(Received for publication, February 3, 1997, and in revised form, March 6, 1997)
From the Department of Pathology, University of
Tennessee Health Science Center, Memphis, Tennessee 38163 and the
¶ Department of Protein Chemistry, Berlex Biosciences,
Richmond, California 94804
We have characterized the functional properties
of the short form of the human interferon /
receptor chain 2 (IFNAR2), denoted IFNAR2.1. IFNAR2.1 contains a shortened cytoplasmic
domain when compared with the recently cloned full-length IFNAR2 chain
(IFNAR2.2). We show that IFN
8 and IFN
1b induce antiviral and
antiproliferative activity in mouse cell transfectants expressing the
human IFNAR1 chain of the receptor and induce the formation of
STAT1/STAT2 dimers in IFN-stimulated response element
(ISRE)-dependent gel shift assays. In contrast,
coexpression of IFNAR2.1 with IFNAR1 reduces the IFN-induced antiviral,
antiproliferative and ISRE-dependent gel shift binding
activity conferred by IFNAR1 alone. No antiviral or antiproliferative
response to IFN, nor IFN-induced ISRE-dependent gel shift
binding activity, was observed when IFNAR2.1 was expressed alone in
murine cells. Therefore, IFNAR2.1 acts as a dominant negative for these
IFN-induced activities. Our results suggest that IFNAR2.1
represents a nonfunctional version of the full-length chain
(IFNAR2.2).
Type I interferons (IFNs),1 consisting
of and
subtypes, are a multifunctional cytokine family capable
of inhibiting cell proliferation and viral replication, in addition to
modulating cellular immune functions (1). Type I IFNs compete with one another for binding to a common multisubunit receptor present on the
surface of target cells. The type I IFN receptor consists of at least
two distinct subunits IFNAR1 and IFNAR2 (2-5). The IFNAR1 chain
appears to be involved primarily in signal transduction (2, 6, 7),
while the IFNAR2 chain plays a role both in ligand binding and signal
transduction (3-5). The IFNAR2 chain was originally identified as a
50-kDa protein (IFNAR2.1) (3), having a truncated
cytoplasmic domain due to alternative mRNA splicing (8). A 100-kDa
form of the IFNAR2 chain (IFNAR2.2) has been cloned which reconstitutes
biological activity in both murine (4) and human cells (5). In most
human cells IFNAR2.1 is expressed at low levels relative to the
full-length IFNAR2.2 chain (9). These findings call into question the
biological function of IFNAR2.1.
We have examined the biological consequences of expression of the human
IFNAR1 and IFNAR2.1 chains on the sensitivity of murine cells to
several type I IFNs (IFN2, IFN
8, and IFN
1b). Murine L929 cells
are completely unresponsive to human IFN (6, 7). We previously
established that the IFNAR1 subunit acts as a species-specific signal
transduction component of the human type I IFN receptor complex, but it
does not directly bind IFN (6, 7).
In the present study, we demonstrate that IFNAR1 expression in L929
cells confers sensitivity to the antiviral, antiproliferative, and
ISRE-dependent gel shift binding activities of IFN8 and
to a lesser extent IFN
1b. However, expression of IFNAR1 in L929 cells does not confer sensitivity to the biological activities of
IFN
2. IFNAR2.1 expression in murine cells does not confer sensitivity to any IFN subtype examined. Interestingly, cells expressing both receptor chains exhibited a markedly reduced
sensitivity to the antiviral, antiproliferative, and
ISRE-dependent gel shift binding activities of IFN
8 and
IFN
1b, when compared with cells expressing only IFNAR1. These
results suggest that IFNAR2.1 acts in a dominant negative manner for
the induction of biological activity by several type I IFNs.
Isolation of mouse L929 cells expressing IFNAR1, IFNAR2.1, or both IFNAR1 and IFNAR2.1 together has been described previously (10). IFNAR1 and IFNAR2.1 transfectants were maintained in Dulbecco's modified Eagle's medium supplemented with 10% bovine calf serum and geneticin (1.5 mg/ml) or hygromycin (0.3 mg/ml), respectively. Stable transfectants expressing both receptor chains were maintained in Dulbecco's modified Eagle's medium supplemented with 10% bovine calf serum, geneticin (1.5 mg/ml), and hygromycin (0.3 mg/ml).
Recombinant IFN2 (4 × 107 IU/mg), IFN
8 (2 × 107 IU/mg), and IFN
1b (4 × 107
IU/mg) was kindly provided by P. Trotta, Schering-Plough (Kenilworth, NJ), H. Hochkoppel, Ciba-Geigy AG (Basel, Switzerland), and Berlex Biosciences (Richmond, CA), respectively. IFN activities are expressed in international reference units/ml as assayed by protection against the cytopathic effect of vesicular stomatitis virus (VSV) on human fibroblasts, using the NIH human IFN
standard for reference.
For determining antiviral activity, cell cultures (5 × 105 cells/25 cm2 flask) were preincubated overnight with IFN, followed by infection with VSV (Indiana strain) for 1.5 h at 0.1 plaque-forming units (pfu) per cell. At 24 h post-infection, the virus yield in the medium was assayed by plaque formation on Vero cells (11). Assay of the antiproliferative action of IFN was performed by treating cells (1 × 105 cells/25 cm2 flask) with IFN. After 4 days, the cells were trypsinized and enumerated in a Coulter Counter (12, 13).
Nuclear Extracts and Gel Shift AssaysNuclei were extracted
from control and IFN-treated cells (3000 IU/ml, 15 min) with buffer
containing 20 mM Tris-HCl, pH 7.85, 250 mM
sucrose, 0.4 M KCl, 1.1 mM MgCl2, 5 mM -mercaptoethanol, and 0.4 mM
phenylmethylsulfonyl fluoride, and extracts were frozen on dry ice and
stored at
80 °C (11, 14, 15). For gel shift analysis, the nuclear
extracts were incubated with a 32P-end-labeled ISG15 ISRE
promoter probe (5
-GATCCATGCCTCGGGAAGGGAAACCGAAACTGAAGCC-3
) (16, 17)
at 25 °C for 30 min, and the unbound radiolabeled probe was
separated from protein·DNA complexes on 5% polyacrylamide gels. Gels
were quantitated by phosphorimage autoradiography (Molecular Dynamics).
We
have previously described the isolation of murine L929 cells stably
expressing the cloned human IFNAR1 chain (L929R1 cells), the 51-kDa IFNAR2.1 chain (L929R2 cells), or both IFNAR1
and IFNAR2.1 (L929R12 cells) (10). IFNs are defined by
their ability to inhibit the replication of a wide variety of RNA and
DNA viruses. Mouse L929 cells are resistant to the antiviral action of
huIFN, but are highly responsive to mouse IFN (6). Expression of the
human IFNAR1 (huIFNAR1) chain in L929 cells has been found to
result in antiviral protection by some human type I IFNs, without
affecting sensitivity to mouse IFN/
(6, 7). In the present study, we determined the effect of huIFNAR2.1 expression on IFN
s antiviral activity in cells expressing IFNAR1. Cells were treated overnight with
human IFN, then infected with VSV (at a multiplicity of infection of
0.1 pfu/cell) and the virus released into the culture medium at one day
post-infection was titered by plaque assays on the indicator Vero cell
line.
Expression of huIFNAR1 or huIFNAR2.1 had no effect on the ability of
L929 cells to support VSV replication, since virus replication was
similar in all cells lines examined (L929, L929R1,
L929R2, or L929R12) varying between 0.8 and
1.3 × 108 pfu/ml. As shown in Fig. 1
even at a concentration as high as 3,000 IU/ml, IFN2, IFN
8, or
IFN
1b did not markedly inhibit VSV replication in L929 cells or in
L929R2 cells. Consistent with our previous studies (6),
L929R1 cells, which only express huIFNAR1, were sensitive
to the antiviral action of human IFN
8 and to a lesser extent
IFN
1b. In contrast, coexpression of huIFNAR2.1 with huIFNAR1 in
L929R12 cells markedly reduced sensitivity to the antiviral
action of human IFN
8 or IFN
1b. For example, IFN
8 treatment
(3000 IU/ml) induced only a 225-fold reduction in viral titer in
L929R12 cells, as compared with a >26,000-fold reduction in L929R1 cells (>100-fold reduction in antiviral
activity). A dose-dependent reduction of viral titer in
L929R1 and L929R12 cells was observed with
IFN
8 concentrations varying from 100 to 10,000 IU/ml (Fig.
1B). However, at all concentrations the antiviral effect of
IFN
8 was markedly reduced in cells coexpressing IFNAR2.1. These
results show that the expression of IFNAR2.1 reduces the sensitivity of
IFNAR1-expressing cells to the antiviral action of human IFN
8 and
IFN
1b.
Antiproliferative Activity of Human IFNs in Murine Transfectants
Besides their broad antiviral action, IFNs also
inhibit cell proliferation. Since L929R1 transfectants were
highly sensitive to the antiviral action of human IFN, we extended our
studies to determine if they were also sensitive to IFN's
antiproliferative action. Although cells sensitive to the antiviral
action of IFN are usually also sensitive to the antiproliferative
action, there are instances where there is a divergence in sensitivity
to these actions of IFNs (18). Thus, the various L929 transfectants
were treated with human type I IFNs and cell numbers quantitated after 4 days. As shown in Fig. 2, treatment with human IFN8
or IFN
1b inhibited the proliferation of L929R1 cells
which express IFNAR1 alone, or of L929R12 cells, which
coexpress IFNAR1 and IFNAR2.1. However, the antiproliferative effect in
L929R12 cells was markedly less when compared with the
effect with L929R1 cells. IFN
2, IFN
8, or IFN
1b did
not affect the proliferation of L929 cells or L929R2 cells
that express the IFNAR2.1 chain. IFN
2 had no significant antiproliferative activity in any of the transfectant
(L929R1, L929R2, or L929R12) cell
lines. Although an antiproliferative effect could be detected in
L929R1 cells at 100 IU/ml IFN
8 (Fig. 2B) or
IFN
1b (data not shown), the maximal antiproliferative effect was
observed at 10,000 IU/ml, and no further inhibition of proliferation
was observed at higher IFN concentrations. Most importantly, IFN's
antiproliferative effect in L929R12 cells was reduced at
all IFN concentrations when compared with the effect in
L929R1 cells. For example, the IC50 of the
antiproliferative effect of IFN
8 was 125 IU/ml in L929R1
transfectants as compared with an IC50 of >30,000 IU/ml
for IFN
8 in L929R12 cells. Thus, expression of
huIFNAR2.1 in murine cells reduced the sensitivity of cells expressing
huIFNAR1 to the antiviral and antiproliferative effects of IFN
8 and
IFN
1b.
Induction of ISGF3 Activity by Human IFNs in Murine Transfectants
An early event induced by the binding of IFNs to
their receptor is specific gene induction through the activation of the
STAT-containing ISGF3, which binds to the conserved ISRE present in the
promoter of many IFN-responsive genes. Because of the importance of
ISGF3 in IFN signaling, we determined the IFN-induced formation of the ISGF3 transcription factor in the various murine transfectants. Nuclear
extracts prepared from control and IFN-treated (3,000 IU/ml, 15 min)
transfectants were incubated with a labeled ISRE probe, and the
resultant protein·DNA complexes were analyzed by the gel shift assay.
Fig. 3 shows that treatment of L929R1 cells with human IFN8 or IFN
1b induced ISGF3 activity. IFN
2 did not induce ISGF3 activity in L929R1 cells. In contrast, the
level of ISGF3 activity induced by IFN
8 and IFN
1b was markedly
reduced in L929R12 cells coexpressing IFNAR1 and IFNAR2.1.
No ISGF3 activity was induced in L929R2 cells by any type I
IFN tested. Quantitative analysis of the autoradiograms showed that
there was a
75% reduction in the formation of ISGF3 in
L929R12 transfectants exposed to either IFN
8 or IFN
1b
when compared with ISGF3 formation in L929R1 transfectants
(Fig. 3). To confirm the presence of specific STAT proteins in the
IFN-inducible protein·DNA complexes, we performed gel supershift
assays with STAT-specific antisera and showed that the
ISRE-dependent gel shift binding activity contains only
STAT1 and STAT2 (Fig. 3B). These results show that IFNAR2.1
coexpression reduces ISRE-dependent gel shift binding
activity in IFNAR1-expressing transfectants induced by human IFN
8 or
IFN
1b.
General Discussion
The data presented here indicate that
IFNAR1 is able to transduce signals not only for the antiviral action
of some type I IFNs as shown previously (6, 7), but also for the
antiproliferative action of human IFN. This presumably occurs through
the interaction of the human IFNAR1 chain with components of the mouse
type I IFN receptor (19), since IFNAR1 expression does not increase human IFN binding to murine cells. It is unclear why particular human
type I IFNs such as IFN2 are unable to transduce signals through the
available murine IFN receptor subunits, while IFN
8, and to a lesser
extent IFN
1b, are able to elicit biological effects through the
endogenous receptor. These results suggest that type I IFN subtypes may
differentially interact with their common multisubunit receptor. For
example, while IFN
8 productively interacts with human type I IFN
receptor components in a murine context, IFN
2 does not.
In addition coexpression of IFNAR1 and IFNAR2.1 in L929 cells markedly
reduced sensitivity to the biological activities of human IFN8 and
IFN
1b relative to cells expressing only IFNAR1, although the
sensitivity to murine IFN
/
was similar in all transfectants (data
not shown). Therefore, IFNAR2.1 acts in a dominant negative manner for
the antiviral and antiproliferative actions of human IFN
8 and
IFN
1b, as well as for the induction of ISRE-dependent gel shift binding activity. The level of human IFNAR2.1 chain expression relative to that of the endogenous murine IFNAR2 chain in
the various transfectants is unknown. Presumably murine cells do not
express a short form of IFNAR2, because human IFNAR2.1 is encoded by an
Alu cassette, which is a sequence present solely in primates (8).
Therefore, we interpret our results to mean that human IFNAR2.1
competes with the endogenous mouse full-length IFNAR2 for interaction
with type I IFNs to bind ligand and transduce biological effects.
We thank P. Trotta, Schering-Plough
(Kenilworth, NJ) and H. Hochkoppel, Ciba-Geigy AG (Basel, Switzerland)
for providing recombinant IFN2 and IFN
8, respectively.