were grown in RPMI 1640 medium (GIBCO)/5% (vol/vol) heat-inactivated fetal calf serum (Sigma)/2 mM L-glutamine/ penicillin-streptomycin at 20 units per ml ...
Proc. Natl. Acad. Sci. USA Vol. 90, pp. 2335-2339, March 1993 Cell Biology
Soluble tumor necrosis factor receptor: Inhibition of human immunodeficiency virus activation (AIDS) 0. M. ZACK HOWARD*, KATHLEEN A. CLOUSEt, CRAIG SMITHt, R. G. GOODWINt, AND WILLIAM L. FARRAR§¶ *Biological Carcinogenesis and Development Program, Program Resources, Inc./DynCorp, and §Laboratory of Molecular Immunoregulation, Biological Response Modifiers Program, National Cancer Institute-Frederick Cancer Research and Development Center, P.O. Box B, Frederick, MD 21702; tDivision of Cytokine Biology, Center for Biologics Evaluation and Research, Food and Drug Administration, Bethesda, MD 20892; and tlmmunex Research and Development Corp., 51 University Street, Seattle, WA 98101
Communicated by Lewis H. Sarett, December 18, 1992
The inflammatory cytokine tumor necrosis ABSTRACT factor a (TNF-a) has been shown to stimulate human immunodeficiency virus type 1 (HIV-1) replication in both chronically and acutely infected T lymphocytes and monocytes. Transcriptional activation of the HIV long terminal repeat and subsequent increase in virus production are linked to TNF activation of the cellular transcription factor NF-KB. Here we report the use of two forms of soluble recombinant type 1 (p8O) TNF receptor to inhibit TNF-induced HIV activation in vitro. One receptor form is a monomer containing the entire 236 residues of the extracellular (ligand-binding) portion of p80. A second receptor form is a chimeric homodimer containing these residues fused to a truncated human IgGl immunoglobulin heavy chain and, thus, resembles a bivalent antibody without light chains. These recombinant receptor proteins were tested for their ability to inhibit TNF-a-induced expression of HIV-1 in chronically infected human celi lines. We also examined the ability of the soluble receptors to limit the activation of the HIV-long terminal repeat transcription. The soluble TNF receptor dimer was most effective at blocking TNF-a-induced HIV-1 expression in both monocytoid and lymphoid celis. The molar ratio of TNF-receptor dimer to TNF-a found to be most effective was, at least, 5:1. We conclude that at specific TNF/soluble TNF-receptor dimer ratios, TNF-a-induced HIV-1 transcription and expression can be limited in vitro.
to endotoxin, IL-1, IL-6, and TNF itself (11). TNF biosynthesis is tightly regulated by both transcriptional and posttranscriptional mechanisms (11-14). TNF gene transcription is not only regulated by NF-KB binding but can induce NF-KB in both T cells and monocytes, thus providing one signal for HIV activation in virus-infected cells. The biological response of TNF is mediated through cellsurface receptors, two forms of which have been molecularly cloned (15-19). These two forms share striking sequence homology in their extracellular domains and bind both TNF-a and TNF-,f with similar affinities. Two recombinant soluble forms of the type I TNF receptor (TNF-R) have been recently described (19, 20). One form, a monomer (mTNF-R1), contains all 236 extracellular residues and binds TNF-a with a Kd of 5 nM. A second form is a chimeric homodimer composed of the 236-amino acid extracellular residues fused to a human IgG Fc region. This soluble receptor (FcTNF-R1) binds TNF-a and TNF-jB with, at least, 50-fold greater affinity than the monomeric form. Naturally occurring soluble forms of type I and type II TNF-Rs have been found in human urine and sera and have been shown to be proteolytic cleavage products of the extracellular portions (20). These naturally occurring soluble receptors appear to bind TNF-a with affinities similar to the recombinant p80 monomer. In normal individuals, circulating serum levels of soluble TNF-Rs are -1-2 ng/ml (21-23). These levels are elevated in individuals in response to infection and some types of cancer (24). Engelmann et al. (25) showed that the soluble receptor can block the toxicity of TNF. A more recent study done in mice demonstrated that a recombinant soluble TNF-R could protect mice from the lethal effects of endotoxemia (26, 27). Various studies of TNF-a levels in HIV patients have shown an increase in TNF-a levels during the early stages of the viral infection (22). A study of the acid-labile interferon a found in AIDS patients shows that this form of interferon can induce TNF-R (23). Our own studies are focused on the ability of recombinant, soluble p80 TNF-Rs to block the TNFmediated induction of HIV expression in vitro.
Human immunodeficiency virus (HIV) expression is concurrent with both lymphocytic and monocytic activation, both of which require the participation of cytokines (1-4). For these reasons, extensive research has been conducted investigating the role of cytokines in the amplification or inhibition of HIV expression in lymphocytic and monocytic cell lines. Among the many cytokines tested, tumor necrosis factor (TNF) appears to be the principal cytokine that can amplify HIV expression in human lymphocyte and monocyte cell lines (3-5). TNF-a, classically characterized as the product of activated macrophages, is a highly pleiotropic cytokine that plays a central role in mediating the inflammatory response to injury and infection (6-8). Activities of TNF-a include cytotoxic and cytostatic effects against tumors and virusinfected cells, stimulation of interleukin (IL) 1 secretion, and activation of various immune effector cells. The biologically active form of TNF-a is a homotrimer (9, 10). Chronic TNF production has been implicated in cachexia; acute TNF production appears to be involved in the pathogenesis of septic shock and endotoxin-induced tissue injury. TNF is produced by phagocytic and nonphagocytic cells in response
MATERIALS AND METHODS Reagents. Recombinant human TNF-a was supplied by the Biological Response Modifiers Program repository (Genentech). Soluble recombinant chimeric FcTNF-Rl and soluble Abbreviations: HIV-1, human immunodeficiency virus type 1; TNF, tumor necrosis factor; TNF-R, TNF receptor; mTNF-R1, monomeric type 1 TNF-R containing entire 236 extracellular residues; FcTNF-R1, homodimer of 236 extracellular residues of type 1 TNF-R fused to IgG Fc region; RT, reverse transcriptase; CAT, chloramphenicol acetyltransferase; LTR, long terminal repeat; IL, interleukin. $To whom reprint requests should be addressed.
The publication costs of this article were defrayed in part by page charge payment. This article must therefore be hereby marked "advertisement" in accordance with 18 U.S.C. §1734 solely to indicate this fact.
2335
2336
Cell
Biology: Howard et al.
recombinant mTNF-R1 were provided by Immunex (Seattle). Cells. U38 [AIDS Research and Reference Reagent Program (ARRRP) no. 1297] and A3.01 (ARRRP no. 166) cells were grown in RPMI 1640 medium (GIBCO)/5% (vol/vol) heat-inactivated fetal calf serum (Sigma)/2 mM L-glutamine/ penicillin-streptomycin at 20 units per ml (GIBCO). The chronically HIV-infected cell lines ACH-2 (ARRRP no. 349) and U1 (ARRRP no. 165) were cultured in RPMI 1640 medium/10 mM Hepes buffer (Biofluids, Rockville, MD), pH 7.0/2 mM L-glutamine/penicillin-streptomycin at 100 units/ ml/10% (vol/vol) heat-inactivated fetal calf serum (complete medium). Soluble Receptor. Two forms of a soluble human p80 TNF-R were expressed. The monomeric form contains the entire 236 residues of the extracellular region; the dimeric form consists of a chimeric homodimer in which these same 236 residues were fused to a truncated human IgGl heavy chain. These constructs were inserted in an expression vector and transfected into DXB-11 CHO cells (dihydrofolate reductase deficient). Cells were selected for expression of dihydrofolate reductase by subculturing in selective medium lacking glycine, hypoxanthine, and thymidine. The highest expressing cultures were amplified by exposure to increased concentrations of methotrexate. mTNF-R1 was purified from supernatants of the highest-expressing clone by TNF affinity chromatography, whereas FcTNF-R1 was purified by using protein A chromatography. Purified receptor concentration was determined by amino acid composition and stored at 4°C (sterile) for up to 2 mo. Details of the constructions and binding characteristics of these recombinant proteins will be described in detail elsewhere. Induction of HIV Expression. ACH-2 or Ul cells (0.5 x 106 cells per ml) were cultured in complete medium alone or in complete medium containing various concentrations of recombinant human TNF-a and/or recombinant human soluble TNF-Rs in 24-well tissue-culture plates in a total vol of 1 ml (28). Cells were kept at 37°C in 5% C02; then aliquots of culture supernatants were removed at 24, 48, and/or 72 hr, frozen at -80°C, and subsequently analyzed for reverse transcriptase (RT) activity. RT Assay. The RT assay used for these experiments is a modification of a method described by Hoffman (29). Briefly, 60 1.l of harvested supematants was diluted with 60 ,ul of Tris buffer (pH 7.8)/0.05% Triton X-100. Replicate 50-,ul samples were added to 100 ,ul of a solution containing poly(rA) (Pharmacia LKB), oligo(dT) (Pharmacia LKB), MgCl2, and 3H-labeled dTTP (NEN) in a 96-well U-bottomed microtiter plate (Falcon 3910) and incubated 2 hr at 37°C. After incubation, 100 ul of a solution containing 10% trichloroacetic acid and 10% sodium pyrophosphate were added to each well. The individual wells were transferred to glass-fiber filters (Wallac) by using a cell harvester (Pharmacia LKB) connected to two fluid reservoirs containing 5% trichloroacetic acid/5% sodium pyrophosphate and 70% ethanol, which ran in sequence. The filters were dried and counted on a /3 scintillation counter (Beta Plated Reader; Pharmacia LKB). Values shown reflect the average of duplicate samples (cpm per 25 ,l4) that differ by not more than 15%. Chloramphenicol Acetyltransferase (CAT) Assay. U38 cells were washed three times with AIM-V medium (GIBCO) and plated in 24-well tissue-culture plates at 0.5-1.0 x 106 cells per ml. These cells were treated with combinations of TNF-a and soluble TNF-R for 24 hr, after which the cells were harvested, and a solution CAT assay was done. The A3.01 cells were electroporated with HIV-long terminal repeat (LTR)-CAT plasmid (30) in a Bio-Rad gene pulser at 960-,uF and 250-V settings. The electroporated cells were cultured overnight in RPMI 1640 medium/10% fetal calf serum. The next day the cells were centrifuged over a Ficoll gradient to
Proc. Natl. Acad. Sci. USA 90 (1993)
separate the living cells. These cells were pooled, washed, and then plated in 24-well tissue-culture plates at 0.3-1.0 x 106 cells per ml. The transfected cells were cultured with various combinations of TNF-a and soluble receptor for 24 hr, after which the cells were harvested, and a solution CAT assay was done. The phase-extraction CAT assay was performed as recommended in ref. 31. Cell extracts were prepared by multiple cycles of freeze-thaw in 200 ,ul of Tris buffer, pH 7.8. The cell extracts were heat-inactivated for 7 min at 65°C. The assay reactions contained 0.2 ,uCi of [3H]chloramphenicol (NEN), 25 ,ug of butyryl-CoA (Sigma), 100 mM Tris (pH 8.0), and between 18 ,ug and 0.4 ,g of cell extraction in a total volume of 100 ,ul. The assay was incubated for 4 hr followed by extraction of the acylated chloramphenicol with 200 ul of tetramethylpentadecane/ xylene, 2:1 (vol/vol). The recovered 150-pl sample was overlaid with biodegradable counting scintillation (phase combining system) mixture (Amersham), and radioactivity was counted on a liquid scintillation counter (Pharmacia LKB 1410).
RESULTS TNF Induction of ACH-2 and Ul Cells. To demonstrate the inducibility of HIV-1 expression by TNF-a and to determine the optimum tissue-culture conditions, recombinant human TNF-a was titrated on the chronically HIV-1-infected T-lymphocytic (ACH-2) and promonocytic (Ul) cell lines. Complete tissue culture medium was placed in 24-well tissueculture plates with various concentrations of TNF-a. To these prepared plates, 0.5 x 106 Ul or ACH-2 cells were added in a final volume of 1 ml. Aliquots of cell supematants harvested 72 hr after stimulation were assayed for HIV-1 expression by RT activity. A representative 72-hr titration curve is shown in Fig. 1. The maximum stimulation of HIV-1 expression in ACH-2 cells occurred with TNF-a at 1 ng/ml (19 pM of bioactive TNF-a trimer). At this TNF-a concentration in ACH-2 cells, there is a 20-fold increase in HIV-1 expression above that seen in medium control cultures. The maximum TNF-a-mediated HIV-1 expression in Ul cells occurs with TNF-a at 5 ng/ml (96 pM bioactive TNF-a trimer). At this TNF-a concentration there is an 18-fold increase in HIV-1 expression above that seen in the medium control (Fig. 1).
0
ACH-2 Cells Ul Cells
C#)
x
E
CL 0
I-.
TNF Concentration
(ng/ml)
FIG. 1. Titration of recombinant TNF-a on ACH-2 and Ul cells. ACH-2 and Ul cells were cultured separately (0.5 x 106 cells per ml) in complete medium with (U, *) and without (o, o) various concentrations of recombinant TNF-a. Assuming all TNF-a is in the bioactive trimer form, the TNF concentration is 10, 19, 48, and 96 pM, respectively. Aliquots of culture supernatants were removed after 72 hr of incubation, stored at -80°C, and then assayed for RT activity. Data are expressed as counts per min per 25 ,ul of supernatant.
Cell
Biology: Howard et al.
Proc. Natl. Acad. Sci. USA 90 (1993)
Effect of Soluble FcTNF-Rl on ACH-2 and Ut HIV-1 Induction. The ability of the soluble FcTNF-R1 to inhibit TNF-a-induced HIV-1 virus expression was determined. Complete tissue culture medium was placed in tissue-culture plates with specific concentrations of TNF-a, in the presence or absence of various concentrations of soluble FcTNF-R1. To these prepared plates, 0.5 x 106 Ul or ACH-2 cells were added. After 24, 48, or 72 hr the cell supernatants were harvested, and an RT was assayed. When ACH-2 cells, the T-lymphocyte model, were treated with TNF-a at 0.5 ng/ml (10 pM TNF-a trimer) two concentrations of FcTNF-R1, 5 ng/ml (49 pM) and 10 ng/ml (97 pM), could limit induction of HIV-1 (Fig. 2). This result represents a 5- or 10-fold molar excess of FcTNF-R1 to TNF-a. Ul cells were used as a monocytic cell model to determine the effectiveness of soluble FcTNF-R1 in modulating the TNF-a-induced HIV-1 expression. When Ul cells were treated with TNF-a at 2.5 ng/ml (48 pM TNF-a trimer), 10and 20-fold weight excesses of soluble FcTNF-R1 [25 ng/ml (240 pM) or 50 ng/ml (480 pM)] inhibited TNF-a-induced HIV-1 expression, in a manner comparable to that seen for ACH-2 cells (Fig. 3). This result represents a 5- or 10-fold molar excess of FcTNF-R1 to TNF-a. Down-regulation of TNF-a-mediated HIV expression by soluble FcTNF-R1 in either ACH-2 or Ul cell systems could not be attributed to toxic effects on the cells (data not shown). At least a 5- to 10-fold molar excess of soluble FcTNF-R1 appeared necessary for reduction of TNF-a-mediated HIV-1 expression. Lower concentrations either failed to reduce HIV-1 expression or appeared to augment virus production, as reflected by RT activity. To determine which form of the soluble TNF-R is most effective at blocking TNF-a-induced HIV-1 expression, we compared the blocking abilities of FcTNF-Rl (dimer) or mTNF-R1 (monomer) in the ACH-2 and Ul cell systems. Fig. 4 shows that FcTNF-R1 could inhibit TNF-a-induced HIV-1 expression in a dose-dependent manner for both ACH-2 and Ul cells. In contrast, no inhibition ofTNF-a-mediated HIV-1 expression was seen when mTNF-Rl was tested at equivalent wt/vol concentrations to the dimer (Fig. 4). Furthermore, 10-fold excess (wt/vol) concentrations of soluble TNF-R, which efficiently reduce HIV-1 expression when the dimer is used, actually augment virus expression when the monomeric form is used (Fig. 5). This result represents a FcTNFRI None
FcTNFRI None
co 0
x
E 25 ng/ml
0.
50 ng/ml
I-
Medium control
24
x
E
/>~~~~~~~5ng/ml
> /
0 I-
48 Hours post induction
72
FIG. 3. Effect of soluble FcTNF-R1 on TNF-a-induced expression of HIV-1 in Ul cells. Ul cells were cultured, as indicated, in medium (o) or recombinant TNF-a (2.5 ng/ml, 48 pM) alone (e) or with FcTNF-R1 [25 ng/ml (240 pM) or 50 ng/ml (480 pM)] (U). Aliquots of culture supernatants were removed after 24, 48, and 72 hr of incubation, stored at -80°C, and then assayed for RT activity. Data presented reflect levels of RT activity in duplicate samples that differed by not more than 15%.
20-fold molar excess of mTNF-R1 to TNF-a, whereas that of the FcTNF-Rl to TNF-a was only a 5-fold molar excess. HIV-LTR CAT Activity. To determine the stimulation of HIV-LTR transcription by TNF-a, CAT activity was measured in U38 and transfected A3.01 cells. This stimulation followed the same curve as seen in Fig. 1. U38 cells, the monocytic model, were derived from U937 by stable integration of copies of HIV-1 LTR promoter linked to a CAT reporter gene (32). U38 cells were placed into tissue-culture plates in the presence of TNF-a at 2.5 ng/ml (48 pM TNF-a trimer) followed immediately by addition of various
.o ._
.D
90. 80. 70. 60.
ACH-2 cells + FcTNFRI
Ul cells + FcTNFRI
50. 40. 30 20.
10. I
