VIRAL IMMUNOLOGYVolume 13, Number 4, 2000Mary Ann Liebert, Inc.Pp. 481-495

CD8+ Cell Lines Isolated from HIV-1-Infected ChildrenHave Potent Soluble HIV-1 Inhibitory Activity That

Differs from /3-ChemokinesA. MOSOIAN,1* A. TEIXEIRA,1* E. CARON,1 J. PIWOZ,2 and M.E. KLOTMAN1*

ABSTRACT

CD8+ cells from human immunodeficiency virus type 1 (HIV-1) infected individuals havebeen shown to suppress HIV-1 replication both through a major histocompatibility complex(MHC)-restricted cytolytic pathway as well as through a noncytolytic pathway mediatedthrough soluble factors. To characterize this soluble activity and its potential role in diseaseprogression further, we studied the HIV-1 inhibition by supernatants derived from her-pesvirus saimiri-transformed CD8+ cells isolated from infected children. Three of the sixCD8+ cell lines derived had a phenotype consistent with an unusual natural killer (NK) cellsphenotype with low CD3, high CD56, and low CD16. Supernatants from some of the celllines derived from children with rapid progression as well as long-term nonprogressors ex-

hibited broad HIV-1-inhibitory activity in primary CD4+ cells as well as in primarymacrophages. In contrast to a cocktail of /3-chemokines, the supernatants inhibited T-tropicas well as M-tropic viruses, efficiently inhibited infection in primary macrophages, and in-hibited HIV-1 activation in the chronically infected Ul cell line. The HIV-1-inhibitory ac-

tivity was heat stable and active over a broad pH range. Fractionation of the supernatantby size and ion exchange chromatography demonstrated activity in the complete absence ofRANTES as well as interferons-a, ß, and y and in a size range of less than 10 kD and greaterthan 3 kD. CD8+ cell supernatants contain additional unidentified factors that have anti-HIV activity to account for this broad phenomenon.

INTRODUCTION

Asymptomatic long-term survival with human immunodeficiency virus type 1 (HIV-1) is clearly recog-nized in both the adult and pédiatrie population (1,7,44). Studies focused on factors that correlate with

progression in adults and children have repeatedly demonstrated that viral burden as measured utilizing a

number of different assays correlates with disease progression (2,15,35,36). The determinants of viral bur-den are most likely multifactorial and heterogeneous involving an interplay between host (genetic and im-munologie) and viral factors. CD8+ cells from HIV-1-infected individuals have been shown to suppress

Departments of 'Medicine and 2Pediatrics, Mt. Sinai School of Medicine, New York, New York.*These individuals contributed equally to this work.

481

MOSOIAN ET AL.

HIV-1 replication both through a major histocompatibility complex (MHC)-restricted cytolytic pathway as

well as through a noncytolytic pathway that is independent of human leukocyte antigen (HLA) and is me-

diated through soluble factors (52,53). The MHC-restricted cytotoxic T-cell response early in infection hasbeen associated with a more favorable viral burden (24,39). Furthermore, studies suggest that soluble CD8+cell antiviral activity (caf) correlates with the clinical status of an infected individual (33,47).

At least some of the soluble CD8+ cell inhibitory activity is due to the )3-chemokines RANTES (regu-lated on activation, normal T-cell expressed and secreted), MlP-la and MIP-1/3, which effectively blockreplication of HIV in CD4+ lymphocytes (9), especially when present in combination. The mechanism ofHIV-1 inhibition by this cocktail is now known to be via interaction with the HIV-1 coreceptor, CCR-5,which is a ligand for each of these inhibitory chemokines. This coreceptor is a receptor for entry of M-tropic isolates into T-cells but not for entry of T-cell tropic laboratory strains that primarily use the CXCchemokine receptor, CXCR4, or Fusin and can be blocked with the ligand for CXCR4, SDF-1 (14,16,17,41).Although the coreceptor ligands can block infection in vitro via interactions with a coreceptor, they can

also enhance replication through a number of pathways (19,23).Inhibitory profiles of the soluble CD8+ inhibitory factor(s) suggest that the c-c class of jß-chemokines

are not the only potent inhibitors released from CD8+ cells. The inhibitory activity appears to be againstT-tropic as well as M-tropic viruses (37). The mechanism of action appears to be, at least partially, throughthe inhibition of either basal or Tat-induced transcription (8,11,12,27,31 ). Several groups have demonstratedthat CD8-derived HIV-1-inhibitory activity from either primary or transformed cells is not completelyblocked by neutralizing antibodies against jß-chemokines and does not correlate with levels of MlP-la,MIP-10, and RANTES levels released from the CD8+ cells (4,20,37,49). The native form of an additionalchemokine, macrophage-derived chemokine (MDC), has been shown to have HIV-1-inhibitory activity. Atleast one receptor for MDC is CCR-4, not a known coreceptor for HIV-1 (21,43). However, a recent re-

port found that HIV-1 inhibitory activity of synthetic MDC against HIV-1 Bai was seen only in monocyte-derived macrophages, not in T-cell blasts and did not involve early steps prior to reverse transcription (13).The full inhibitory potential of native or modified MDC is not clear (51). Interleukin-16 (IL-16) also re-leased from CD8+ cells has been shown to have modest anti-HIV activity as well (3). Increasing evidencesuggest that there are additional potent factors expressed by primary or transformed CD8+ cells that arenatural inhibitors of HIV-1.

To characterize CD8+ cell HIV-inhibitory activity and its potential role in disease progression in HIV-1-infected children further, we studied the HIV-1 inhibition of supernatants derived from herpesvirus saimiri-transformed CD8+ cells isolated from HIV-infected children with differing disease progression.

MATERIALS AND METHODS

The cohort. Two groups of HIV-infected children were identified for these studies, children with rapidlyprogressive disease having acquired immunodeficiency syndrome (AIDS)-related symptoms within 2 yearsof infection and long-term survivors consisting of children who have remained asymptomatic for at least 8years after infection and have CD4+ counts of greater than 500 cells/mm3. Plasma HIV-1 RNA levels were

determined using quantitative reverse transcriptase/polymerase chain reaction amplification (Amplicor As-say, Roche Molecular Diagnostics).

CD8+ cell lines. Peripheral blood mononuclear cells (PBMCs) from each of the HIV-infected childrenwere separated from whole blood by Ficoll-Hypaque centrifugation and infected in bulk with herpesvirassaimirí (HVS) strain C-488 (kindly provided by R. Desrosiers, New England Regional Primate Center) or

infected after 2-3 days of stimulation with phytohemagglutin antigen (PHA) (Sigma, St. Louis, MO). HVS-transformed cells retain many of the characteristics of primary cells including the requirement of IL-2 forcontinuous growth, their original T-cell receptor expression and specific antigen recognition after prolongedin vitro growth (6,54). After 6-8 weeks in culture with conditioned medium (RPMI) containing 10% fetalcalf sera and IL-2 at a concentration of 20-50 U/mL (Gibco BRL, Grand Island, NY), viable cells were re-covered by Ficoll-Hypaque separation. Single cell clones from some of the established cell lines were ob-tained by limiting dilution. Cell lines and clones were characterized by FACS utilizing a panel of antibod-

482

CD8+ CELL LINES WITH POTENT HIV-1 INHIBITORY ACTIVITY

ies including Leu 2a and Leu 3a, (Becton Dickinson, Bedford, MA), CD25, HLA-DR, CD8, and CD3(Biosource International, Camarillo, CA) CD16 (Immunotech, Marseille, France) and CD56 (Caltag, SanFrancisco, Ca). Both CD4+ and CD8+ cell lines were obtained and further characterized.

HIV-1 inhibitory activity. Purified CD8+ cell lines and clones were screened for soluble HIV-1 in-hibitory activity in transformed and primary CD4+ cells (purified by magnetic beads from PBMC) and inprimary macrophages utilizing both primary and laboratory-adapted strains of HIV-1. Macrophage cultureswere prepared from Ficoll-Hypaque (Sigma) purified PBMCs with and without CD 14 selection with mag-netic beads and purified by adherence after 14 days in culture in RPMI containing 20% FCS. The CD4+transformed cell line used for some of these infections was transformed by HVS and was broadly permis-sive for laboratory-adapted as well as primary HIV isolates (40). The CD4+ cell line or primary cells were

incubated with virus for 2 hours, the cells washed, and subsequently cultured in the presence of conditionedmedia (RPMI) with IL-2 and CD8+ cell supernatant. To obtain the supernatants, CD8+ cells were platedat a density of 1 X 106 cells per well and cultured in RPMI with fetal calf serum (FCS) and IL-2. After3-4 days, CD8+ cell supernatant was harvested and residual cells removed by centrifugation and filtrationusing 0.45-j_m filters. In the HIV-1-inhibition assays, the CD8+ supernatant represented 10%-25% of theculture media for primary macrophages and CD4+ cells, respectively. HIV-1 infection was monitored bymeasuring p24 by enzyme-linked immunosorbent assay (ELISA) (Du-Pont, Wilmington, DE). CD4+ cellswere infected with different strains of HIV-1 (at a multiplicity of infection of 0.01-0.1) including HIV-1MB, HIV-1 MN, HIV-1 Bal, as well as primary isolates. The primary viral isolates were obtained from pa-tients and minimally passed in culture using allogeneic PBMCs. The CD8+ cell inhibitory activity was com-

pared to activity of a cocktail of the recombinant chemokines including RANTES, MlP-la and MIP-1/3(R&D, Minneapolis, MN) each at a concentration of 200 ng/mL. Media was changed every 3^4- days inculture with fresh CD8+ cell supernatant or chemokine cocktail added each time.

Chemokine and interferon measurements. Cells were cultured in RPMI containing 10% FCS and IL-2 at 50 U/mL at an initial concentration of 0.5 X 106 per milliliter. After 72 hours in culture, supernatantswere collected and the following were measured by ELISA using commercially available assays with therespective cutoffs in parentheses: RANTES (50 pg/mL, R&D), MlP-la (50 pg/ml, R&D), IFN-a (78 pg/mL,Biosource International), IFN-/3 (0.25 IU/mL, Biosource International), and IFN-y (100 pg/mL, BiosourceInternational).

Characterization of HIV-1 inhibitory activity. The heat stability of the HIV-1 inhibitory activity was

tested as follows. Aliquots of serum-free caf 10 supernatant (200 ph) were placed in 1.5 mL Eppendorftubes and placed into a preheated waterbath of 50°, 60°, 70°, and 80°C for 10 minutes and then allowed tocool to room temperature. Treated medium (150 juL) was diluted with serum-containing medium to a finalconcentration of 10% and applied to primary macrophages after 2 hours of viral absorption as previouslydescribed.

To test pH stability of the activity, 10-mL aliquots of serum-free caf 10 supernatant were stirred gentlyin beakers with continuous monitoring of pH via a small-volume pH electrode. The pH was lowered or

raised from 7 to a set pH between 2 and 12 in two pH unit steps and held at the designated pH with stir-ring for 10 minutes. The pH was then quickly returned to 7.5. An aliquot was taken from each and testedfor inhibition of HIV-1 Bai in primary macrophages as previously described.

Partial protein isolation. To exclude any contribution of the jß-chemokines or interferons, partial pro-tein fractionation was performed as described below and fractions tested for HIV-1 inhibitory activity as

well as the presence of RANTES and interferon.Caf 10 supernatant was prepared from cells grown for 4 days at 1 X 106 cells per milliliter in serum-

free, phenol red-free RPMI-1640 (Life Technologies, Rockville, MD) supplemented with IL-2, insulin (50ng/mL), transferrin (50 ng/mL), and sodium selenite (50 pg/mL). The medium was clarified by centrifuga-tion and filtered through low protein binding 0.22 j_m cellulose acetate filters. The cell-conditioned cell-free medium was lyophilized to dryness in 200-mL aliquots in 600-mL Labconco lyophilizer flasks.

The lyophilate was dissolved in a total of 90 mL of 20 mM Tris, pH 8, and size-fractionated in 15-mLaliquots in Centriprep (Amicon, Bedford, MA) centrifugal ultrafiltration membranes with a molecular weightcutoff (MWCO) of 10 kD (Centriprep 10). The Centriprep 10 membrane was used to clarify the concen-

trated medium for application to Centriprep 3 membranes. Previous work had shown that most of the HfV-1

483

MOSOIAN ET AL.

inhibitory activity was retained by the 3 kD MWCO membrane and that it appeared in the filtrate of all thehigher MWCO Centriprep membranes. However, direct application of the concentrated cell-free cell-con-ditioned medium to the Centriprep 3 quickly clogged the membranes and prevented concentration of themedium.

The reténtate of the Centriprep 3 (final volume, 20 mL) was applied in 2-mL aliquots to a MonoQ HR5/5 column (Amersham Pharmacia Biotech, Piscataway, NJ) using a 30-minute gradient of 0 to 1 M NaClin 20 mM Tris, pH 8. The fractions that were found to inhibit HIV-1 Bal on primary macrophages were

tested for the presence of RANTES and IFN-y.Ul assay. The HIV-1 latently infected promonocytic cell line, Ul, can be induced to express HIV with

activating agents including phorbol myristate acetate (PMA) (18). Ul cells were exposed to PMA at a con-

centration of 10~10 M followed by treatment with CD8+ cell supernatants. HIV-1 p24 was measured atdays 2 and 4 after activation in the presence of CD8+ supernatants.

Statistics. In the in vitro culture experiments, each culture condition was done in triplicate and the re-

sults expressed as a mean of all three wells with the indicated standard deviation.

RESULTS

For these studies, HIV-1-infected children with rapid progression as well as long-term nonprogressionwere identified and plasma RNA levels determined. Both CD4+ and CD8+ cell lines were established fromchildren in each group and supernatants from the CD8+ cell lines screened for HIV-1-inhibitory activity(Table 1). As shown in Figure 1, cell lines with inhibitory activity against HIV-1 MB (A) and HIV-1 Bal(B) were isolated from HIV-infected children with rapid disease progression (closed symbols) as well as

from HIV-infected children with long-term nonprogression (open symbols). In this assay, inhibitory activ-ity was tested in a broadly permissive HVS-transformed CD4+ cell line. Each of the cell lines and cloneswere IL-2 dependent for continuous growth and all were CD8+ and CD4~. HLA DR and CD25 were de-termined for caf 10, 16, and 6 and found to be 25%-35%, respectively. Three of the cell lines, caf 10, 17,and 18, had low or nondetectable CD3 with 99% positive for CD56 and negative for CD 16, an unusual nat-ural killer (NK) cell phenotype. While caf 15 and 16 had inhibitory activity against HIV-1 MB and no ac-

tivity against Bal, the remaining supernatants from these cell lines had significant inhibitory activity forboth of these laboratory-adapted strains. Caf 10, 6, 17, and 18 inhibited a primary isolate grown in primaryCD4+ cells as well (Fig. 2). There was no apparent correlation between the ability to isolate these trans-formed cell lines and the clinical status or plasma RNA level of the source of the cells (Table 1 ).

The anti-HIV activity of one of the better growing single cell clones, caf 10, was further characterized.This clone was CD4 , CD8+, CD3 low (16%), CD16 negative and CD56 positive (91%). Supernatant fromthis cell line, originally isolated from a child with rapid disease progression, potently inhibited HIV-1 MN,HIV-1 Bal, HIV-1 MB, and an additional primary isolate (Fig. 3). To contrast this activity with that of the

Table 1. Source of CD8+ Cell Lines and Clones

Plasma HTV-1 RNA levelsCell line/clone Clinical status of source (copies/mL)

Caf 10 (clone) Rapid progression 459,688Caf 15 (clone) Long-term nonprogression 1,291Caf 16a (line)Caf 6 (line) Long-term nonprogression 2,264Caf 17 (clone) Rapid progression 968,000Caf 18 (clone)b

"Same donor source as Caf 15.bSame donor source as Caf 17.

484

CD8+ CELL LINES WITH POTENT HIV-1 INHIBITORY ACTIVITY

Days Post Infection

7 9

Days Post Infection

FIG. 1. Human immunodeficiency vims type 1 (HIV-1 )—inhibitory activity of supernatants from CD8+ cell linesestablished from HIV-1-infected children with rapid disease progression (closed symbol) or children with long-termnonprogression (open symbols) against HIV-HUB (A) and HIV-1 Bal (B). Supernatant from each cell line was incu-bated with herpesvirus saimirí (HSV)-transformed CD4+ cells after infection with the indicated vims. HIV replicationwas monitored by measuring HIV-1 p24 in the cell supernatants on the indicated days after infection.

ß-chemokines, inhibition of both HIV-1 Bal and HIV-1 MB were evaluated in primary CD4+ cells. TheCD8+-cell supernatant were compared to a cocktail of RANTES, MlP-la, and MIP-1/3 each at a concen-tration of 200 ng/mL. As expected, the cocktail potently inhibited HIV-1 Bal and showed no inhibition ofHIV-1 MB consistent with the utilization of the coreceptor CCR5 by Bal and CXCR4 by MB. In contrast,

150-,

Ec<*CMQ.

>

100 -\

50 A

0

^T

§§t__

V,__

T¿___KS_

untreated

_3 caf 10

H caf 15

__ caf 16

H caf6

II caf 17

caf 18

Day 10 Post InfectionFIG. 2. Human immunodeficiency vims type 1 (HlV-l)-inhibitory activity of supernatants from CD8+ cell lines

against a primary isolate of HIV-1 in primary CD4+ cells. Supernatant from each cell line was incubated with primaryCD4+ cells after infection with a primary isolate and HIV-1 p24 measured on day 10 after infection.

485

MOSOIAN ET AL.

100n untreatedCaf 10

Virus: HIV-1 MN Virus: HIV-1 primary isolate

Virus: HIV-1 MB

Days Post Infection

FIG. 3. Human immunodeficiency vims type 1 (HlV-l)-inhibitory activity of the supernatant from the caf 10 cellline against a panel of HIV-1 isolates from in an herpesvims saimirí (HVS)-transformed CD4+ cell line. HIV-1 p24was measured in the supernatant at the indicated times after infection.

the caf 10 supernatant inhibited both laboratory-adapted strains in the primary cells (Fig. 4). Further evi-dence of differing activity was demonstrated in primary macrophages. The CD8+ supernatant consistentlyand potently inhibited HIV-1 Bai in primary macrophages derived from different donors while the chemokinecocktail at 200 ng/mL showed little inhibitory activity (Fig. 5). When the dose of each jß-chemokine was

increased to 400 ng/mL some inhibitory activity was observed, however, this activity seemed to vary withthe preparation of primary macrophages (Figs. 5 and 6). In the same experiments, the CD8+ cell super-natant diluted 14 with complete media almost completely inhibited virus replication (Figs. 5 and 6).

All of the cell lines made significant amounts of IFN-y and RANTES but there appeared to be no cor-relation of these levels with inhibitory activity. The concentration of RANTES in the caf 10 supernatantprior to dilution in complete media ranged from 5.7 to 14.3 ng/mL and the MlP-la level was 6.9 ng/ml.The nonsuppressive cell line, caf 16, had some of the highest concentrations of the /3-chemokine RANTES

486

CD8+ CELL LINES WITH POTENT HIV-1 INHIBITORY ACTIVITY

20

154

c

CMa.

104

'IHIV-1 BaL Day 6 HIV-1 1MB Day 7

untreated

chemokine cocktail

L_ caf 10

FIG. 4. Supernatant from the CD8+ cell line caf 10 inhibited both human immunodeficiency vims type 1 (HIV-1) HIBand HJV-1 Bal in primary CD4+ cells while the chemokine cocktail inhibited only HIV-1 Bal. Primary CD4+ cells were

infected with the indicated virus and treated with either caf 10 supernatant or a cocktail of RANTES, MlPl-a, and MIP1-ß each at a concentration of 200 ng/mL. HTV-1 p24 was measured in the supernatant at the indicated time after infection.

5 7 9 12

Days post Infection

- - K- - untreated

—•— caf 10

—o— recombinant chemokines

60

50

_) 40c

çjj. 30

> 20-X

10-

6 10

Days post Infection

6 10

Days post Infection

FIG. 5. Supernatant from the CD8+ cell line caf 10 inhibited replication of human immunodeficiency vims type1 (HIV-1) Bai in primary macrophages. Each graph represents infection of macrophages isolated from a different donor.Primary macrophages were infected with HIV-1 Bal and treated with either caf 10 supernatant or a cocktail of RANTES,MlPl-a and MIP1-/3 each at a concentration of 200 ng/mL. HIV-1 p24 was measured in the supernatant at the indi-cated time after infection.

487

MOSOIAN ET AL.

200

-X— media-o— 50 ng/ml-A— 200 ng/ml-•— 400 ng/ml-•— caf 10, 25%

1 1

Days Post Infection

FIG. 6. Inhibition of human immunodeficiency vims type 1 (HIV-1) Bal in primary macrophages by supernatantfrom the CD8+ cell line caf 10 at 25% concentration compared to RANTES, MlPl-a, and MIP1-/3 each at a concen-

tration of 50, 200 or 400 ng/mL. HIV-1 p24 was measured in the supernatant at the indicated time after infection.

at 17.6 ng/mL. The preparation of supernatant for purification involved culturing the cell line in serum-freemedia for 4 days. RANTES levels under these conditions fell significantly (1.321 ng) but inhibitory activ-ity remained (Fig. 7). Similarly, the suppressive line caf 10 had the same level of IFN-y as the nonsup-pressive caf 16 line with 16.5 ng/mL and 17.0 ng/mL, respectively. Caf 18 had the lowest concentration ofIFN-y at 1.4 ng/mL but efficiently inhibited both HIV-IIIIB and Bal. IFN-y levels for caf 6, 15, and 17were 12.2, 0.5, and 14.2 ng/ml, respectively. For all of the cell lines, IFN-/3 and IFN-a levels were belowthe measured cutoffs of 0.25 IU and 100/pg/mL respectively. Therefore, neither /3-chemokine levels nor

levels of IFN-a, IFN-jß, or IFN-y correlated with inhibitory activity of the unfractionated supernatants.

BOO

I5>X

medium

Caf 10, conditioned,serum free medium

Reténtate, 10 kDa MWCO

E3 Reténtate, 3 kDa MWCO

Filtrate, 3 kDa MWCO

0.01

FIG. 7. Size fractionation of human immunodeficiency vims type 1 (HIV-1) Bal inhibitory activity in caf 10 cellsupernatant. Utilizing centrifugation ultrafiltration, the indicated fractions were tested for HIV-1 inhibitory activity inprimary macrophages. The HIV-1 p24 values were measured in the culture supernatant at day 11 after infection.

488

CD8+ CELL LINES WITH POTENT HIV-1 INHIBITORY ACTIVITY

The cell-conditioned medium containing caf 10 was concentrated by lyophilization and initially size frac-tionated on centrifugal ultrafiltration membranes of varying molecular weight cutoffs (MWCOs). The fil-trate of the 10-kd MWCO membrane that had previously been shown to have HIV-1-inhibitory activity(data no shown) was applied to a 3-kd MWCO membrane. The reténtate of the 3-kd MWCO membraneshowed enhanced activity over the filtrate of this same membrane, unfractionated supernatant, and the re-

téntate of the 10-kd MWCO membrane (Fig. 7). This suggests that under the conditions applied the opti-mum activity is in the size range of 3 kd to 10 kd. The 3-kd MWCO retenate was further separated by ionexchange chromatography. Both RANTES and IFN-y eluted in the initial peaks (1 and 2). In the third elu-tion peak, IFN-y and RANTES were below the level of detection however, HIV-1-inhibitory activity was

greater than any other elution peak (Fig. 8).There was no significant loss of HIV-1-inhibitory activity against HIV-1 Bal when caf 10 supernatant

was subjected to heating up to 80°C or to a change in pH over a broad range of 2-12, consistent with thesmall size of the protein.

The supernatants had modest inhibitory effects against activation of HIV when tested in the latently in-fected cell line, Ul. In this cell line, activation of the virus is from integrated proviral DNA and thereforeinhibitory effects would primarily be on steps after integration including transcription. Although activationin this cell line is resistant to reverse transcriptase inhibitors such as azidothymidine (AZT), blocking of re-

infection in this cell line could account for some of the inhibitory activity of the supernatant (46). As shownin Figure 9, the supernatant derived from caf 10 had the greatest inhibitory effect after PMA activationwhile caf 15 and caf 16 had modest inhibitory effects.

IFN-y: 128 ng 31.2 ng <100pg/ml <100 pg/ml <100 pg/ml <100pg/ml

RANTES: 261 pg 3.568 ng <15 pg/ml <15 pg/ml <15 pg/ml <15 pg/ml

media ,1 2 3 4 5 6

elution peaksFIG. 8. Elution peaks from ion exchange chromatography of caf 10 supernatant initially fractionated by centrifu-

gation ultrafiltration. The reténtate of the 3-kd molecular weight cutoff (MWCO) Centriprep 3 was separated on a

MonoQ H/R 5/5 column (Amersham Pharmacia Biotech) and the peaks analyzed for HIV-1 Bal inhibitory activity on

primary macrophages. The HIV-1 p24 values were measured in the culture supernatant at day 11 after infection. Thepeaks were also analyzed for the presence of RANTES and interferon-y (IFN-y).

489

MOSOIAN ET AL.

30-1

Ë 20

s,c

o.

>X

10H

untreatedcaf 10 (RP)caf 15(LTS)caf 16(LTS)

Days Post Activation with PMA

FIG. 9. Inhibition of activation of human immunodeficiency vims type 1 (HIV-1 ) in the chronically infected lym-phocyte line, Ul. Ul cells were treated with phorbol myristate acetate at a concentration of 10~10 M followed by treat-ment with CD8+ cell supernatants. HIV-1 p24 was measured at days 2 and 4 after activation.

DISCUSSION

CD8+ cell lines and clones demonstrating significant and broad HIV-1 inhibitory activity were easilyisolated from HIV-1-infected children. Consistent with previous reports on this activity derived from pri-mary CD8+ cells; the activity was soluble, did not require cell to cell contact and was not HLA-restricted(30,52,53). The isolation of these cells appeared to be independent of the state of HIV-infection of the hostas cell lines with potent inhibitory activity were isolated from children with rapid progression as well as

from the long-term nonprogressors. Furthermore, within the same patient, there were cell lines that inhib-ited HIV-1 MB and appeared to enhance HIV-1 Bal. Therefore the suppressive activity profile was cellline-specific. Prior work on caf from non-transformed CD8+ cells derived from HIV-infected-individualssuggests that soluble CD8+ CAF activity correlates with the clinical status of infected adults as well as in-fants (33,34,47). The apparent lack of correlation between production of these factors and the clinical sta-tus of the host in this study does not preclude such an association in light of the transformation processused to derive the lines. In our previous work, utilizing HVS-transformed cells from an uninfected donor,we demonstrated a similar pattern of broad, soluble antiviral activity (38). Although soluble inhibitory ac-

tivity reported by others in nontransformed CD8+ cells has been predominantly seen in infected and/or ex-

posed individuals when an acute infection model is used to screen activity, some have reported this activ-ity in CD8+ cells derived from uninfected individuals as well (5,28,29,48,50). HVS-transformed cell linesafter infection and transformation with HSV are all activated and could induce expression of the solublefactor(s) and mask a clinical association. It has been reported that antigen recognition through an HLAclass-1-restricted pathway may be one way to trigger this soluble activity (55).

Attempts to isolate soluble HIV-1-inhibitory cell factors have been limited by the number of primarycells available and by the limited life span of these cells in tissue culture. The isolation of transformed cell

490

CD8+ CELL LINES WITH POTENT HIV-1 INHIBITORY ACTIVITY

lines has resulted in further characterization of this inhibitory activity (9,12,26,32). Some of the cell lines(caf 10, 17, and 18) isolated in this report have a unique phenotype relative to previous reports of caf ac-

tivity in both transformed and nontransformed CD8+ cell lines. This phenotype is an uncommon pheno-type of NK cells and has been reported to occur in 0.66% to 1.11% of peripheral blood cells of men andwomen respectively (22). Soluble suppressor activity has been reported in CD56+/CD16+ nontransformedNK cells derived from infected individuals and largely attributed to the j3-chemokines, although in some

instances not all activity could be neutralized by antibodies to these chemokines (42). The j3-chemokinesRANTES, MlPl-a and MIP-1/3 are also made in CD8+ T cells and their HIV-1-inhibitory capacity was

first identified in an HTLV-1 transformed CD8+ cell line (9). The activity present in the HVS-transformedCD8+ cell lines derived from the HIV-1-infected children reported above, as well as that described by oth-ers, has features that cannot be attributed to /3-chemokines. These features include the broad inhibition ofstrains using either CCR-5 or CXCR-4 receptors and the potent inhibitory activity of caf against HIV-1 inmacrophages (25,37). Although CCR-5 is a receptor for entry of strains such as HIV-1 Bal into primarymacrophages, inhibition by recombinant )3-chemokines is variable, whereas caf consistently inhibited HIV-1 in primary macrophages from a panel of donors and appears to be more potent that inhibition in primaryor transformed CD4+ cells. Lastly, inhibitory activity of these cell lines does not correlate with differencesin MlP-la, MIP-1/3, RANTES, or interferon levels (27,49). Recently, it has been demonstrated that cell su-

pernatants from PBMC stimulated with influenza A virus had broad HIV-1 suppressive effects against HIV-1 and this activity was partially blocked with antibodies to IFN-a (45). The HVS-transformed cell super-natants described here had no detectable IFN-a and IFN-/3 and as was the case with RANTES, inhibitoryactivity could be separated from any detectable IFN-y. In fact, in contrast to previous reports (26), sepa-rating out the jß-chemokines and IFN-y resulted in no loss of inhibition of HÍV-l Bai in primarymacrophages. The size profile for activity was similar to chemokines in that activity appeared to be asso-

ciated with low molecular weight proteins less than 10 kd as has been reported by Lacey et al. (26).Because previous reports indicated that CD8+ cell supernatant inhibition may work at the level of tran-

scription, we tested the suppressive effects of caf 10 supernatant on activation of HIV-1 in Ul cells. Thiscell line is resistant to AZT, suggesting that viral production is largely from integrated proviras rather thanfrom reinfection of the chronically infected cell (46). We found modest inhibitory effects in this assay, sim-ilar to what was reported using transformed CD8+ cell lines by Moriuchi et al. (37). Copeland et al. (10)has shown that CD8+ T-cell supernatant effects on transcription are target cell-line dependent with sup-pression in T-cell lines and enhancement in monocyte-derived lines. Utilizing different assays, work by oth-ers has also suggested that CAF inhibits either basal or Tat-induced transcription (8,31). The inhibition ofthe Ul assay was maximally 50%-85% whereas in the virus replication assays, inhibition often exceeded95% suggesting additive mechanisms of inhibition. It is now apparent that these supernatants contain a num-

ber of soluble factors that have anti-HIV activity that when combined might account for this broad phe-nomenon. Because, these unfractionated supernatants do contain a number of factors that can enhance as

well as suppress HIV-1, further understanding of the nature of the unknown factors and their mechanismof action will be achieved with purified proteins through the use of these transformed cell lines.

ACKNOWLEDGMENTS

This work was presented in abstract form at the 4th Conference on Retroviruses and Opportunistic In-fections, January 24-26, 1997, Washington D.C. #744.

This work was supported by the Pédiatrie AIDS Foundation grant #50592-18-PG.

REFERENCES

1. 1994. Features of children perinatally infected with HIV-1 surviving longer than 5 years. Italian Register for HIVInfection in Children Lancet. 343:191-195.

2. Abrams, E.J., J. Weedon, R.W. Steketee, G. Lambert, M. Bamji, T. Brown, M.L. Kalish, E.E. Schoenbaum, P.A.

491

MOSOIAN ET AL.

Thomas, and D.M. Thea. 1998. Association of human immunodeficiency vims (HIV) load early in life with dis-ease progression among HIV-infected infants. New York City Perinatal HIV Transmission Collaborative StudyGroup. J. Infect. Dis. 178:101-108.

3. Baier, M., A. Werner, N. Bannert, K. Metzner, and R. Kurth. 1995. HIV suppression by interleukin-16 [letter; com-

ment] [see comments]. Nature 378:563.

4. Barker, E., K.N. Bossart, and J.A. Levy 1998. Primary CD8+ cells from HIV-infected individuals can suppressproductive infection of macrophages independent of beta-chemokines. Proc. Nati. Acad. Sei. USA. 95:1725-1729.

5. Barker, T.D., D. Weissman, J.A. Daucher, K.M. Roche, and A.S. Fauci. 1996. Identification of multiple and dis-tinct CD8+ T cell suppressor activities: Dichotomy between infected and uninfected individuals, evolution withprogression of disease, and sensitivity to gamma irradiation. J. Immunol. 156:4476-4483.

6. Biesinger, B., I. Muller-Fleckenstein, B. Simmer, G. Lang, S. Wittmann, E. Platzer, R.C. Desrosiers, and B. Fleck-enstein. 1992. Stable growth transformation of human T lymphocytes by herpesvirus saimirí. Proc. Nati. Acad. Sei.USA. 89:3116-3119.

7. Cao, Y., L. Qin, L. Zhang, J. Safrit, and D.D. Ho. 1995. Virologie and immunologie characterization of long-termsurvivors of human immunodeficiency vims type 1 infection [see comments]. N. Engl. J. Med. 332:201-208.

8. Chen, C.H., K.J. Weinhold, J.A. Bartlett, D.P. Bolognesi, and M.L. Greenberg 1993. CD8+ T lymphocyte-medi-ated inhibition of HIV-1 long terminal repeat transcription: A novel antiviral mechanism. AIDS Res. Hum. Retro-viruses. 9:1079-1086.

9. Cocchi, F., A.L. DeVico, A. Garzino-Demo, S.K. Arya, R.C. Gallo, and P. Lusso. 1995. Identification of RANTES,MIP-1 alpha, and MIP-1 beta as the major HIV-suppressive factors produced by CD8+ T cells [see comments].Science. 270:1811-1815.

10. Copeland, K.F., J.G. Leith, P.J. McKay, and K.L. Rosenthal. 1997. CD8+ T cell supernatants of HIV type 1-in-fected individuals have opposite effects on long terminal repeat-mediated transcription in T cells and monocytes.AIDS Res. Hum. Retroviruses. 13:71-77.

11. Copeland, K.F., P.J. McKay, and K.L. Rosenthal. 1995. Suppression of activation of the human immunodeficiencyvims long terminal repeat by CD8+ T cells is not lentivims specific. AIDS. Res. Hum. Retroviruses. 11:1321-1326.

12. Copeland, K.F., P.J. McKay, and K.L. Rosenthal. 1996. Suppression of the human immunodeficiency vims longterminal repeat by CD8+ T cells is dependent on the NFAT-1 element. AIDS. Res. Hum. Retroviruses. 12:143-148.

13. Cota, M., M. Mengozzi, E. Vincenzi, P. Panina-Bordignon, F. Sinigaglia, P. Transidico, S. Sozzani, A. Mantovani,and G. Poli. 2000. Selective inhibition of HIV replication in primary macrophages but not T lymphocytes bymacrophage-derived chemokine. Proc. Nati. Acad. Sei. USA. 97:9162-9167.

14. Deng, H., R. Liu, W. Ellmeier, S. Choe, D. Unutmaz, M. Burkhart, P. Di Marzio, S. Marmon, R.E. Sutton, CM.Hill, C.B. David, S.C. Peiper, T.J. Schall, D.R. Littman, and N.R. Landau. 1996. Identification of a major co-re-

ceptor for primary isolates of HIV-1 [see comments]. Nature. 381:661-666.

15. Dickover, R.E., M. Dillon, K.M. Leung, P. Krogstad, S. Plaeger, S. Kwok, C. Christopherson, A. Deveikis, M.Keller, E.R. Stiehm, and Y.J. Bryson. 1998. Early prognostic indicators in primary perinatal human immunodefi-ciency vims type 1 infection: importance of viral RNA and the timing of transmission on long-term outcome. J.Infect. Dis. 178:375-387.

16. Dragic, T., V. Litwin, G.P. Allaway, S.R. Martin, Y. Huang, K.A. Nagashima, C. Cayanan, P.J. Maddon, RA.Koup, J.P. Moore, and W.A. Paxton. 1996. HIV-1 entry into CD4+ cells is mediated by the chemokine receptorCC- CKR-5 [see comments]. Nature. 381:667-673.

17. Feng, Y., CC. Broder, P.E. Kennedy, and E.A. Berger. 1996. HIV-1 entry cofactor: Functional cDNA cloning ofa seven-transmembrane, G protein-coupled receptor [see comments]. Science. 272:872-877.

18. Folks, T., D.M. Powell, M.M. Lightfoote, S. Benn, MA Martin, and A.S. Fauci. 1986. Induction of HTLV-III/LAVfrom a nonvims-producing T-cell line: Implications for latency. Science. 231:600-602.

19. Gordon, C.J., M.A. Muesing, A.E. Proudfoot, C.A. Power, J.P. Moore, and A. Trkola. 1999. Enhancement of hu-man immunodeficiency vims type 1 infection by the CC- chemokine RANTES is independent of the mechanismof vims-cell fusion. J. Virol. 73:684-694.

492

CD8+ CELL LINES WITH POTENT HIV-1 INHIBITORY ACTIVITY

20. Greenberg, M.L., S.F. Lacey, C.H. Chen, D.P. Bolognesi, and K.J. Weinhold. 1997. Noncytolytic CD8 T cell-me-diated suppression of HIV replication. Springer Semin Immunopathol. 18:355-369.

21. Imai, T., D. Chantry, C.J. Raport, C.L. Wood, M. Nishimura, R. Godiska, O. Yoshie, and P.W. Gray. 1998.Macrophage-derived chemokine is a functional ligand for the CC chemokine receptor 4. J. Biol. Chem.273:1764-1768.

22. King, A., N. Balendran, P. Wooding, N.P. Carter, and Y.W. Loke. 1991. CD3-leukocytes present in the humanuterus during early placentation: Phenotypic and morphologic characterization of the CD56++ population. Dev.Immunol. 1:169-190.

23. Kinter, A., A. Catanzaro, J. Monaco, M. Ruiz, J. Justement, S. Moir, J. Arthos, A. Oliva, L. Ehler, S. Mizell, R.Jackson, M. Ostrowski, J. Hoxie, R. Offord, and A.S. Fauci. 1998. CC-chemokines enhance the replication of T-tropic strains of HIV-1 in CD4(+) T cells: Role of signal transduction. Proc. Nati. Acad. Sei. USA. 95:11880-11885.

24. Koup, R.A., J.T. Safrit, Y. Cao, CA. Andrews, G. McLeod, W. Borkowsky, C. Farthing, and D.D. Ho. 1994. Tem-poral association of cellular immune responses with the initial control of viremia in primary human immunodefi-ciency virus type 1 syndrome. J. Virol. 68:4650-4655.

25. Lacey, S.F., C.B. McDanal, R. Homk, and M.L. Greenberg. 1997. The CXC chemokine stromal cell-derived fac-tor 1 is not responsible for CD8+ T cell suppression of syncytia-inducing strains of HIV-1. Proc. Nati. Acad. Sei.USA. 94:9842-9847.

26. Lacey, S.F., K.J. Weinhold, C.H. Chen, C. McDanal, C. Oei, and M.L. Greenberg. 1998. Herpesvirus saimirí trans-formation of HIV type 1 suppressive CD8+ lymphocytes from an HIV type 1-infected asymptomatic individual.AIDS. Res. Hum. Retrovimses. 14:521-531.

27. Leith, J.G., K.F. Copeland, P.J. McKay, CD. Richards, and K.L. Rosenthal. 1997. CD8+ T-cell-mediated sup-pression of HIV-1 long terminal repeat-driven gene expression is not modulated by the CC chemokines RANTES,macrophage inflammatory protein (MIP)-l alpha and MIP-1 beta. AIDS. 11:575-580.

28. Levy, J.A., F. Hsueh, D.J. Blackboum, D. Wara, and P.S. Weintmb. 1998. CD8 cell noncytotoxic antiviral activ-ity in human immunodeficiency vims-infected and -uninfected children. J. Infect. Dis. 177:470-472.

29. Levy, J.A., CE. Mackewicz, and E. Barker 1996. Controlling HIV pathogenesis: The role of the noncytotoxic anti-HIV response of CD8+ T cells. Immunol. Today. 17:217-224.

30. Mackewicz, C, and J.A. Levy. 1992. CD8+ cell anti-HIV activity: Nonlytic suppression of vims replication. AIDS.Res. Hum. Retrovimses. 8:1039-1050.

31. Mackewicz, CE., D.J. Blackboum, and J.A. Levy. 1995. CD8+ T cells suppress human immunodeficiency vimsreplication by inhibiting viral transcription. Proc. Nati. Acad. Sei. USA. 92:2308-2312.

32. Mackewicz, CE., R. Orque, J. Jung, and J.A. Levy. 1997. Derivation of herpesvirus saimiri-transformed CD8+ Tcell lines with noncytotoxic anti-HIV activity. Clin. Immunol. Immunopathol. 82:274-281.

33. Mackewicz, CE., H.W. Ortega, and J.A. Levy. 1991. CD8+ cell anti-HIV activity correlates with the clinical stateof the infected individual. J. Clin. Invest. 87:1462-1466.

34. Mackewicz, C.E., L.C. Yang, J.D. Lifson, and J.A. Levy. 1994. Non-cytolytic CD8 T-cell anti-HIV responses inprimary HIV-1 infection. Lancet. 344:1671-1673.

35. Mellors, J.W., L.A. Kingsley, C.R. Rinaldo, Jr., J.A. Todd, B.S. Hoo, R.P. Kokka, and P. Gupta. 1995. Quantitä-ten of HIV-1 RNA in plasma predicts outcome after seroconversion. Ann. Intern. Med. 122:573-579.

36. Mellors, J.W., C.R. Rinaldo, Jr., P. Gupta, R.M. White, J.A. Todd, and L.A. Kingsley. 1996. Prognosis in HIV-1infection predicted by the quantity of vims in plasma [see comments] [published erratum appears in Science 1997Jan 3;275(5296):14]. Science. 272:1167-1170.

37. Moriuchi, H., M. Moriuchi, C, Combadiere, P.M. Murphy, and A.S. Fauci. 1996. CD8+ T-cell-derived solublefactor(s), but not beta-chemokines RANTES, MIP-1 alpha, and MIP-1 beta, suppress HIV-1 replication in mono-

cyte/macrophages. Proc. Nati. Acad. Sei. USA. 93:15341-15345.

38. Mosoian, A., A. Teixeira, and M.E. Klotman. 1998. Soluble HIV-1 suppressive effects of herpesvirus saimiri-trans-formed CD8+ cells are distinct from RANTES, MIP-la and MIP-lb. Proceedings of Colloque Des Cent Gardes,Retroviruses of Human AIDS and Related Animal Diseases, Paris, 1998, pp. 151-158.

493

MOSOIAN ET AL.

39. Musey, L., J. Hughes, T. Schacker, T. Shea, L. Corey, and M.J. McElrath 1997. Cytotoxic-T-cell responses, viralload, and disease progression in early human immunodeficiency virus type 1 infection [see comments]. N. Engl.J. Med. 337:1267-1274.

40. Nick, S., H. Fickenscher, B. Biesinger, G. Bom, G. Jahn, and B. Fleckenstein. 1993. Herpesvirus saimirítransformed human T cell lines: A permssive system for human immunodeficiency vimses. Virology. 194:875-877.

41. Oberlin, E., A. Amara, F. Bachelerie, C. Bessia, J.L. Virelizier, F. Arenzana-Seisdedos, O. Schwartz, J.M. Heard,I. Clark-Lewis, D.F. Legier, M. Loetscher, M. Baggiolini, and B. Moser. 1996. The CXC chemokine SDF-1 is theligand for LESTR/fusin and prevents infection by T-cell-line adapted HIV-1 [published erratum appears in Nature1996 Nov 21;384(6606):288]. Nature. 382:833-835.

42. Oliva, A., A.L. Kinter, M. Vaccarezza, A. Rubbert, A. Catanzaro, S. Moir, J. Monaco, L. Ehler, S. Mizell, R. Jack-son, Y. Li, J.W. Romano, and A.S. Fauci. 1998. Natural killer cells from human immunodeficiency virus (HIV)-infected individuals are an important source of CC-chemokines and suppress HIV-1 entry and replication in vitro.J. Clin. Invest. 102:223-231.

43. Pal, R., A. Garzino-Demo, P.D. Markham, J. Bums, M. Brown, R.C. Gallo, and A.L. DeVico. 1997. Inhibition ofHIV-1 infection by the beta-chemokine MDC [see comments]. Science. 278:695-698.

44. Pantaleo, G., S. Menzo, M. Vaccarezza, C. Graziosi, O.J. Cohen, J.F. Demarest, D. Montefiori, J.M. Orenstein, C.Fox, L.K. Schräger, J.R. Margolick, S. Buchbinder, J.V. Giorgi, and A.S. Fauci. 1995. Studies in subjects withlong-term nonprogressive human immunodeficiency vims infection [see comments]. N. Engl. J. Med. 332:209-216.

45. Pinto, L.A., V. Blazevic, B.K. Patterson, C Mac Tmbey, M.J. Dolan, and G.M. Shearer. 2000. Inhibition of hu-man immunodeficiency vims type 1 replication prior to reverse transcription by influenza vims stimulation. J. Vi-rol. 74:4505-4511.

46. Poli, G., J.M. Orenstein, A. Kinter, T.M. Folks, and A.S. Fauci. 1989. Interferon-alpha but not AZT suppressesHIV expression in chronically infected cell lines. Science. 244:575-577.

47. Pollack, H., M.X. Zhan, J.T. Safrit, S.H. Chen, G. Rochford, P.Z. Tao, R. Koup, K. Krasinski, and W. Borkowsky.1997. CD8+ T-cell-mediated suppression of HIV replication in the first year of life: Association with lower viralload and favorable early survival. AIDS. 11:F9—13.

48. Rosok, B., P. Voltersvik, B.M. Larsson, J. Albeit, J.E. Brinchmann, and B. Asjo. 1997. CD8+ T cells from HIVtype 1-seronegative individuals suppress vims replication in acutely infected cells. AIDS. Res. Hum. Retrovimses.13:79-85.

49. Rubbert, A., D. Weissman, C. Combadiere, K.A. Pettrone, J.A. Daucher, P.M. Murphy, and A.S. Fauci. 1997. Mul-tifactorial nature of noncytolytic CD8+ T cell-mediated suppression of HIV replication: Beta-chemokine-depen-dent and -independent effects. AIDS. Res. Hum. Retrovimses. 13:63-69.

50. Stranford, S.A., J. Skumick, D. Louria, D. Osmond, S.Y. Chang, J. Sninsky, G. Ferrari, K. Weinhold, C. Lindquist,and J.A. Levy. 1999. Lack of infection in HIV-exposed individuals is associated with a strong CD8(+) cell non-

cytotoxic anti-HIV response. Proc. Nati. Acad. Sei. USA. 96:1030-1035.

51. Struyf, S., P. Proost, S. Sozzani, A. Mantovani, A. Wuyts, E. De Clercq, D. Schols, and J. Van Damme. 1998. En-hanced anti-HIV-1 activity and altered chemotactic potency of NH2-terminally processed macrophage-derivedchemokine (MDC) imply an additional MDC receptor. J. Immunol. 161:2672-2675.

52. Walker, CM., A.L. Erickson, F.C. Hsueh, and J.A. Levy. 1991. Inhibition of human immunodeficiency vimsreplication in acutely infected CD4+ cells by CD8+ cells involves a noncytotoxic mechanism. J. Virol.65:5921-5927.

53. Walker, CM., and J.A. Levy. 1989. A diffusible lymphokine produced by CD8+ T lymphocytes suppresses HIVreplication. Immunology. 66:628-630.

54. Weber, F., E. Meinl, K. Drexler, A. Czlonkowska, S. Huber, H. Fickenscher, I. Muller-Fleckenstein, B. Flecken-stein, H. Wekerle, and R. Hohlfeld. 1993. Transformation of human T-cell clones by Herpesvirus saimirí: Intactantigen recognition by autonomously growing myelin basic protein-specific T cells. Proc. Nati. Acad. Sei. USA.90:11049-11053.

494

CD8+ CELL LINES WITH POTENT HIV-1 INHIBITORY ACTIVITY

55. Yang, O.O., S.A. Kalams, A. Trocha, H. Cao, A. Luster, R.P. Johnson, and B.D. Walker, 1997. Suppression of hu-man immunodeficiency vims type 1 replication by CD8+ cells: Evidence for HLA class I-restricted triggering ofcytolytic and noncytolytic mechanisms. J. Virol. 71:3120-3128.

Address reprint requests to:Mary E. Klotman

Box 1090Mt. Sinai School of Medicine

1 Gustave L. Levy PI.New York, NY. 10029

E-mail: Mary.Klotman@MSSM.eduReceived April 17, 2000; accepted August 17, 2000

495