Abstract
TNF-related apoptosis-inducing ligand (TRAIL) is able to kill many transformed cells of diverse tissue types. We show that TRAIL is inducible by IFN-γ, by TNF-α, and by infection with human CMV, and has potent antiviral activity in vitro. CMV infection and IFN-γ also reciprocally modulate TRAIL receptor (TRAIL-R) expression. CMV infection increased the expression of TRAIL-R1 and -R2, whereas IFN-γ down-regulated the expression of TRAIL-Rs on uninfected fibroblasts. Moreover, IFN-γ significantly decreased the basal level of NF-κB activation, a known survival factor that inhibits apoptosis. Thus, TRAIL selectively kills virus-infected cells while leaving uninfected cells intact, and IFN-γ potentiates these effects by dynamic modulation of TRAIL and TRAIL-R expression and by sensitizing cells to apoptosis. The regulation of TRAIL and TRAIL-R expression may represent a general mechanism that contributes to the control of TRAIL-mediated apoptosis.
Interferon-γ is produced by T and B lymphocytes, NK cells, and monocyte/macrophages, and can influence a vast number of cellular responses (1). However, IFN-γ is perhaps best known for its direct antiviral activities, which are mediated primarily through induction of the dsRNA-activated protein kinase, 2′-5′ oligoadenylate synthase, and RNaseL (2). A role for IFN-γ in apoptotic cell death and immune suppression has also become apparent (3, 4). In fact, IFN-γ up-regulates the expression of a number of apoptosis-associated proteins in vitro, including Fas (5), and induces apoptosis of mouse monocytes (6). IFN-γ mediates its affects by binding to a specific cellular receptor that is ubiquitously expressed on most cells, especially nonlymphoid cells (7).
TNF-related apoptosis-inducing ligand (TRAIL)2 is capable of inducing apoptosis of a wide range of human tumor cells, but generally not normal cells (8). TRAIL induces apoptosis by binding and cross-linking death-domain containing receptors, TRAIL-R1 (also known as DR4 (9)) and TRAIL-R2 (10) (also called DR5 or TRAIL receptor inducer of cell killing 2 (TRICK2) (11, 12, 13, 14, 15)). Apoptotic signaling occurs via recruitment of adapter proteins, which results in the activation of caspases (reviewed in Refs. 16 and 17). Two non-death domain-containing TRAIL receptors (TRAIL-Rs) also exist: TRAIL-R3 (18) (or decoy receptor-1 DcR1/TRAIL receptor without an intracellular domain (TRID) (11, 13, 14, 15)) and TRAIL-R4 (19) (or DcR2/TRAIL receptor with a truncated death domain (TRUNDD) (20, 21)). TRAIL-R3 is an extracellular GPI-linked protein; TRAIL-R4 is a transmembrane receptor that is similar to TRAIL-R1 and -R2 but lacks an intact death-domain and therefore cannot signal cell death. It has been proposed that TRAIL-R3 and -R4 act predominately as decoy receptors to inhibit the cytotoxic effects of TRAIL (11, 15, 17, 18, 19). However, their physiological role is still uncertain and all three transmembrane receptors (TRAIL-R1, TRAIL-R2, and TRAIL-R4) can transduce signals that result in the activation of NF-κB proteins in vitro (19, 22).
The events that are required for TRAIL-resistant cells to become susceptible to TRAIL-mediated apoptosis are not well understood. Many viruses have a significant impact on host cell metabolism; hence, it might be predicted that cells infected with viruses might acquire sensitivity to TRAIL. That apoptosis is an important innate antiviral response of the host organism is attested to by the fact that many viruses have incorporated open reading frames encoding potent regulators of cell death (reviewed in Refs. 23, 24, 25). In fact, viruses with targeted disruptions or naturally occurring mutations in these genes often exhibit replication defects in vitro as well as reduced virulence in vivo (26, 27, 28). Therefore, apoptosis of virus-infected cells and viral resistance to apoptosis-inducing ligands are important factors that can determine the outcome of virus infection in vivo.
With these facts in mind, we studied the effects of TRAIL on virus infection. Given that TRAIL does not induce apoptosis of most normal cells in vitro (8) or when administered in vivo as a potently active trimeric protein (29), we used primary human fibroblasts that are not normally susceptible to TRAIL but are capable of supporting productive virus infection. In this setting, we have demonstrated the efficient killing of virally infected fibroblasts by TRAIL. We found that human CMV (HCMV)-infection directly up-regulated the expression of TRAIL and TRAIL-Rs on virus-infected fibroblasts; IFN-γ potentiated these effects by inducing TRAIL expression and down-regulating the expression of TRAIL-Rs on uninfected fibroblasts. Collectively, these data describe a novel mechanism by which IFN-γ mediates antiviral activity through the regulation of TRAIL and TRAIL-R expression.
Materials and Methods
Cells
Primary human foreskin fibroblast (HFF) cells were cultured in DMEM with 10% FBS, l-glutamine, and antibiotics at 37°C in 10% CO2. Primary normal human bronchial airway epithelial (NHBE) cells (Clonetics, San Diego, CA) were grown in bronchial epithelial growth medium (Clonetics) at 37°C in 5% CO2. CV-1 cells and HEp-2 cells (American Type Culture Collection, Manassas, VA) were grown in DMEM supplemented with 5% FBS as described above.
Virus
HCMV strain AD169 was a gift of Thomas Jones (Wyeth-Ayerst Research, Pearl River, NY). HCMV stocks were titrated on HFF cells, vaccinia virus (VV)-WR strain was titrated on CV-1 cells, and respiratory syncytial virus (RSV) (Long strain) was titrated on HEp-2 cells, by standard techniques.
Cytokines and Abs
FACS
Cells were removed from tissue culture dishes by gently pipetting in 2× SSC. Nonspecific binding was blocked by incubating HFF cells in FACS blocking buffer for 30 min at 4°C. Cells were incubated with anti-TRAIL M181 or anti-TRAIL-R-specific Abs (2 μg/ml), washed, and incubated with biotinylated goat anti-human Ig (1 μg/ml) and finally with streptavidin (SA)-PE (PharMingen). Two-color flow cytometry was performed by incubating cells with PE-mouse anti-human TNFR-p55 or PE-anti-human TNFR-p75 (1 μg/ml) in the final step reagent. Cells were fixed in 1% paraformaldehyde/PBS. Data were collected (10,000 or 30,000 events) using a Becton Dickinson FACScan or FACScalibur and analyzed using CellQuest software (Becton Dickinson, San Jose, CA).
Confocal microscopy
HFF cells were seeded into 8-chamber coverslips, infected with HCMV at various multiplicities of infection (moi), and subsequently cultured in media with or without IFN-γ and TRAIL. Cultures were incubated for 24 or 48 h at 37°C, at which time 10× binding buffer (PharMingen) and biotinylated annexin-V were added directly to the culture media. Cultures were incubated for 20 min at 37°C, washed, and incubated in avidin-Texas Red (PharMingen) for an additional 20 min at room temperature. Cells were fixed in 1% paraformaldehyde/PBS and incubated with a 1/5 dilution of FITC-HCMV (Dako) in 0.1% saponin for 45 min. Cells were washed again in fresh media containing 0.1% saponin, covered with glycerol/PBS containing 10% DABCO (Sigma), and examined by confocal laser scanning microscopy. The percentage of annexin-V+ FITC+ cells was scored from randomly selected low-power fields, representative of the whole culture.
RT-PCR
HFF cells were seeded into 10-cm dishes and treated with cytokines or were infected with HCMV at an moi of 2. At 24 or 48 h posttreatment and/or virus infection, HFF cells were washed in PBS and harvested directly into RNA lysis buffer. Total cellular RNA was prepared using RNeasy minicolumns (Qiagen, Valencia, CA). RNA was quantitated by absorbance at 260 nm, and 2 μg was used as the template for first strand random-primed cDNA synthesis using the preamplification system (Life Technologies, Grand Island, NY). PCR amplification of cDNAs specific to TRAIL, TRAIL-R1, TRAIL-R2, TRAIL-R3, and TRAIL-R4 as well as to TNFR-p55, TNFR-p75, dihydrofolate reductase (DHFR), and HCMV open reading frames UL-18 and UL-67 was performed using specific primers. Plasmids containing the specific target cDNAs were used as PCR controls to indicate amplification of the specific cDNA. To ensure that detection was in the linear range, PCRs were performed at various cycles and repeated at least twice.
Plasmids and transfections
An NF-κB-responsive luciferase reporter plasmid, IL8p-LUC, was constructed by ligating the IL-8 promoter sequence into the ASP718-XhoI site of pGL-Basic (Promega, Madison, WI). An IFN-γ-responsive luciferase reporter plasmid pSTAT-LUC was similarly generated by cloning the FcγRI promoter into the ASP718-XhoI sites immediately upstream of the minimal thymidine kinase promoter (−35 to +10 of the HSV thymidine kinase promoter). Therefore, this promoter contains four copies of a functional γ activation site (CGTATTTCCCAGAAAAGGAA) spaced 20 bp apart. HFF cells (105/well) were seeded into 24-well tissue culture plates, and triplicate cultures were transfected with reporter plasmids by DEAE-dextran. HFF cells were allowed to recover for 12 h before infection (or mock infection) with 100 μl of HCMV at an moi of 2 for 1 h at 37°C. Cells were cultured in DMEM in the absence or presence of various cytokines (IFN-γ, CD40L, TRAIL, IFN-γ and TRAIL, or CD40L and TRAIL). Lysates were prepared at either 24 or 48 h postinfection (p.i.) and assayed independently for luciferase activity using a Lumat LB9507 luminometer (EG&G Berthold, Wildbad, Germany).
Lactate dehydrogenase (LDH) assays
HFF cells (104/well) were seeded in 96-well flat-bottom plates, infected with HCMV at an moi of 2 for 1 h, and incubated in the presence or absence of cytokines for 24 (data not shown) or 48 h. Aliquots of supernatants (100 μl) were then assayed for LDH release using the Cytotoxicity Detection kit (Boehringer Mannheim, Indianapolis, IN). Minimum lysis was determined from cells cultured in medium alone; maximum lysis was induced by adding 100 μl of media containing 0.1% Triton X-100. Percent specific lysis was calculated using the following formula: % specific LDH release = ([experimental release − minimum release]/[maximal release − minimum release]) × 100.
Results
TRAIL synergizes with IFN-γ to induce death of virus-infected cells
TRAIL was first described as a TNF-related ligand that is capable of inducing apoptosis of transformed cells of diverse origins (8). To test whether TRAIL can also induce apoptosis of virus-infected cells, we infected HFF cells with HCMV and cultured these cells in the presence of TRAIL or TRAIL plus IFN-γ. Cultures were examined by phase contrast light microscopy at 12, 24, 48, and 72 h p.i. A rounded cytopathic effect was detectable in infected cultures in the absence of cytokines by 24 h p.i. but was more prominent by 48 h p.i. (data not shown). In contrast, extensive cell death was apparent in HCMV-infected HFF cultures treated with TRAIL or TRAIL plus IFN-γ at 48 h p.i. To quantitate the extent of cell death, we measured the release of LDH in the supernatants of these cultures. LDH release was found in the supernatants of HCMV-infected HFF cultures treated with IFN-γ, TRAIL, or IFN-γ and TRAIL but was not found in uninfected HFF cultures treated with these cytokines (Fig. 1⇓). In fact, uninfected HFF cultures treated with IFN-γ and TRAIL appeared healthy throughout the experiment. The ability of TRAIL to induce the death of HCMV-infected cells was comparable with the antiviral cytokine TNF-α (31, 32) in that HCMV-infected cells cultured with TNF-α or TNF-α plus IFN-γ were also killed by 48 h p.i. (Fig. 1⇓). Importantly, the death-inducing effects of TRAIL were not specific to HCMV-infected HFF cells, because TRAIL also induced cell death in RSV-infected NHBE cells, especially in the presence of IFN-γ (data not shown). These data indicate that TRAIL, either alone or in combination with IFN-γ, can induce cell death in virus-infected cultures in vitro.
LDH release from HFF cells infected with HCMV and treated with TRAIL and IFN-γ. HFF cell death was measured by LDH release in supernatants of six replicate cultures, mock-infected or HCMV-infected, and treated with IFN-γ, TRAIL, TNF-α, IFN-γ and TRAIL, or IFN-γ and TNF-α for 48 h. Data shown represent means of individually assayed supernatants. Statistically significant differences (Student’s t test, p = 0.0001) in LDH release in IFN-γ or TRAIL-treated HCMV-infected cultures compared with untreated HCMV-infected cultures are indicated (*). Statistically significant differences (p = 0.0001) also exist between the LDH assayed in HCMV-infected cultures treated with IFN-γ or TRAIL alone compared with HCMV-infected cultures incubated with TRAIL and IFN-γ (**). SDs were ≤5% and have been omitted for clarity. Results shown are representative of three independent experiments.
To delineate that TRAIL was specifically killing virus-infected cells, we cultured uninfected or HCMV-infected HFF cells in the presence of TRAIL and IFN-γ for 24 or 48 h. We subsequently stained these cells with an FITC-conjugated anti-HCMV Ab and with annexin-V-biotin and SA-Texas Red. Using confocal microscopy, we were therefore able to specifically detect the virus-infected cells and determine whether they were undergoing apoptosis. By 24 h p.i., HCMV-infected cells cultured in the presence of TRAIL and IFN-γ exhibited rounded morphology, HCMV-specific staining, and annexin-V surface staining, indicating that the virus-infected cells were undergoing early stages of apoptotic cell death (Fig. 2⇓A). By 48 h p.i., many of the FITC+ cells also appeared to be in the late stages of apoptosis, because they stained very brightly for annexin-V, were reduced in size, and showed drastically altered morphology consistent with apoptotic cell death (Fig. 2⇓B). In contrast, uninfected HFF cells cultured in the presence of TRAIL and IFN-γ were negative for annexin-V staining (Fig. 2⇓C), and HCMV-infected cells cultured in media alone showed cytoplasmic FITC staining indicative of HCMV gene expression but exhibited no signs of apoptosis as indicated by annexin-V staining (data not shown). As an estimate of the extent of cell death, the percentage of annexin-V+ HCMV+ cells in each culture was calculated from replicate randomly selected low power fields (Table I⇓). We found that 90% of the HCMV-infected FITC+ cells were annexin-V+ in the TRAIL- and IFN-γ-treated cultures, but ≤5% of uninfected cultures were annexin-V+ when cultured with these cytokines for 48 h (Table I⇓). These data indicate that TRAIL can selectively induce apoptosis of virus-infected HFF cells but not uninfected HFF cells.
Confocal microscopic images of mock-infected or HCMV-infected HFF cells after culture in IFN-γ and TRAIL. HCMV-infected HFF cells were mock-infected or infected with HCMV at an moi of 2 and cultured in IFN-γ and TRAIL. A, HCMV-infected HFF cells showed a cytopathic rounding morphology due to virus infection by 24 h p.i. and stained with FITC anti-HCMV Abs (green) as well as cell surface annexin-V/Texas Red, indicating that HCMV-infected cells were undergoing early stages of apoptosis. B, By 48 h p.i., cells undergoing later stages of apoptosis were evident. C, Mock-infected HFF cells cultured in IFN-γ and TRAIL for 48 h exhibited no annexin-V staining or FITC anti-HCMV staining. The scale bar represents 5 μm.
Percentage of annexin V+ HCMV+ cells in TRAIL-treated culturesa
TRAIL-mediated killing precedes maximal virus replication
To assess whether the TRAIL-induced apoptosis of the virus-infected cells occurred before productive virus replication and whether TRAIL has antiviral activity on other viruses in different cell types, we measured the amount of infectious progeny virus recoverable in TRAIL-treated cultures infected with HCMV, RSV, and VV. IFN-γ alone caused a notable decrease in recoverable HCMV, RSV, and VV, consistent with its known antiviral activity (Table II⇓). Exposure of cells to TRAIL alone also resulted in reduced recovery of HCMV and RSV but not of VV. Importantly, treating cells with IFN-γ plus TRAIL resulted in a further reduction of titratable HCMV, RSV, and VV compared with cells treated with IFN-γ or TRAIL alone (Table II⇓). Although both RSV-infected NHBE cells and VV-infected HFF cells appeared to be less sensitive to the apoptosis mediated by TRAIL than did HCMV-infected HFF cells, the combination of TRAIL and IFN-γ still had an impact on the replication of these viruses (Table II⇓). Therefore, TRAIL-mediated apoptosis of virus-infected cells was not specific only to HCMV-infected cells and occurred before maximal virus replication. Thus, TRAIL has antiviral activity by inducing apoptosis of virus-infected cells in vitro. It is unknown whether TRAIL can exert additional antiviral activity distinct from apoptosis induction.
Effect of TRAIL on virus replication
HCMV and cytokines regulate TRAIL and TRAIL-R mRNA expression
Next, we investigated the mechanism by which TRAIL selectively kills only the virus-infected cells. Semiquantitative RT-PCR analysis was used to study the relative abundance of distinct TRAIL-R mRNA in HCMV-infected or uninfected HFF cells or after exposure to cytokines. IFN-γ clearly decreased the abundance of TRAIL-R3 and TRAIL-R4 mRNA in uninfected or infected HFF cells by 24 h posttreatment. By 48 h after HCMV infection, TRAIL-R4 mRNA was present at levels similar to those seen in untreated or uninfected cells, even in the presence of IFN-γ (Fig. 3⇓). In contrast, HCMV infection resulted in an increase in mRNA for TRAIL-R2 by 24 h p.i. and in an increase for both TRAIL-R1 and TRAIL-R2 by 48 h p.i. By comparison, TNFR-p55 and the housekeeping gene DHFR mRNA were detectable at relatively similar levels in all samples irrespective of treatment with cytokines; TNFR-p75 mRNA was inducible by HCMV infection under all culture conditions (Fig. 3⇓). We also performed RT-PCR for TRAIL mRNA and found increased TRAIL mRNA after HCMV infection, either in the presence or absence of IFN-γ. As an indicator of viral infection, RT-PCR for the HCMV early gene UL-18 and late gene UL-67 was also performed (data not shown). Taken together, the RT-PCR data indicate that clear differences exist in the relative abundance of TRAIL and TRAIL-R mRNA in HFF cells after treatment with IFN-γ or after HCMV infection. Thus, IFN-γ and HCMV infection either transcriptionally regulates the expression of TRAIL and TRAIL-Rs in HFF cells or alters the stability or half-life of these mRNAs by unknown mechanisms.
RT-PCR analysis of TRAIL and TRAIL-R mRNA in HFF cells after HCMV infection and exposure to IFN-γ, CD40L, and TRAIL. RNA was harvested from mock-infected (virus: −) HFF cells cultured in the presence of IFN-γ, CD40L, or TRAIL for 24 h or from HCMV-infected (virus: +) HFF cells cultured in the same cytokines for 24 and 48 h p.i. First strand cDNAs were used for PCR analysis to detect transcripts specific for human TRAIL-R1, TRAIL-R2, TRAIL-R3, TRAIL-R4, TRAIL, TNFR-p55, TNFR-p75, and the housekeeping gene DHFR. Data shown are ethidium bromide-stained agarose gels that are representative of at least two independent PCR performed at various cycles to ensure that detection was in the linear range. Plasmids encoding TRAIL-R cDNAs (right panel) or TNFR cDNAs (data not shown) were used as template controls to indicate primer specificity.
Modulation of TRAIL and TRAIL-R proteins correlates with mRNA expression
It was important to determine whether the relative changes in abundance of mRNA transcripts for these receptors correlated with their expression as cell surface proteins and whether these or other cytokines could similarly influence TRAIL expression. First, we assessed the expression of each distinct TRAIL-R on uninfected HFF cells using TRAIL receptor-specific mAbs. HFF cells expressed TRAIL-R2 and low but detectable levels of TRAIL-R1, but not TRAIL-R3 or TRAIL-R4, as assessed relevant to staining with a control Ab (Fig. 4⇓A). Treatment with IFN-γ for 24 h resulted in decreased expression of TRAIL-R1 and TRAIL-R2 on uninfected HFF cells (Fig. 4⇓B). Furthermore, we found that IFN-γ also down-regulated the expression of TRAIL-R3 and -R4 on primary human B cells (data not shown). In contrast, HCMV infection directly up-regulated the expression of death domain-containing TRAIL-R1, and TRAIL-R2 by 24 h p.i. (Fig. 4⇓B). Interestingly, HCMV-infected HFF cells cultured in IFN-γ did not exhibit decreased TRAIL-R2 expression. Thus, in the presence of IFN-γ, the expression of TRAIL-R1 and-R2 was significantly higher on HCMV-infected HFF cells than on uninfected HFF cells (Fig. 4⇓B). For comparison, we also stained HFF cells with specific Abs to TNFR-p55 and TNFR-p75 and found that the expression of TNFR-p55 was not significantly altered after treatment with IFN-γ, IFN-αδ, CD40L, or TRAIL (data not shown), whereas TNFR-p55 and TNFR-p75 were both up-regulated on HCMV-infected HFF cells (data not shown). We were particularly interested in determining whether TRAIL itself was expressed on virus-infected HFF cells, because we had found that HCMV infection induced TRAIL mRNA. Using an Ab specific to human TRAIL, we found that both IFN-γ and TNF-α, but not a control reagent, RANKL, induced a small but reproducible increase in TRAIL expression on uninfected HFF cells (Fig. 4⇓C). Furthermore, HCMV infection resulted in a dramatic induction of TRAIL expression on HFF cells (Fig. 4⇓C). These results are consistent with the relative abundance of TRAIL and TRAIL-R mRNA detected by RT-PCR (Fig. 3⇑) and confirmed that exposure to cytokines or HCMV infection directly influences both TRAIL and TRAIL-R expression on primary human fibroblasts.
FACS analysis of TRAIL and TRAIL-R expression on HFF cells cultured in the presence of cytokines or after infection with HCMV. A, Normal basal expression of TRAIL-Rs on HFF cells using specific mouse mAbs (open histograms) compared with staining with a control Ab (shaded histograms). B, Surface protein expression levels of TRAIL-R1 and TRAIL-R2 on (i) untreated HFF cells (solid line) vs 24 h after culture in IFN-γ (dotted line), (ii) HCMV infection at an moi of 2 (dashed line), or (iii) HCMV infection in the presence of IFN-γ (thick solid line) overlayed on uninfected cells in IFN-γ alone (dotted line). C, Detection of TRAIL on HFF cells cultured in normal growth media (solid line); 24 h after culture with IFN-γ (dotted line), TNF-α (dotted line), RANKL (dotted line); or 24 h after HCMV infection (dotted line). Data shown are representative of three repeat experiments.
HCMV and cytokines influence NF-κB activation
TRAIL-R1, TRAIL-R2, and TRAIL-R4 are capable of inducing the activation of NF-κB proteins in vitro (19, 22). Because HFF cells normally express detectable TRAIL-R2 protein but are not normally killed by TRAIL, we assessed the extent of NF-κB activation in unstimulated HFF cells after the ligation of TRAIL-R2 by TRAIL, after exposure to IFN-γ or to the control reagent CD40L, and after HCMV infection. We found that treating HFF cells with IFN-γ or IFN-γ plus TRAIL resulted in a 10-fold decrease in basal NF-κB activation (Fig. 5⇓A), whereas transfection with the parent promoterless plasmid did not induce luciferase activity (data not shown). For control purposes, HFF cells were transfected with a γ activation site containing, IFN-γ-responsive luciferase reporter plasmid pSTAT-LUC; as expected, IFN-γ induced a 4-fold increase in pSTAT-LUC-mediated luciferase activity (data not shown). Thus, TRAIL does not activate NF-κB proteins in uninfected HFF cells, even though they express TRAIL-R2 and are not normally killed by TRAIL, whereas IFN-γ clearly suppressed basal NF-κB levels in uninfected HFF cells.
NF-κB activation in HFF cells after treatment with cytokines or HCMV infection. A, NF-κB activity as measured by luciferase assay in uninfected HFF cells after 24 h of culture in IFN-γ, CD40L, TRAIL, or combinations of these cytokines. B, NF-κB activity in HFF cells at 24 h after HCMV infection plus or minus treatment with the same cytokines. HFF cells (104 cells/well) were transfected in triplicate and mock-infected or infected with HCMV at an moi of 2; lysates were assayed for luciferase activity at 24 h p.i. Results shown are means of independently assayed triplicates. For clarity, SDs are not shown but were ≤7%. The relative increase or decrease in NF-κB activity with respect to that measured in HFF cells cultured in media alone is also indicated numerically above the graph bars. Note that HCMV infection results in increased NF-κB activity; hence, there are differences in the scale on the y-axis on B.
We also found vastly increased basal NF-κB activity in HCMV-infected HFF cells (Fig. 5⇑B), which is consistent with the fact that HCMV is known to be a potent activator of NF-κB proteins (33). However, treating HCMV-infected HFF cells with TRAIL and IFN-γ, either individually or in combination, decreased NF-κB activity by ∼50% (Fig. 5⇑B). Treatment with the control reagent CD40L significantly augmented NF-κB activity to almost 4-fold above that found in HCMV-infected fibroblasts, which is consistent with the fact that CD40 is known to be expressed on most fibroblasts, including foreskin fibroblasts, and can activate NF-κB (33, 34) Furthermore, CD40L-treated HCMV-infected HFF cells cultured in TRAIL were found to have a 10-fold induction of NF-κB activity compared with infected cells cultured in TRAIL alone, indicating that CD40L was sufficient to overcome the decrease in NF-κB activity induced by TRAIL. Similar to uninfected cells, control pSTAT-LUC-transfected HCMV-infected HFF cells showed increased IFN-γ-inducible luciferase activity over uninfected cells (data not shown). Taken together, these results confirm the finding that HCMV infection induced the activation of NF-κB proteins, but this was insufficient to prevent TRAIL-mediated apoptosis of infected HFF cells.
Discussion
We have demonstrated that primary human fibroblasts, which are normally resistant to TRAIL-mediated death, become susceptible to TRAIL after virus infection. Since TRAIL does not kill most normal cells, we believe that our results may be closely analogous to events that occur in vivo. It is in this setting that we have documented that TRAIL is inducible by IFN-γ, TNF-α, and CMV infection and has potent antiviral activity in vitro by selectively killing virus-infected fibroblasts. Furthermore, we have demonstrated that virus infection and cytokines alter the expression of TRAIL-Rs.
It has been suggested that the reason why some cells are susceptible to TRAIL while others appear to be resistant lies in the endogenous presence of the non-death-signaling TRAIL-Rs (35). However, examination of various tumor cell types has shown that basal expression of TRAIL-Rs does not correlate with susceptibility to TRAIL (30, 36). Alternatively, the presence or absence of intracellular inhibitors of apoptosis such as the cellular inhibitor of caspase 8/FLICE-inhibitory protein (cFLIP) may be important (37), and in some circumstances, cFLIP expression does correlate with the susceptibility of tumor cells to TRAIL (36). In contrast, we have shown here that cytokines and virus infection differentially modulate TRAIL and TRAIL-R expression at the cell surface, and that these alterations directly correlate with susceptibility versus resistance to TRAIL. Therefore, we propose that modulation of TRAIL-R expression and the relative abundance of each distinct receptor at the time of exposure to TRAIL significantly contribute to regulating susceptibility to TRAIL-mediated apoptosis. In the case of HCMV-infected HFF cells, virus infection induced the expression of the death-domain containing receptors TRAIL-R1 and TRAIL-R2, whereas IFN-γ down-regulated TRAIL-Rs on uninfected HFF cells. Thus IFN-γ influences the susceptibility of HFF cells to TRAIL, in part by regulating the abundance of specific TRAIL-Rs at the cell surface. IFN-γ might also have effectively concentrated soluble TRAIL to target the virus-infected cells. Moreover, both HCMV infection and IFN-γ directly induced TRAIL protein expression on HFF cells, indicating that endogenous expression of TRAIL itself may act to induce apoptosis via suicide and/or fratricide.
The modulation of TRAIL-R expression need not exclude a role for intracellular factors in contributing to susceptibility to TRAIL. In this regard, it is curious that global activation of NF-κB by HCMV infection was itself insufficient to prevent TRAIL-mediated apoptosis, yet activation of NFκB has been widely described as a major prosurvival factor, capable of overcoming the apoptosis mediated by death-inducing ligands in vitro (38). This might be explained in terms of the order in which survival or proapoptotic events occur, the magnitude of NF-κB activation, or differential induction of distinct NF-κB/Rel proteins. The presence and abundance of different NF-κB proteins has recently been shown to explain the ability of CD40L versus LPS to initiate Ig class switching in B cells (39). Interestingly, IFN-γ not only modulated TRAIL-R expression but also significantly decreased NF-κB-activation, which presumably enhanced susceptibility to TRAIL by decreasing the survival threshold of these cells. Moreover, the viral impact on host cell metabolism may also sensitize cells to TRAIL in an analogous manner to protein or nucleic acid inhibitors such as cycloheximide or actinomycin D. These agents are often capable of sensitizing cells to TRAIL-mediated apoptosis in vitro (data not shown). Hence, our results are consistent with a situation whereby the balance of apoptotic versus survival signals cumulatively dictates the final outcome of susceptibility versus resistance to TRAIL-mediated cell death. However, our data are also consistent with the idea that modulation of TRAIL-R expression at the cell surface is a crucial upstream event that influences this process.
It is important to consider the question of whether TRAIL is normally expressed in a manner relevant to virus infection in vivo. We have demonstrated that TRAIL is inducible by HCMV infection as well as by the antiviral cytokines IFN-γ and TNF-α (31, 32). Moreover, we have also shown recently that IFN-γ induces functional TRAIL expression on freshly isolated human peripheral blood monocytes (40), and TRAIL has also been found to be a functional cytotoxic effector molecule expressed on primary human NK cells (41). Induction of IFN-γ and the role of NK cells and macrophages during virus infection in both humans and mice have been well documented, although characterization of TRAIL expression and its role during virus infection in vivo remains to be investigated. Interestingly, a role for TRAIL in the activation-induced cell death of T cells in HIV-infected individuals has also been proposed (42). Therefore, a role for TRAIL in virus infection in vivo is strongly implied. Finally, the ability of IFN-γ to modulate both TRAIL and TRAIL-R expression on the same cell is currently undefined at a molecular level, and further studies will be required to dissect which signaling proteins mediate these events.
In conclusion, we have demonstrated that TRAIL is directly inducible on primary human fibroblasts by IFN-γ, TNF-α, and HCMV infection and can selectively kill virally infected cells. We have also documented that virus infection and IFN-γ differentially regulate the expression of TRAIL-Rs, which directly correlates with the susceptibility of virus-infected cells to TRAIL-mediated apoptosis. Therefore, the differential regulation of TRAIL and TRAIL-R expression on virus-infected cells represents a novel mechanism by which IFN-γ mediates antiviral activity.
Acknowledgments
We thank Phil Morrissey, Laurent Galibert, Hilary McKenna, Eilidh Williamson, and Doug Williams for critical reading of the manuscript. We also thank Daniel Hirschstein, Steve Braddy, and Alan Alpert for FACS assistance; Charles Rauch for purification of LZ proteins; Sally Painter, Kim Harrington, and Mei-Ling Hsu for assistance with virus work; Beth Daro, John Doedens, and Adel Youakim for help and suggestions with confocal microscopy; and Anne Aumell for editorial assistance.
Footnotes
-
↵1 Address correspondence and reprint requests to Dr. Lisa M. Sedger, Department of Molecular Immunology, Immunex Corporation, 51 University Street, Seattle, WA 98101. E-mail address: sedgerl{at}immunex.com
-
↵2 Abbreviations used in this paper: TRAIL, TNF-related apoptosis-inducing ligand; TRAIL-R, TRAIL receptor; HCMV, human CMV; RSV, respiratory syncytial virus; moi, multiplicity of infection; TNFR, TNF receptor; CD40L, CD40 ligand; DcR, decoy receptor; HFF, human foreskin fibroblast; NHBE, normal human bronchial airway epithelial; VV, vaccinia virus; DHFR, dihydrofolate reductase; LDH, lactate dehydrogenase; SA, streptavidin; p.i., postinfection; rhu, recombinant human; LZ, leucine zipper; RANKL, receptor activator of NF-κB ligand.
- Received February 8, 1999.
- Accepted April 30, 1999.
- Copyright © 1999 by The American Association of Immunologists
References
In this issue
