti-Leu-8 (Beckton-Dickenson) subdivides human regulatory T-cell populations by reacting with approximately 75% of Leu-3+ T helper cells and 60% of Leu-2+ T suppressor cells. The function of cells coexpressing the Leu-3 and Leu-8 markers is not known since Leu-3+8- cells mediate helper function. However, Leu-2+8+ cells have been shown to collaborate with Leu-2+8- cells in generating suppression of B-cell differentiation (13). PBL were incubated for 30 min on ice with the MC antibodies in RPM1 1640 and 0.5% y-globulin-free calf serum @G-free CS). The concentrations of MC antibodies were 0.3 pg of anti&u-7 and anti&u-g, 0.1 pg of OKM 1, and 0.125 pg of OKT3 per lo6 PBL. These concentrations were determined to be in slight excessof that required for plateau killing. Sensitized lymphocytes were lysed by treatment with rabbit C (Pel-Freeze, Rogers, Ark.) at a final dilution of 1:3. The cells were incubated with C for 45 min at 37”C, washed, and subjected to a second cycle of C treatment. The cells were then washed twice and resuspended in assay medium (RPM1 1640 plus 10% GG-free CS). Cell counts were adjusted to reflect the viability of cells receiving C only. Panning. In certain experiments PBL were depleted of Leu-2+ (suppressor/cytotoxic) T lymphocytes. Twenty million PBL were suspended in 2 ml of phosphate-buffered saline (PBS) containing 20 pg of anti-Leu-2a (Becton-Dickenson Monoclonal Center Inc.). The cells were incubated for 20 min on ice, washed, and depleted of Leu-2+ cells by panning as previously described (14). Briefly, sensitized PBL were incubated for 2 hr at 4°C in a petri dish (Fisher, Itasca, Ill.) that had been previously coated with the F(abh fragment of affinity-purified goat anti-mouse immunoglobulin (GAMIg) antibody. The nonadherent (enriched Leu-2-) cells were then removed by gently washing the surface of the dish five times with 1% CS/PBS, pooling the washes.Due to technical difficulties in retrieving the adherent cells, this procedure was used for negative selection only.. The efficiency of depletion of Leu-2+ cells was determined by retreating the nonadherent cells with anti-Leu-2a and fluoresce&conjugated GAMIg (Fl-GAMIg) followed by microscopic examination. The enriched I&u-2- cell prep arations were O-5% Leu-2+. All lymphocyte preparations were adjusted to reflect Leu2- cells only. Isolation of lymphocyte subpopulations with a fluorescence activated cell sorter (FACS). Positive and negative selection of lymphocyte subpopulations was accomplished with a FACS (“Epics 5,” Coulter Electronics Inc., Hialeah, Fla.). PBL were treated with anti-Leu-2a (0.5 pg/106 PBL) or anti-Leu-7 (1 &lo6 PBL) for 30 mitt, washed, and treated for 30 min with Fl-GAMIg ( 1.O pg/ 1O6PBL). The cells were then washed three times and resuspended at 2 X 106/ml in PBS. The fluorescencepositive and -negative populations were separated with the FACS. As controls, PBL treated with anti-Leu-2a or anti-Leu-7, but no Fl-GAMIg, were collected from the FACS as fluorescence-negative cells. The efficiency of separation was determined by retreating the sorted cells with anti-Leu-2a or anti-Leu-7 and Fl-GAMIg followed by flow analysis on the FACS. In all experiments reported, the enriched marker-positive cells were 85-90% positive while the marker-negative cells were l-5% positive.

CHARACTERIZATION

OF NK-HSV EFFECTOR CELLS

265

Target cells. HSV-l-infected HEp-2 cells (HEp-ZHSV) were used as targets to measure NK-HSV activity. A monolayer of HEp-2 cells was infected with HSV-1 at a MO1 of 10 PPU per cell for 6 hr. The infected monolayer was then trypsinized and washed, and 2 X lo6 viable cells were suspendedin 3 ml of assaymedium. The HEp2-HSV suspension was incubated overnight to regenerate surface structures. K562 cells were used as targets for measuring NKK562 activity. Prior to assay, 2 X lo6 HEp-ZHSV or K562 cells were labeled for 1 hr with 100 &i of Naz5’CrG, (sp act 350-600 mCi/mol, Amersham Corp., Arlington Heights, Ill.). The “Cr-labeled cells were washed three times and suspended at 1 X lo4 viable cells/ml in assaymedium. The proportion of HEpZHSV-expressing HSV-1 antigens routinely exceeded 90% as assessedby immunofluorescent staining with fluorescein-conjugated antiserum to HSV-1 (M. A. Bioproducts, Walkersville, Mass.). Cytotoxicity assay. The Cr-release cytotoxicity assay has been described in detail (15). Effector cells and lo3 targets at various effector to target (E/T) ratios were added to triplicate wells of a U-bottomed microtiter plate to determine experimental 5’Cr release. Controls included triplicate cultures of lo3 targets in assay medium (spontaneous 51CrRel) or in 1%Nonidet-P40 (NP-40, Particle Data Laboratories, Elmhurst, Ill.). in water (maximum “Cr Rel). The plate was centrifuged at 2008 for 1 min and then incubated stationary for 4 hr at 37°C in a humid 5% CO*-95% air atmosphere. The radioactivity in 100 ~1 of supematant fluid was determined with a Beckman 7000 liquid scintillation counter, using Ready-Solv HP cocktail (Beckman Instruments, Lincolnwood, Ill.). The percentage of 51Crrelease(% Sp. Rel.) was determined using the formula: experimental ?r Rel. - spontaneous “Cr Rel. % Sp. Rel. = x 100. maximum “Cr Rel. - spontaneous “Cr Rel. The spontaneous release for most experiments was less than 10% of the maximum release and never exceeded 20% of the maximum release. In some experiments the data are recorded as lytic units (LU). This value was calculated using the regression formula of the percentage of 5’Cr release versus the log of the E/T ratio. One lytic unit is defined as the number of lymphocytes required for 10% Sp Rel. The data recorded are lytic units per lo6 lymphocytes, and the correlation coefficient (r) of the regression line. Values less than .Ol LU per lo6 lymphocytes are recorded as 0. In vitro stimulation of cytotoxic cells. The cytotoxic activity of PBL was retested after in vitro stimulation with HSV- 1. One hundred thousand PBL in 0.1 ml of assay medium were dispensed into the wells of a U-bottomed microtiter plate. One-tenth milliliter of each serial lo-fold dilution (lo-‘- 10w5)of stock HSV-1 ( 1.7 X lo8 PPU/ ml) was dispensed into triplicate wells containing PBL. After 48 hr incubation, the cells in each culture were washed twice and resuspended in 0.1 ml of assay medium. Incubation with HSV-1 did not result in decreased viability, as assessedby trypan blue dye exclusion. Cytotoxic activity in these cultures was tested by adding “0-labeled target cells ( 103) to each well and performing the cytotoxicity assay as described above. RESULTS Characterization of the Cytotoxic Cells Since PBL from Group I donors exhibited low levels of NK-HSV activity, it was of interest to characterize the effector cells. PBL obtained from Group II donors

266

HENDRICKS AND SUGAR

exhibited high levels of NK-HSV activity, and were used to determine the surface phenotype of lymphocytes mediating this activity. Monoclonal antibodies were used to lyse with C, to deplete by “panning”, or to isolate by FACS sorting lymphocytes bearing the Leu-7, OKM l,OKT3, Leu-2, or Leu-8 markers. Each monoclonal antibody was tested in two to six experiments with similar results. The effect of lysing lymphocytes bearing the Leu-7, Leu-8, OKMl, and OKT3 surface markers on NK-HSV and NK-K562 activities of freshly isolated PBL is shown in Table 1. Treatment with anti-Leu-7 + C’ resulted in a mean reduction in viability of 15% (SD = 6%). Lysis of Leu-7+ cells significantly reduced both NK-HSV and NK-K562 activities. Lysis of Leu-7+ cells consistently resulted in greater reduction of NK-HSV activity than of NK-K562 activity. FACS depletion of Leu-7+ cells (Table 2) resulted in greater diminution of NK-HSV activity than was observed following treatment with anti-Leu-7 + C’. Treatment of PBL with anti-OKMl + C reduced their viability by a mean of 17% (SD = 2.8%). Lysis of OKMl’ cells resulted in 79100% reduction of NK-HSV activity, and 57-97% reduction of NK-K562 activity. Treatment of PBL with anti-OKT3 + C reduced viability by a mean of 72% (SD = 5.8%). Lysis of OKT3 lymphocytes had a differential effect, signihcantly diminishing the NK-HSV activity while not affecting or slightly enriching the NK-K562 activity. Lysis of Leu-8+ lymphocytes, while reducing viability by 60%, either had no effect or resulted in a slight increase in both NK-HSV and NK-K562 activities of PBL. Table 1 shows that the NK-HSV effecters expressed both NK markers and the pan-T-cell marker OKT3. Thus it was of interest to determine whether they also expressthe Leu-2 marker that defines the T suppressor/cytotoxic population. Depletion of Leu-2+ lymphocytes by panning did not significantly affect either the NK-HSV or NK-K562 activity of freshly isolated PBL (Table 2). Similarly, depletion of Leu-2+ TABLE 1 NK Activity following Antibody and Complement Treatment NK-K562

NK-HSV Treatment of effector cells

LU/lO6 effecters”

rb

LU/lO6 effecters

r

1

C’ Anti-Leu-7 + C’

69 17

0.9662 0.7410

309 193

0.9384 0.9004

2

C’ Anti-Leu-7 + C’

73 7

0.9797 0.9738

351 180

0.9083 0.8829

3

C’ OKMl OKT3

37 0 9

0.8720 0.250 1 0.8867

87 3 113

0.9870 0.8767 0.9604

4

C’ OKMl OKT3

132 28 17

0.9050 0.9474 0.8197

118 51 118

0.9728 0.9741 0.9076

5

C’ Anti-Leu-8 + C’

44 56

0.9762 0.9492

112 128

0.9719 0.9883

Expt

u 1 LU is the number of effector cells required to produce 10% lysis. bCorrelation coefficient for the regressionanalysis of the percentage specific s’Cr release versus the log of the effecter/target ratio.

CHARACTERIZATION

261

OF NK-HSV EFFECTOR CELLS TABLE 2

NK Activity of Marker Positive and Marker Negative PBL NKX562

NK-HSV Expt

Treatment of effector cells

LU/106 effectors”

rb

LU/106 e&tots

r

1

Unpanned Plate control’ Leu-2-

45.43 47.37 46.45

0.9180 0.9487 0.8678

129 137 119

0.7853 0.8990 0.6127

2

Controld Leu-2- (FACS) Leu-2+ (FACS)

19.73 21.40 0.04

0.7803 0.8592 0.7822

64 61 53

0.8593 0.9783 0.8822

3

Unsorted Controld Leu-7- (FACS)

114.12 96.54 8.48

0.8154 0.7241 0.7276

378 366 153

0.9299 0.8447 0.8529

u 1 LU is the number of effector cells required to produce 10% lysis. b Correlation coefficient for the regression analysis of the percentage specific “Cr release versus the log of the effecter/target ratio. c Untreated PBL that did not adhere to the GAMIg-coated petri dish. d PBL were treated with anti-Leu-2a or anti-Leu-7 but no Fl-GAMIg and collected from the FACS as fluorescence-negativecells.

cells by FACS sorting (Table 2) did not significantly affect the NK-HSV or NK-K562 activity of freshly isolated PBL. Surprisingly, the Leu-2+ lymphocytes that lacked NK-HSV activity showed only slightly diminished NK-K562 activity compared to unseparated PBL. NK Activity of Leu-2-Depleted PBL following in Vitro Cultivation We reported earlier that incubation of PBL obtained from Group II donors for 48 hr with HSV-1 resulted in augmented NK-HSV activity (submitted for publication). Similar treatment of PBL from Group I donors did not significantly affect NK-HSV activity. To determine the effect of Leu-2+ cells on the NK-HSV activity of cultured PBL from Group I and II donors, PBL were depleted of Leu-2+ cells followed by in vitro incubation. The results of two representative experiments in which PBL obtained from Group I and II donors were depleted of Leu-2+ cells by panning, incubated for 48 hr in the presence of HSV-1, and tested for NK-HSV activity are shown in Fig. 1. Depletion of Leu-2+ cells from the PBL of Group I donors consistently resulted in significantly increased NK-HSV activity after 48 hr of incubation in the presence of HSV- 1. A slight increase in NK-K562 activity was observed in PBL from some, but not all, of our Group I donors and, overall, the increase was not significant (data not shown). In contrast, depletion of Leu-2+ cells from the PBL of Group II donors either had no significant effect or diminished the NK-HSV activity (Fig. 1) or the NK-K562 activity (data not shown) after 48 hr of incubation in the presence of HSV- 1. Enriched Leu-2+ lymphocytes freshly isolated from Group II donors were shown to express high levels of NK-K562 activity but very little NK-HSV activity (vide supra). To determine if NK-HSV activity could be induced, the Leu-2+ cells were

268

HENDRICKS AND SUGAR 70.00 63.00 56.00 49.00 I 42.00 2 35.00

d is

1.

28.00 I 21.00 14.00

t

7.00 0.00 GROUP

I

GROUP

GROUP

II

I

GROUP

II

1. Two representative experiments in which PBL obtained from Group I and II donors were depleted of Leu-2+ cells by panning, incubated for 48 hr in the presenceof HSV- 1, and tested for NK-HSV activity. The data recorded are the mean 70Sp Rel. and standard error for triplicate cultures of unseparated (0) or Leu-2- (I) PBL. FIG.

cultivated with various doses of HSV-I and their NK-HSV and NK-K562 activities were retested. The NK-HSV activity of Leu-2+ cells was not a.ffectedand remained at the low level observed in the freshly isolated Leu-2+ cells (Fig. 2). The NK-HSV 100.00 90.00 80.00

t

70.00

1

60.00 e

50.00

Ii 2

40.00 30.00

0.00

1 .70

0.17 PFU’s

of

HSV-I

170.00

17.00 /

Culture


1000)

FIG. 2. PBL isolated from a Group II donor were treated with anti-Leu-2a plus Fl-GAMIg and sorted with the FACS into enriched Leu-2+ (0) or Leu-2- (m) lymphocytes. Unseparated cells (Cl) were PBL that had been treated with anti-Leu-2a only and collected from the FACS as fluorescence-negativecells. The responseof untreated PBL was not significantly different from that of the unseparated controls and is not shown. The lymphocytes were incubated for 48 hr with various doses of HSV-1 and then tested for NKHSV activity.

CHARACTERIZATION

OF NK-HSV EFFECTOR CELLS

269

activity of the enriched Leu-2- cells was significantly higher than that of the Leu-2+ cells but lower than that of the unseparated cells at all doses tested. The enriched Leu-2+ cells, while lacking NK-HSV activity, expressed high levels of NKX562 activity (Fig. 3). DISCUSSION In this study we have further defined the surface phenotype of the lymphocytes responsible for NK-HSV and NK-K562 activities in our assay. In freshly isolated PBL, most of the NK-HSV activity is mediated by cells with the surface phenotype Leu-7+, OKMl’, OKT3+, Leu-2-, Leu-8-. In contrast, the lymphocytes responsible for the NK-K562 activity express the surface phenotype Leu-7+, OKMl’, OKT3-, Leu-2+/-, Leu-8-. Thus the OKT3+ lymphocytes responsible for most of the NKHSV activity can be differentiated from the OKT3- lymphocytes mediating the NKK562 activity. These data suggest that the NK-HSV and NK-K562 activities may not merely be two functions of the same population of lymphocytes, but may be mediated by distinct subpopulations of NK cells. The two subpopulations may be of the same lineage since they share the Leu-7 and OKM 1 surface markers. Abo et al. (16) have demonstrated that 95% of human bone marrow Leu-7+ cells possessedthe surface phenotype Leu-7+, OKT3+, OKM l-, and expressedvery low levels of NK-K562 activity. In contrast, 70% of Leu-7+ cells in the peripheral blood are Leu-7+, OKT33, OKMl’, and expresshigh levels of NKK562 activity. Basedon these findings the authors hypothesize that NK cells undergo differentiation from immature to mature NK cells. In the course of differentiating

0.17 FU’s

I .70 of

sv-I

17 00 /

Culture

170
00

1700.00

1000)

FIG. 3. PBL isolated from a Group II donor were treated with anti-Leu-2a plus Fl-GAMIg and sorted with the FACS into enriched Leu-2+ (0) or Leu-2- (m) lymphocytes. Unseparated cells (0) were PBL that had been treated with anti-Leu-2a only and collected from the FACS as fluorescence-negativecells. The responseof untreated PBL was not significantly different from that of the unseparated controls and is not shown. The lymphocytes were incubated for 48 hr with various doses of HSV-I and then tested for NKK562 activity.

270

HENDRICKS AND SWGAR

the OKT3 marker is lost and the OKMl marker expressed.If this pathway is correct, the Leu-7+, OKT3+, OKM l+ cells responsible for most of the NK-HSV activity may represent an intermediate stage of differentiation which occurs prior to the loss of the OKT3 marker but subsequent to the acquisition of the OKMl marker. We previously demonstrated that PBL from Group I donors exhibit reduced levels of NK-HSV activity. It was not clear whether this reduced activity reflected a reduced number of cytotoxic cells or active suppression of the activity of these cells. Our current finding that depletion of the Leu-2+ cells elevates the NK-HSV activity of PBL from Group I donors but not Group II donors favors the latter hypothesis. Streilein et al. (17) first described a deviant immune response that develops when certain antigens are introduced into the ocular anterior chamber. This deviant response, referred to as anterior chamber-associatedimmune deviation (ACAID), is comprised of normal or enhanced antibody production but depressedcell-mediated immunity. The depressedcell-mediated immunity resulting from inoculation of HSV- 1 into the ocular anterior chambers of mice appears to be associated with the appearance of antigen-specific suppressor T lymphocytes in the spleen (18). We note certain similarities to ACAID in our Group I donors. These patients exhibit high serum titers of anti-HSV-1 antibody but impaired NK activity. The impaired NK activity appears to be associated with the presence of suppressor T lymphocytes in the blood. Perhaps the repeated introduction of HSV-1 antigen into the anterior chamber during periods of virus shedding and active corneal lesions has resulted in an ACAID-like deviant immune response in these patients. ACKNOWLEDGMENTS The authors gratefully acknowledge the technical assistanceof Barbara Grajewski and Christine Rolnik.

REFERENCES 1. Kaufman, H. E., Brown, D. C., and Ellison, E. D., Amer. J. Ophthalmol. 65, 32, 1968. 2. Montgomerie, J. Z., Becroft, D. M. O., Croxson, M. C., Doak, P. B., and North, J. D. K., Luncet 2, 867, 1969. 3. St. Geme, J. W. Jr., Prince, T., Burke, B. A., Good, R. A., and Krivit, W., N. Engl. J. Med. 273,226, 1965. 4. Sutton, A. L., Smithwick, E. M., Seligman, S. T., and Kim, D., Amer. J. Med. 56, 545, 1974. 5. Ching, C., and Lopez, C., Infect. Imrnun. 26, 49, 1979. 6. Lopez, C., Ryshke, R., and Bennet, M., Infect. Immun. 28, 1028, 1980. 7. Zarling, J. M., Clouse, K. A., Biddison, W. E., and Kung, P. C., J. Immunol. 127, 2575, 1981. 8. Ault, K. A., and Springer, T. A., J. Immunol. 126, 359, 1981. 9. Abo, T., and Balch, C. M., J. Immunol. 127, 1024, 1981. 10. Fitzgerald, P. A., Evans, R., Kirkpatrick, D., and Lopez, C., J. Immunol. 130, 1663, 1983. 11. Pertoft, H., Pharmacia Fine Chemicals, Separation News 3, (1980). 12. Levy, P. C., Shaw, G. M., and Lo Buglio, A. F., J. Immunol. 123, 595, 1979. 13. Gatenby, P. A., Kansas, G. S., Chen, Y. X., Evans, R. L., and Engleman, E. G., J. Immunol. 129, 1997, 1982. 14. Engleman, E. G., Benike, C. J., Grumet, F. C., and Evans, R. L., J. Immunol. 127,2124, 1981. 15. Brunner, K. T., Engers, H. D., and Cerottini, J. G., In “In Vitro Methods in Cell-Mediated and Tumor Immunity” (B. R. Bloom and J. R. David, Ed%), pp. 423-428. Academic Press,New York, 1976. 16. Abo, T., Cooper, M. D., and Balch, C. M., In “NK Cells and other Natural Effector Cells” (R. B. Heberman, Eds.), pp. 31-38. Academic Press,New York, 1982. 17. Streilein, J. W., Neiderkorn, J. Y., and Shadduck, J. A., J. Exp. Med. 152(4), 1121, 1980. 18. Whittum, J., Neiderkom, J., McCulley, J., and Streilein, J. W., ARVO Abstracts. Invest. Ophthalmol. Vis. Sci. 24(Supple), 2 13, 1983.