Bethesda, Maryland 20892,1 and Department ofPathology, George Washington University Medical Center,. Washington, D.C. 20007. Received 13 September ...
Vol. 68, No. 4
JOURNAL OF VIROLOGY, Apr. 1994, p. 2598-2604
0022-538X/94/$04.00+0
Copyright X 1994, American Society for Microbiology
Cytokine-Mediated Induction of Human Immunodeficiency Virus (HIV) Expression and Cell Death in Chronically Infected Ul Cells: Do Tumor Necrosis Factor Alpha and Gamma Interferon Selectively Kill HIV-Infected Cells? PRISCILLA BISWAS,1* GUIDO POLI,1t JAN M. ORENSTEIN,2 AND ANTHONY S. FAUCI' Laboratory of Immunoregulation, National Institute of Allergy and Infectious Diseases, National Institutes of Health, Bethesda, Maryland 20892,1 and Department of Pathology, George Washington University Medical Center, Washington, D.C. 20007 Received 13 September 1993/Accepted 5 January 1994
Infection with several DNA or RNA viruses induces a state of increased sensitivity to cell lysis mediated by tumor necrosis factor (TNF), particularly in the presence of gamma interferon (IFN-y). Infection of human cells with the human immunodeficiency virus (HIV) may induce a similar phenomenon. However, TNF and IFN-'y are known upregulators of HIV replication, raising the question of the potential role of these cytokines in the selective elimination of cells infected with this virus. The present study demonstrates that chronically infected Ul cells were killed with much greater efficiency by costimulation with TNF-a and IFN-y than their uninfected parental cell line U937. However, synergistic induction of viral expression also occurred in Ul cells as a consequence of treatment with the two cytokines. Cell death in Ul cells was not caused by the massive production of virions, in that costimulation with glucocorticoid hormones and TNF-ae or IFN-y resulted in high levels of virion production without cytopathicity. To investigate the nature of the selective cytotoxic effect observed in Ul cells costimulated with TNF-a plus IFN--y, a panel of uninfected cell clones was generated by limiting dilution of U937 cells and tested for response to TNF-a and/or IFN-,y. In contrast to the uncloned bulk parental U937 cell line, most uninfected cell clones showed a very high susceptibility to being killed by TNF-oa and IFN-y. Similar findings were obtained when both infected Ul cells and several uninfected U937 cell clones were costimulated with an anti-Fas monoclonal antibody in the presence of IFN-'y, although, unlike cells stimulated with TNF-a, cells treated with anti-Fas antibody did not express virus. Therefore, the increased susceptibility to cytokine-mediated lysis observed in cell lines infected with HIV is likely due to the selection of preexisting cell clones rather than viral infection.
with inconclusive results in terms of whether cytokines were beneficial or accelerated disease progression (2). Other in vitro studies (19) confirmed that TNF could mediate lysis of several HIV-infected cell lines but not of their parental uninfected counterparts. However, in contrast to an earlier report (25), cell lysis was associated with increased rather than decreased virus expression (19). Furthermore, the promonocytic U937 cell line appeared to be more susceptible to TNF-induced lysis before rather than after infection with HIV (19). Recently, the possibility of triggering lysis of TNF-sensitive target cells without inducing the regulatory effects associated with cell stimulation with this cytokine (such as activation of NF-KB) has been demonstrated by use of monoclonal antibodies (MAbs) directed at the cell surface Fas antigen (27). This antigen belongs to the TNF-receptor family, and its expression appears to play a major role in different models of immune dysfunction and cell death (7). With regard to HIV infection, it has been previously reported that anti-Fas MAbs can kill HIV-infected cells without inducing viral expression (16). In the present study, we have investigated the effects of the simultaneous treatment of HIV-infected Ul cells and their parental uninfected U937 cell line with TNF-ao and IFN--y. Our results support the interpretation that selection of cell clones which were highly susceptible to cytokine-mediated cell death and were already present in the bulk cell line before HIV infection is the most probable explanation of the apparently increased lysis of chronically infected Ul cells compared with that of their uninfected counterpart.
Several cytokines have been implicated in the immunopathogenesis of human immunodeficiency virus type 1 (HIV-1) infection, both as potential mediators of the immune dysfunction typical of this disease and as direct modulators of virus expression (22). In particular, tumor necrosis factor alpha (TNF-a) (13, 18) is known to induce HIV-1 transcription and expression via activation of the cellular transcription factor NF-KB, which binds to the promoter-enhancer region of the virus long terminal repeat (9, 15, 20). Interferon-gamma (IFN-y) can also activate HIV expression in several cells (17, 22), including the chronically infected promonocytic cell line Ul (3). However, earlier reports indicated that treatment of HIV-1-infected cell lines with TNF resulted in a decrease of viral RNA synthesis, particularly in the presence of IFN--y (25). This possibility was consistent with earlier studies that demonstrated that infection of cells with several DNA or RNA viruses induced a state of increased susceptibility to cell lysis by TNF (6, 14, 24). On the basis of these early findings, clinical trials involving the individual and combined use of TNF-ot and IFN--y have been conducted with HIV-infected individuals,
* Corresponding author. Mailing address: Laboratory of Immunoregulation, National Institute of Allergy and Infectious Diseases, National Institutes of Health, Building 10A, Room 6A33 A, 9000 Rockville Pike, Bethesda, MD 20892. t Present address: AIDS Immunopathogenesis Unit, DIBIT, San Raffaele Scientific Institute, Milano, Italy.
2598
HIV-INFECTED CELL KILLING BY TNF-a AND IFN--y
VOL. 68, 1994
MATERUILS AND METHODS Cell lines. Uninfected U937 cells were purchased from the American Type Culture Collection (ATCC CRL 1593) and were maintained at low density (1 x 105 to 5 x 105 cells per ml) in RPMI 1640 (Whittaker M. A. Bioproducts, Walkersville, Md.) supplemented with 2 mM glutamine (Biofluids, Inc., Rockville, Md.), 10 mM HEPES (N-2-hydroxyethylpiperazineN'-2-ethanesulfonic acid) (Biofluids, Inc.), and 10% fetal calf serum (Whittaker M. A. Bioproducts) (complete medium). U937 cells were cloned by limiting endpoint dilution as previously described (11). Briefly, 0.2 cells per well in 0.2 ml of complete medium were seeded in 96-well round-bottom plates for tissue culture (Costar, Cambridge, Mass.) and were incubated at 37°C in 5% CO2. Cell clones, which began to be detectable as microclusters of a few cells after approximately 14 days, were further expanded, first in 1 ml of medium in 24-well plates (Costar) and then in to tissue culture flasks. The chronically infected Ul cell line was originally derived from the expansion of a single cell clone obtained by limiting dilution of U937 cells surviving an acute infection with HIV-1 (11) and is characterized by a state of relative viral latency, as defined by the constitutive presence of spliced, but not unspliced, mRNAs (23). These cells can be converted to high levels of virus production after cell stimulation with phorbol myristate acetate (11) or several cytokines (for a review, see reference 22). No evidence of mycoplasma contamination of the different U937 and Ul cell cultures was observed by both a commercially available mycoplasma detection kit (Gene-Probe) and by transmission electron microscopy. Cytokines and reagents. Recombinant human (rHu) TNF-a and IFN--y were purchased from Genzyme (Genzyme Corp., Boston, Mass.). The glucocorticoid hormone dexamethasone (DEX) was purchased from Sigma (Sigma Chemical Corp., St. Louis, Mo.). Anti-Fas monoclonal antibody (immunoglobulin M) was generously donated by Shin Yonehara (The Tokyo Metropolitan Institute of Medical Science, Tokyo, Japan) and was used in some experiments in the same range of concentrations previously shown to affect HIV-infected cells (16). Flow cytofluorometric analysis. The presence of cell surface receptors for TNF and IFN-y was evaluated on U937 clones 17 and 30, as well as on the Ul cell line, by cytofluorometric analysis on an Epic profile (Coulter, Hialeah, Fla.). For determination of IFN-y receptors, cells were stained with a mouse anti-human IFN--y receptor MAb (Genzyme), washed and stained with a fluorescein isothiocyanate-conjugated sheep anti-mouse Ab. For determination of TNF receptors, cells were directly stained with recombinant human TNF-a labeled with phycoerythrin (Fluorokine, R&D Systems, Minneapolis, Minn.). Cells were also stained with the fluorescein isothiocyanate-conjugated anti-CD14 MAb (MY4; Coulter). Cytotoxicity studies. Viability and proliferative capacity of the different cells were monitored by Trypan blue dye exclusion and 3H-labeled thymidine incorporation, respectively. For Trypan blue dye exclusion determinations, cell suspensions were diluted 1:1 (vol/vol) in a 0.4% solution of Trypan blue (Sigma) at different time points after the beginning of the cultures and counted in a hemocytometer chamber. Both total cell numbers and percentage of viable cells (i.e., the fraction of cells unstained by Trypan blue) were evaluated. For studies of [3H]thymidine uptake, cell cultures (1 x 105 to 2 x 105 cells per ml), carried out in duplicate or triplicate in 96-well flat-bottom microtiter plates, were pulsed overnight with 0.5 ,uCi of [3H]thymidine (DuPont, NEN Products, Boston, Mass.) per well and harvested by an automatic 96-well
2599
plate harvester-washer (Tomtec, Orange, Conn.), and the glass fiber filters were counted in a liquid scintillation counter (1205 Betaplate; Wallac Inc., Gaithersburg, Md.). The percentage of [3H]thymidine uptake was calculated by dividing the counts per minute of the samples by the counts per minute of unstimulated cells and then multiplying by 100 (100% being the arbitrary value given to the counts per minute of unstimulated cells). Reverse transcriptase (RT) activity assay. Culture supernatants of Ul cells were collected at various time points after stimulation and stored at - 70°C until tested. Five microliters of supernatant was added in duplicate or triplicate to 25 ,ul of a mixture containing poly(A), oligo(dT) (Pharmacia Fine Chemicals, Piscataway, NJ), MgCl2, and 32P-labeled dTTP (Amersham Corp., Arlington Heights, Ill.), and the mixture was incubated for 2 h at 37°C. Six microliters of the mixture was spotted onto DE81 paper, air dried, washed five times in 2 x SSC buffer (1 x SSC is 0.15 M NaCl plus 0.015 M sodium citrate) and twice in 95% ethanol. The paper was then dried, cut, and counted in a liquid scintillation counter (LS 5000; Beckman Instruments., Inc., Fullerton, Calif.). The variability of culture replicates was always less than 15%. Western blot (immunoblot) analysis of HIV proteins. A total of 107 Ul cells either unstimulated or stimulated with TNF-a, IFN--y, or TNF-a plus IFN--y for 72 h were pelletted and lysed, and the total protein content of each lysate was measured with the use of a spectrophotometer (DU-64; Beckman) and referred to a standard curve of plasma gamma globulin (4). Twenty-five microliters of each lysate, standardized at 10 jig/pul, was added to each lane and subjected to vertical electrophoresis through 3 to 27% gradient polyacrylamide gels (Integrated Separation Systems; Enprotech Co., Natick, Mass.) for 6 to 8 h. The migrated proteins were transferred from the gels onto nitrocellulose filters, which were then saturated with a 5% milk solution for 1 h and incubated for 1 h with a 1:500 (vol/vol) dilution of pooled AIDS patients' serum samples containing Abs recognizing most of the major HIV proteins (3). The filters were then washed and incubated for 1 h with sheep anti-human immunoglobulin F(ab')2 fragments conjugated with horseradish peroxidase (Amersham). Finally, the filters were covered for 1 min with the detection reagents (ECL, Amersham), dried, and exposed to X-ray film for 1 to 2 s. Northern (RNA) blot analysis of HIV mRNAs. Ul cells (0.5 x 106 cells/ml) were either unstimulated or stimulated with TNF-a (100 U/ml) or IFN--y (1,000 U/ml) or costimulated with TNF-ot plus IFN--y for 12 and 24 h. Total RNA was extracted from 2 x 107 Ul cells by the guanidine thiocyanate phenol method with an RNA isolation kit (Stratagene, La Jolla, Calif.). Ten micrograms of the extracted total RNA was separated by 0.8% agarose formaldehyde gel electrophoresis and transferred to nitrocellulose filters which were baked, hybridized to a 32P-labeled HIV DNA fragment (SST-BssHII), washed, and exposed to X-ray film. The labeled probe was stripped by washing the filters at 85°C for 10 min in 0.1 x SSC containing 0.1% sodium dodecyl sulfate. The filters were then rehybridized with a 32P-labeled ,B-actin cDNA probe as a control (3). Nuclear run-on analysis of HIV transcription. Nuclei from 108 Ul cells were isolated by sucrose gradients and incubated 30 min with 32P-UTP (Amersham) and a mixture of nucleotides (ATP, CTP, and TTP; Amersham) according to a published procedure (21). 32P-labeled RNA was then extracted from the nuclei and hybridized for 36 h to plasmid DNA probes which were previously linearized and immobilized on nitrocellulose filters. The probes used were pUC19 (plasmid
2600
BISWAS ET AL.
A.
Cytolysis
J. VIROL.
B.
Cytostasis
-0
A.
~z
~~~~~~U-
m
_=3
+
a$
E
cnU-
gpl
LL
z
c
z
20
n
0)
5C3)(a) -.O
* '0
om
.. .i
p24
Unstimulated
TNF-a
IFN-y
(10 U/ml) (1000 U/ml)
TNF-a+lFN-y
Unstimulated
TNF-a TNF-a,IFN-y IFN-y (100 U/rn) (1000 U/Wm)
control), pNL4-3 (which contains a full-length HIV genome [1]), and a human P-actin cDNA. The filters were then extensively washed and exposed to X-ray film. Quantitation of the levels of RNA was obtained by laser scanning densitometry of the autoradiograms with a 2222-10 apparatus (LKB Instruments, Inc., Gaithersburg, Md.).
-o
B.
FIG. 1. Differential sensitivity of HIV-infected Ul cells and their parental uninfected U937 cell line to the cytolytic and antiproliferative effects of TNF-oo plus IFN-y. Cell viability and proliferation in cells stimulated with the two cytokines either alone or in combination were evaluated by Trypan blue dye exclusion (A) and [3H]thymidine uptake (B) criteria. A higher susceptibility of HIV-infected Ul cells to the cytostatic and cytolytic effects of TNF-a plus IFN-y was observed as compared to uninfected U937 cells. Three independent experiments were performed with similar results.
,
kb
,.
_
z-
:3
+
E
a
a
CD
u-L
IL
z_
D
*0
a)
z
.-3
+
.E Co:z
s z
z
I
_L
H
9.0
f
7
45
> >>>t= tiit
2.0-
inuiumi'n0F 24h
1 2h
P-actin C. Unstimulated
IFN-y
TNF-a
TNF-ax + IFN-y -pUC
RESULTS
HIV -
The combination of TNF-a and IFN-y induces death of HIV-infected, but not uninfected, U937 cells. Both HIVinfected Ul cells and their parental uninfected cell line U937 were exposed to TNF-ot (100 U/ml) and IFN--y (1,000 U/ml) either alone or in combination. Treatment of Ul cells with these concentrations of individual cytokines was previously shown to be maximally effective in inducing HIV expression in the absence of cytopathicity (3, 21). A reduction in the proliferative capacity of Ul cells in the presence of IFN-,y alone was noted, although without a decrease in cell viability (Fig. 1A and 1B). TNF-ot alone had no effect on either the viability or proliferative capacity of Ul cells. However, a striking cytopathic effect was observed when Ul cells were costimulated with these two cytokines (Fig. 1). Ultrastructural studies by transmission electron microscopy confirmed that a high degree of cytopathic effects occurred in Ul cells costimulated with TNF-ot plus IFN--y (data not shown). In contrast, uninfected U937 cells were not substantially affected by TNF-aL and IFN--y either alone or in combination (Fig. 1), despite the fact that these cells are usually considered susceptible to TNF-mediated cell lysis (19, 25). These results are consistent with the hypothesis that HIV infection may have induced a state of higher susceptibility to cytokine-dependent cell death in U937 cells, as previously observed in other cell lines (6, 19, 24, 25). TNF-a and IFN-y synergistically induce HIV expression in Ul cells. Concomitant with cytopathicity, costimulation with TNF-ot and IFN-,y resulted in the expression of a greater magnitude of viral proteins than that observed with IFN-,y or TNF-ao alone (Fig. 2A). Furthermore, Ul cells costimulated
13-Actin
FIG. 2. Costimulation of Ul cells with TNF-cx and IFN--y results in a synergistic induction of HIV proteins and RNA. (A) Ul cell lysates were prepared after 72 h of incubation with the different stimuli. IFN--y (1,000 U/ml) and TNF-co (100 U/ml) alone induced detectable levels of cell-associated HIV proteins which were increased in Ul cells costimulated with TNF-as and IFN-y. Enhanced viral proteins in TNF-o plus IFN-,y costimulated Ul cell lysates were observed also at earlier time points (24 and 48 h; data not shown). Three independent experiments were performed with similar results. (B) Northern (RNA) blot analysis was performed with total RNA extracted from Ul cells stimulated with IFN--y (1,000 U/ml), TNF-ot (100 U/ml), or TNF-ot plus IFN--y after 12 and 24 h of incubation. The results shown are obtained from one experiment representative of five independently performed. The synergistic induction of HIV steady-state mRNAs by TNF-ot plus IFN--y was present also at later time points (48 h) and with different concentrations of the two cytokines (TNF-ot [10 U/ml] plus IFN-,y [1,000 U/ml] and TNF-as [100 U/ml] plus IFN--y [100 U/ml]; data not shown). (C) Nuclear run-on analyses were performed on newly synthesized mRNA extracted from nuclei isolated after 24 h of incubation of Ul cells stimulated with IFN--y (1,000 U/ml), TNF-ot (100 U/ml) or the combination of the two cytokines. The autoradiograms were then analyzed by laser scanning densitometry to obtain semi-quantitative values of the levels of HIV transcription.
with TNF-cx and IFN--y showed a clear synergy in terms of accumulation of HIV steady-state mRNAs as early as 12 h poststimulation, when the inductive effect of IFN--y alone was barely detectable (Fig. 2B). Stimulation of Ul cells with TNF-a alone induced an approximately fivefold increase of HIV transcription over unstimulated cells as assessed by nuclear run-on analysis followed by laser scanning densitome-
HIV-INFECTED CELL KILLING BY TNF-ot AND IFN-y
VOL. 68, 1994 A.
Unstimulated
0
c-500
no cytopathicity as indicated by [3H]thymidine incorporation (Fig. 3B). Similar results were obtained when Ul cells were stimulated with IFN--y plus DEX (Fig. 3C and 3D). Thus, the profound cytopathicity observed in Ul cells costimulated with TNF-ot and IFN--y was unlikely to be a consequence of high
B.
0OTNF-a (100 U/ml) A/ Glucocorticoids (10-8 M) 0 TNF-a + Glucocorticoids
600 -
400 t 300 ° '
200
cr 100
O Unstimulated
C.
O IFN-y (1000 U/mi) A Glucocorticoids (10i M) * IFN-y + Glucocoricoids
8(o0 61
E -
-S
40O
.H
2M 0~~~~~~~6
0
2
4 Days
6
8
Jv~-
.n
\0N1
1
0
\
FIG. 3. The combinations of TNF-ot plus glucocorticoids and
IFN-y plus glucocorticoids synergistically induce HIV production but are not cytopathic for Ul cells. Ul cells were stimulated with TNF-cs, IFN--y, or glucocorticoids alone and with the combination of TNF-a plus glucocorticoids (panel A and B) and IFN--y plus glucocorticoids (panel C and D). Culture supernatants were harvested 3 and 6 days
after stimulation and tested for RT activity. At day 6, Ut cells under different treatment conditions were also pulsed overnight with [3H]thymidine. The results shown are obtained from one experiment of three independently performed.
(Fig. 2C). Although IFN--y alone did not induce a clear effect, the combination of TNF-ot and IFN--y resulted in a synergistic effect on HIV transcription estimated to be an approximately 15-fold increase over that of unstimulated cells (Fig. 2C). This synergistic increase of HIV transcription observed in Ul cells costimulated with TNF-ot and IFN--y was confirmed by studies in which a heterologous HIV-1 long terminal repeat linked to the reporter gene chloramphenicol acetyltransferase was transfected into Ul cells (data not shown). Death of Ul cells is selectively caused by costimulation with TNF-a and IFN--y and is not a consequence of the massive production of HIV particles. Because massive cell death was associated with the synergistic induction of HIV expression in Ul cells costimulated with TNF-ot and IFN-y, we investigated whether the observed cytopathicity was caused by the high levels of virus production, as suggested earlier in other in vitro models of HIV infection (28). In this regard, glucocorticoid hormones such as DEX have been shown recently to synergize potently with several cytokines in the induction of HIV expression in Ul cells (5). Therefore, we investigated the correlation between levels of virus expression and cytopathicity in U1 cells costimulated with either TNF-a or IFN--y plus DEX. Supernatant-associated RT activity was determined as a parameter of virion production (10), and [3H]thymidine incorporation of Ul cells under each condition was determined simultaneously. DEX alone had no substantial effect on virus expression in Ul cells but synergized with TNF-ca (Fig. 3A), as reported previously (5). Despite high virus production, cell cultures showed
try
2601
levels of virus production synergistically induced by these two cytokines, since other similar synergistic costimulatory conditions did not substantially affect cell viability. Clonal variability in the susceptibility of U937 cells to cytokine-mediated killing. The observation that Ul cells are much more sensitive to cytokine-mediated killing than their uninfected counterparts is in contrast to previous observations indicating that HIV infection of U937 cells led to a decreased sensitivity to the antiproliferative effect of TNF-a (19). Therefore, we investigated whether clonal variability within different U937 cells could provide an explanation for this discrepancy. Uninfected U937 cells were cloned by a limiting-dilution technique (11). Forty-five clones were obtained, expanded in culture for 14 days, and tested for their sensitivity to the antiproliferative effect of TNF-a and IFN--y, either alone or in combination. Surprisingly, despite the lysis-resistant phenotype of the uncloned U937 cell line and a heterogenous susceptibility to cytokine-mediated antiproliferative effects that was observed among the different clones in response to each cytokine alone, the vast majority of the clones were killed by the costimulation of TNF-a plus IFN--y (Fig. 4A). To investigate this apparent paradox, reconstitution experiments in which one lysis-resistant clone (clone 30) was added in equal proportions to a panel of sensitive clones were performed. The mixture of these different clones showed a substantial resistance to the antiproliferative or cytotoxic effects of TNF-ax plus IFN-y (Fig. 4B). Similarly, mixing experiments in which Ul and clone 30 cells were seeded in a 1:1 ratio resulted in the appearance of a culture phenotype that was relatively resistant to the antiproliferative effect of TNF-a plus IFN-y (data not shown). No evidence was obtained that the resistant phenotype conferred by clone 30 was dependent upon the release of a soluble factor, in that the addition of supernatants from clone 30 cultures to lysis-sensitive U937 clones or Ul cells did not alter their susceptibility to the antiproliferative effect of TNF-at
plus IFN--y (data not shown). Having observed the existence of cytokine-sensitive clones within the bulk U937 cell line, we compared the susceptibility of infected Ul cells to killing by TNF-a plus IFN--y to that of a representative uninfected sensitive clone (clone 34). TNF-a and IFN--y alone at the highest concentrations used (100 and 1,000 U/ml, respectively) showed a mild antiproliferative effect (ranging approximately from a minimum of 10% to a maximum of 40% over the control level) on both cell lines (Fig. 5). The cytolytic effect of the cytokines at these concentrations was
also modest (ranging from 0 to 20%) (Fig. 1A). However, when TNF-ot and IFN--y were combined, a very strong and comparable concentration-dependent antiproliferative effect was observed in Ul cells (Fig. 5, upper panel) and clone 34 (Fig. 5, lower panel). The lysis-resistant U937 clone 30 possesses cell surface receptors for TNF or IFN--y. To rule out the possibility that clone 30 was a trivial contaminant of the U937 cell line, studies of the cell surface phenotype were conducted. No significant expression of T-lymphocytic (CD3), B-lymphocytic (CD20), or natural killer cell (CD56) markers were expressed on the cell surface of either uncloned U937 cells, clone 30, or other clones (data not shown), whereas all the cells expressed comparable levels of the CD14 antigen characteristic of the monocytic lineage (26) (Fig. 6). We also examined the possibility that the resistance to TNF-a-plus-IFN-y-mediated cell lysis manifested
J. VIROL.
BISWAS ET AL.
2602
A.
B.
100
ii
80
c(C: a) C
Li
C
60
.0
.a
E
E
co
.-C
ICZ-
40
20
0 Unstimulated
IFN-,y (1000 U/ml)
TNF-a TNF-a+IFN--y (100 U/mI)
TNF-a+IFN-y
Unstimulated
FIG. 4. Heterogeneous sensitivity of U937 clones to the cytostatic or cytolytic effects of TNF-ox and/or IFN-'y. Cell proliferation was measured by [3H]thymidine incorporation 72 h after cytokine treatment. (A) The relative insensitivity of the bulk U937 cell line is not reflected in the majority of the U937 clones obtained (a selected and representative panel of a total of 45 clones is shown). In particular, the simultaneous addition of TNF-ca plus IFN--y resulted in the lysis of all clones but clone 30, which was essentially unaffected in its proliferative ability by the presence of TNF-ot and/or IFN--y. (B) Mixing of clone 30 (arrow) with different lysis-sensitive clones partially reconstitutes a lysis-resistant phenotype to the cell culture following costimulation with TNF-ot and IFN--y. These results are representative of four (A) and three (B) experiments, respectively, that were independently performed.
Ul
120 r
by clone 30 was due to a lack of surface receptors for these cytokines. However, both lysis-resistant (clone 30) and lysissensitive (clone 17) uninfected clones of U937 showed levels of cell surface TNF and IFN-,y receptors comparable to those present on the plasma membrane of Ul cells (Fig. 6). Anti-Fas MAb synergizes with IFN-y in triggering cell death in Ul cells and several uninfected U937 clones without inducing HIV expression. It has been previously reported that anti-Fas MAb can induce lysis of TNF-sensitive target cells, including HIV-infected cells, without transducing a cell activation signal such as the activation of NF-KB (7, 16, 27). Anti-Fas MAb alone did not cause substantial cytopathicity in either HIV-infected Ul cells or in several uninfected U937 cell clones (Fig. 7). However, the combination of anti-Fas MAb and IFN-,y resulted in a synergistic antiproliferative effect both in Ul cells and in the majority of the uninfected U937 clones (Fig. 7). As previously observed in cells costimulated with TNF-ot and IFN-,y, anti-Fas plus IFN--y did not cause substantial cytopathicity in uninfected bulk U937 and clone 30 (Fig. 7). Consistent with previous reports (16), and in contrast to TNF-ox or IFN-y, anti-Fas MAb did not cause upregulation of HIV expression in Ut cells when used either alone or in combination with IFN-y or TNF-o (data not shown). These data further emphasize that the death of Ut cells observed with stimulation by TNF-ot and IFN-y was not the consequence of massive production of newly synthesized virions. DISCUSSION The present study has demonstrated that costimulation of Ul cells with the HIV-inductive cytokines TNF-ox and IFN-,y
(D ~-
X'
80
a c a)
r o 60 E E
± -.,
40
TNF-a (U/ml)
120
Clone 34
r
100
G
E
80
o 60
-
>, E I
±00.
O
4
40
-
O
*
20
*
A
100
10
1
IFN-yalone (U/ml) TNF-a alone + IFN-y 1000 (U/ml) +IFN-y1OO(U/ml) + IFN-y 10 (U/ml)
0.1
TNF-a (U/ml)
FIG. 5. Concentration-dependent effects of TNF-o and IFN--y alone and in combination on the proliferation of Ul cells and clone 34. Serial 10-fold dilutions were used starting from 100 U/ml of TNF-ax and 1,000 U/ml of IFN--y. The titration curves did not reveal substantially different susceptibilities to cytokine-mediated killing between HIV-infected (Ul) and uninfected (clone 34) promonocytic cells.
Clone 17
CD14
I
0
m
HIV-Infected
Uninfected
IFN-rR
2603
HIV-INFECTED CELL KILLING BY TNF-ot AND IFN-y
VOL. 68, 1994
Clone 30
U1 O U937 11 cl. 7 - cl. 8 E cl. 25 MZ cl. 30 P cl. 34 - cl. 45
Ul
t
t ci
TNF-a-R
E
.-c Lysis + + Phenotype FIG. 6. The insensitivity of U937 clone 30 to the anti-proliferative effect of TNF-a plus IFN--y does not depend on the lack of the cell surface cytokine receptors. Fluorescence-activated cell sorter analyses were performed twice on Ul, U937, and a panel of U937-derived clones. Ul and clone 17 are sensitive to TNF-ac plus IFN--y, whereas U937 bulk cell line (not shown) and clone 30 are resistant. All four cell lines and clones expressed the monocytic marker CD14 (although fluctuations of the levels of expression of this marker have been observed from experiment to experiment) and comparable levels of IFN-y and TNF receptors. Dashed line, isotype Ab control. Solid line, specific fluorescence.
Cl,
a-
(D
co
~~E< EE a, E
:3z
resulted in profound synergistic effects on virus expression as well as lysis of persistently infected Ul cells. Since cytokines such as TNF-a and IFN-y which induce lysis of HIV-infected cells in vitro have been used in clinical trials (2) because of their potential ability to eliminate HIV-infected cells in vivo, it is important to determine whether such cytokines, either alone or in combination, also induce substantial virus production in the process of inducing cytolysis. Our findings are in contrast with earlier observations by Wong, who observed a suppressive effect on viral RNA synthesis in a panel of cell lines infected with HIV and exposed to TNF-oa, either alone or in combination with IFN--y (25). However, U937 cells were not used in that study because of their known susceptibility to TNF-mediated lysis. Furthermore, very high concentrations of TNF were used, suggesting the possibility that cell toxicity occurred before viral mRNAs and proteins could accumulate (25). Matsuyama first described the phenomenon of increased virus expression concomitant with cell lysis in HIV-infected cells after stimulation with TNF-a (19). However, no information was provided on the potential role of IFN--y in those culture conditions. Furthermore, it was reported in the same study that certain cell lines differed from this general pattern and, in particular, HIV-infected U937 cells were less susceptible to TNF-mediated killing than their uninfected counterparts (19). Our comparative analysis of the cytopathic effects induced by TNF-a and/or IFN-y on HIV-infected Ul cells, uninfected U937 cells, and uninfected U937 clones indicates that a high level of heterogeneity exists within the population of U937 cells. In this regard, it is important to underscore that a strong selection process occurs during the acute phase of HIV replication in both primary and tumor CD4+ target cells (such as U937 cells) where a substantial proportion of cells are killed and survivor cells emerge thereafter which contain integrated proviral DNA in their genome (11, 12, 28). The mechanisms underlying this active clonal selection process are
E)
c
CO 0
E-
C,)
., +
z U-
FIG. 7. Anti-Fas MAb kills both Ul cells (without inducing viral expression) and several uninfected cell clones in synergy with IFN-y. The proliferative capacity of HIV-infected Ul cells and uninfected U937 and related cell clones was assessed after 72 h of treatment with IFN--y alone (1,000 U/ml), anti-Fas MAb (1 ,ug/ml), or the combination of these two agents. The antiproliferative effects were confirmed in two independent experiments and by comparative analysis of cell proliferation and viability determined by Trypan blue dye exclusion criteria. Neither induction of RT activity nor enhancement of TNF-oaor IFN-y-mediated upregulation of virus expression was observed in the presence of anti-Fas MAb.
currently unknown. Therefore, functional or phenotypic changes in persistently infected cells may not necessarily reflect the influence of HIV genes, as in the case of adenovirus infection (6, 8, 14), but may be the result of the selective process imposed on target cells during the acute phase of virus replication. Experiments in which anti-Fas MAb (7, 16, 27) instead of TNF-ot was used in combination with IFN-y confirmed that both infected Ul cells and the majority of uninfected U937 cell clones were highly susceptible to cytokinemediated lysis. The presence of such clonal variability dictates that caution should be exercised in comparative analyses of infected and uninfected cell lines. In this regard, the decrease susceptibility to the lytic effect of TNF-ot observed by Matsuyama in infected U937 cells versus their uninfected counterparts (19), which is in contrast with our findings (Fig. 1), may be explained by clonal variability within the bulk U937 cells. Finally, it is very unlikely that lysis of Ul cells was the consequence of high levels of virus expression induced by
2604
TNF-a plus IFN--y since costimulation of Ul cells with TNF-at plus glucocorticoids or IFN--y plus glucocorticoids resulted in the synergistic induction of virus expression that was dissociated from substantial cytopathicity. Furthermore, treatment with IFN--y plus anti-Fas MAb resulted in death of Ul cells without synergistic induction of HIV expression. In conclusion, these data demonstrate that TNF-at and IFN--y in combination can induce substantial expression of HIV from infected cells despite the fact that they may ultimately lead to the elimination of these cells. Hence, caution should be exercised in the design and implementation of therapeutic protocols involving the administration of potentially HIV-inductive cytokines. ACKNOWLEDGMENTS
We thank D. Cohen, G. Pantaleo, and E. Vicenzi for helpful suggestions and critical reading of the manuscript; C. Muro-Cacho for studies of cytokine-induced apoptosis; M. Baseler and J. Adelsberger for performing cytofluorometric analysis; E. Hardy for technical assistance; and Janet S. Bey for typing the manuscript. Priscilla Biswas is funded by a fellowship for AIDS research from the Istituto Superiore di Sanita (I.S.S.), Rome, Italy. This project was partly supported by the I.S.S./A.I.D.S. project no. 8204-99. REFERENCES 1. Adachi, A., H. E. Gendelman, S. Koenig, T. M. Folks, R. Willey, A. Rabson, and M. A. Martin. 1986. Production of acquired immunodeficiency syndrome-associated retrovirus in human and nonhuman cells transfected with an infectious molecular clone. J. Virol. 59:284-291. 2. Agosti, J. M., R. W. Coombs, A. C. Collier, M. A. Paradise, J. K. Benedetti, H. S. Jaffe, and L. Corey. 1992. A randomized, doubleblind, phase I/II trial of tumor necrosis factor and interferongamma for treatment of AIDS-related complex (Protocol 025 from the AIDS Clinical Trials Group). AIDS Res. Hum. Retroviruses 8:581-587. 3. Biswas, P., G. Poli, A. L. Kinter, J. S. Justement, S. K. Stanley, W. J. Maury, P. Bressler, J. M. Orenstein, and A. S. Fauci. 1992. Interferon-y modulates the expression of human immunodeficiency virus in persistently infected promonocytic cells by redirecting the production of virions to intracytoplasmic vacuoles. J. Exp. Med. 176:739-750. 4. Bradford, M. M. 1976. A rapid and sensitive method for the quantitation of microgram quantities of protein utilizing the principle of protein-dye binding. Anal. Biochem. 72:248-254. 5. Bressler, P., G. Poli, J. S. Justement, P. Biswas, and A. S. Fauci. 1993. Glucocorticoids synergize with TNF-a in the induction of HIV expression from a chronically infected promonocytic cell line. AIDS Res. Hum. Retroviruses 9:547-551. 6. Chen, M. J., B. Holskin, J. Strickler, J. Gorniak, M. A. Clark, P. J. Johnson, M. Mitcho, and D. Shalloway. 1987. Induction by ElA oncogene expression of cellular susceptibility to lysis by TNF. Nature (London) 330:581-583. 7. Cohen, P. L., and R. A. Eisenberg. 1992. The lpr and gld genes in systemic autoimmunity: life and death in the Fas lane. Immunol. 8.
J. VIROL.
BISWAS ET AL.
Today 13:427-428. Duersken-Hughes, P., W. S. Wold, and L. R. Gooding. 1989.
Adenovirus ElA renders infected cells sensitive to cytolysis by tumor necrosis factor. J. Immunol. 143:4193-4200. 9. Duh, E. J., W. J. Maury, T. M. Folks, A. S. Fauci, and A. B. Rabson. 1989. Tumor necrosis factor at activates human immunodeficiency virus type 1 through induction of nuclear factor binding to the NF-KB sites in the long terminal repeat. Proc. Natl. Acad. Sci. USA 86:5974-5978. 10. Fernie, B. F., G. Poli, and A. S. Fauci. 1991. Alpha interferon suppresses virion but not soluble human immunodeficiency virus
antigen production in chronically infected T-lymphocytic cells. J. Virol. 65:3968-3971. 11. Folks, T. M., J. S. Justement, A. Kinter, S. Schnittman, J.
12.
13. 14.
15.
16.
17.
18.
19.
20.
21.
22. 23.
24. 25.
26.
27.
28.
Orenstein, G. Poli, and A. S. Fauci. 1988. Characterization of a promonocyte clone chronically infected with HIV and inducible by 13-phorbol-12-myristate acetate. J. Immunol. 140:1117-1122. Folks, T. M., D. M. Powell, M. M. Lightfoote, B. Steven, M. A. Martin, and A. S. Fauci. 1986. Induction of HTLVIII/LAV from a nonvirus-producing T-cell line: implications for latency. Science 231:600-602. Golstein, P., D. Ojcius, and D. J. Young. 1991. Cell death mechanisms and the immune system. Immunol. Rev. 121:29-65. Gooding, L. R., L. W. Elmore, A. E. Tollefson, H. A. Brady, and W. S. Wold. 1988. A 14,700 MW protein from the E3 region of adenovirus inhibits cytolysis by tumor necrosis factor. Cell 53:341346. Griflin, G. E., K. Leung, T. M. Folks, S. Kunkel, and G. J. Nabel. 1989. Activation of HIV gene expression during monocyte differentiation by induction of NF-kappa B. Nature (London) 339:7073. Kobayashi, N., Y. Hamamoto, N. Yamamoto, A. Ishii, M. Yonehara, and S. Yonehara. 1990. Anti-Fas monoclonal antibody is cytocidal to human immunodeficiency virus-infected cells without augmenting viral replication. Proc. Natl. Acad. Sci. USA 87:96209624. Koyanagi, Y., W. A. O'Brien, J. Q. Zhao, D. W. Golde, J. C. Gasson, and I. S. Y. Chen. 1988. Cytokines alter production of HIV-1 from primary mononuclear phagocytes. Science 241:16731675. Larrick, J. W., and S. C. Wright. 1990. Cytotoxic mechanism of tumor necrosis factor-a. FASEB J. 4:3215-3223. Matsuyama, T., Y. Hamamoto, G. I. Soma, D. Mizuno, N. Yamamoto, and N. Kobayashi. 1989. Cytocidal effect of tumor necrosis factor on cells chronically infected with human immunodeficiency virus (HIV): enhancement of HIV replication. J. Virol. 63:2504-2509. Osborn, L., S. Kunkel, and G. J. Nabel. 1989. Tumor necrosis factor a and interleukin 1 stimulate the human immunodeficiency virus enhancer by activation of the nuclear factor KB. Proc. Natl. Acad. Sci. USA 86:2336-2340. Poli, G., P. Bressler, A. Kinter, E. Duh, W. C. Timmer, A. Rabson, J. S. Justement, S. K. Stanley, and A. S. Fauci. 1990. Interleukin 6 induces human immunodeficiency virus expression in infected monocytic cells alone and in synergy with tumor necrosis factor alpha by transcriptional and post-transcriptional mechanisms. J. Exp. Med. 172:151-158. Poli, G., and A. S. Fauci. 1993. Cytokine modulation of HIV expression. Semin. Immunol. 5:165-173. Pomerantz, R. J., D. Trono, M. B. Feinberg, and D. Baltimore. 1990. Cells non productively infected with HIV-1 exhibit an aberrant pattern of viral RNA expression. Cell 61:1271-1276. Wong, G. H., and D. V. Goeddel. 1986. Tumour necrosis factors a and P inhibit virus replication and synergize with interferons. Nature (London) 323:819-822. Wong, G. H., J. F. Krowka, D. P. Stites, and D. V. Goeddel. 1988. In vitro anti-human immunodeficiency virus activities of tumor necrosis factor-alpha and interferon-gamma. J. Immunol. 140: 120-124. Wright, S. D., R. A. Ramos, P. S. Tobias, R. J. Ulvitch, and J. C. Mathison. 1990. CD14: a receptor for complexes of lypopolysaccharide (LPS) and LPS binding protein. Science 249:1431-1433. Yonehara, S., A. Ishii, and M. Yonehara. 1989. A cell-killing monoclonal antibody (anti-Fas) to a cell surface antigen codownregulated with the receptor of tumor necrosis factor. J. Exp. Med. 169:1747-1756. Zagury, D., J. Bernard, R. Cheynier, M. Feldman, P. S. Sarin, and R. C. Gallo. 1986. Long-term cultures of HTLV-III-infected cells: a model of cytopathology of T-cell depletion in AIDS. Science
231:850-853.
