Materials and Methods

Cell culture Human microvascular endothelial cells (HMECs) were cultured in MCDB 131 medium (Gibco) supplemented with L-glutamine (10 mmol/L; Gibco), mouse epidermal growth factor (10 ng/mL; BD Bioscience), hydrocortisone (1 μg/mL; Sigma), and 10% heat-inactivated fetal bovine serum [11]. Cells were grown in a humidified incubator with 5% CO₂ at 37°C. Approximately 24 h before infection, antibiotics were withdrawn from the culture medium

Infection and treatment of cells A plaque-purified seed stock (1×10⁷–5×10⁷ pfu/mL) of R. rickettsii (Sheila Smith strain) was prepared from infected Vero cells [12]. Cells were infected with approximately 5×10⁴ pfu for every square centimeter of cell culture area, as described elsewhere [13]. After 3 h of incubation, to allow for bacterial adhesion and invasion, medium that contained free extracellular organisms was removed and the cell monolayer was quickly washed with fresh medium. Intracellular infection was then continued for another 21 or 45 h (total duration, 24 or 48 h) before the introduction of staurosporine (275 nmol/mL; Sigma) for 24 h before cells were harvested

MTT cell proliferation assay Cell viability was measured by the reduction of tetrazolium salt (3-[4,5-dimethylthiazolyl-2]-2, 5-diphenyltetrazolium bromide) by metabolically active cells. HMECs were plated in triplicate in a 96-well plate and allowed to achieve ∼80%–90% confluence. Increasing concentrations of staurosporine were added in fresh medium (100 μL per well), followed by incubation for 24 h. MTT reagent (10 μL; R&D Systems) was added directly to the cells, which were incubated at 37°C for approximately 4 h. Detergent (100 μL) was subsequently added, and the cells were allowed to incubate at room temperature in the dark for another 4 h. Absorbance was measured at 570 nm with an EL_X800 microplate reader (BioTek Instruments). Average values from triplicate readings were determined by subtracting the average value from the average of the blank, and standard deviations were calculated

Western blot analysis Total cellular protein extracts were prepared in SDS-containing lysis buffer supplemented with protease inhibitors (Sigma). Equal amounts of protein were subjected to SDS–PAGE on 12% (wt/vol) polyacrylamide gels, followed by wet-tank transfer to a nitrocellulose membrane (Bio-Rad Laboratories). The membranes were probed with primary antibodies specific for human cIAP2 (Santa Cruz Biotechnology) in a 5% nonfat dry milk solution. The protein-antibody complexes were visualized using compatible horseradish peroxidase–conjugated secondary antibodies (Santa Cruz Biotechnology) and an enhanced chemiluminescence–based detection system. The blots were then stripped and reprobed with a monoclonal antibody against α-tubulin (Accurate Chemical)

Annexin staining and fluorescence-activated cell sorting analysis HMECs were washed twice with PBS before the addition of 0.025% (wt/vol) trypsin (Gibco), to detach and lift cells from the culture surface. Fresh medium supplemented with 22 μg/mL tetracycline (Sigma) was added to the suspension to inactivate trypsin and intracellular rickettsiae. Cells were centrifuged at 330 g for 10 min at 4°C. The pellet was washed with precooled PBS containing tetracycline and was centrifuged again. Annexin staining solution was prepared from the TACS Annexin V-FITC apoptosis detection kit (R&D Systems) with use of 10× binding buffer (10%), annexin V-FITC (2%), and 20 μg/mL tetracycline in water (88%), and incubated with cells (∼1×10⁶ cells) for 15 min at room temperature in the dark. Cells were washed in binding buffer, fixed with 1% paraformaldehyde, and stored at 4°C until subjected to fluorescence-activated cell sorting analysis. Flow cytometry was performed on ⩾10,000 cells with a FACSCalibur dual laser cytometer, and data analysis was performed using CellQuest Pro (Becton Dickinson)

Oligo microarray analysis HMECs were infected for 48 h and/or challenged with staurosporine for an additional 24 h. RNA from cells was extracted using the Array Grade Total RNA isolation kit (Superarray Bioscience). Total RNA (3 μg) was converted to cDNA at 42°C for 50 min, and cDNA was then transcribed to cRNA in the presence of biotin-16-UTP (Roche Molecular Biochemicals) overnight by use of the TrueLabeling-AMP 2.0 kit (Superarray). The probe was purified using the ArrayGrade cRNA cleanup kit (Superarray), in accordance with the manufacturer’s instructions, and was hybridized with each human apoptosis microarray overnight at 60°C. The membranes were first washed with 2× standard sodium chloride plus sodium citrate and 1% SDS, followed by washing with 0.1× standard sodium citrate and 1% SDS at 60°C for 15 minutes each. After the washes, signals were detected by chemiluminescence by use of alkaline phosphate–conjugated streptavidin and CDP-Star chemiluminescent substrate (SABiosciences). The data were analyzed using GEArray Expression Analysis Suite (SABiosciences). Relative induction or suppression of target genes under different experimental conditions was calculated using normalization to the genes that encode β-2-microglobulin and/or the 90-kDa heat-shock protein, with a threshold setting of a 2.5-fold change in gene expression

Analysis of cIAP2 mRNA expression RNA was isolated from infected and uninfected HMECs with TRI Reagent (Molecular Research Center), in accordance with the manufacturer’s protocol. In brief, cDNA was synthesized from RNA (4 μg) by use of SuperScript II (Invitrogen), in accordance with the manufacturer’s protocol. Polymerase chain reaction (PCR) amplification was performed by using a cIAP2-specific primer pair described elsewhere [14]. All reactions were performed in the linear range and compared with a glyceraldehyde 3-phosphate dehydrogenase (GAPDH) reference control. The PCR products for the GAPDH and cIAP2 genes, 175 and 342 bp, respectively, were resolved on a 2% agarose gel and visualized with ethidium bromide staining. For quantitative real-time PCR, total RNA was further purified using a qPCR-grade RNA isolation kit (SuperArray) and quantified using a Nanodrop ND-1000 spectrophotometer (Thermo Scientific). cDNA synthesis was performed using RT² First strand kit (SuperArray) with 0.5 μg of total RNA. Quantitative PCR reactions were performed in triplicate in a MyiQ cycler (BioRad) by use of human cIAP2 and GAPDH primers with SYBR Green, in accordance with the manufacturer’s instructions (Superarray). The gene expression level for cIAP2 was normalized to the GAPDH gene expression level and analyzed by use of the ΔΔ cycle threshold method

cIAP2 knockdown by siRNA HMECs were seeded into T-150 flasks 3 days before nucleofection and harvested with trypsin when they reached ∼70% confluence. Each nucleofection tube contained approximately 2.5×10⁶ cells and 100 μL nucleofector solution R as well as 2 μg pmaxGFP, 2 μg pmaxGFP and 200 nmol/L control siRNA, or cIAP2 siRNA (Dharmacon). SMARTpool reagents (Dharmacon) combine 4 SMARTselection-designed siRNAs into 1 highly effective siRNA pool. This combination of siRNAs is thought to mimic the natural silencing pathway. The negative control siRNA consisted of the 4 pooled nucleotide sequences with ⩾4 mismatches to any human gene. To minimize the cells’ exposure to nucleofector solution, the samples were prepared and processed in a staggered fashion. The nucleofection equipment was set on program T-016, as recommended by the supplier of the HMECs. After the procedure, the electroporation cuvettes were washed with 1 mL warm media; cells were immediately transferred into 500 μL of additional warmed media in a 12-well tissue culture plate and placed into a 37°C incubator for 4 h

Densitometric and statistical analyses To quantitatively analyze Western blot results, the films were scanned in grayscale mode using an HP ScanJet 6300C scanner (HP) at a resolution of 600 dpi. Volume analysis was performed using ImageQuant (version 3.3; Molecular Dynamics), and band intensities were quantified as densitometric units. For comparison, the normalized band intensity for the uninfected control in each experiment was assigned a value of 1. Data were calculated as the mean value ± SE for at least 3 independent experiments, and comparisons between the study and control groups were performed with Student’s t test. P values ⩽.05 were considered to be statistically significant

Results

Staurosporine as an inducer of apoptosis in HMECs First, we investigated whether staurosporine was able to induce apoptosis in HMECs, as it can in other cell types [15 –17]. MTT assay on HMECs incubated with 0–400 nmol/L staurosporine for 24 h revealed a dose-dependent loss of viability with an IC₅₀ of approximately 275 nmol/L (figure 1 A). Flow cytometric analysis further demonstrated that at 24 h, ∼7% of untreated, control cells displayed annexin V positivity, whereas treatment with 275 nmol/L staurosporine resulted in a phenotype consistent with early apoptosis in ∼34% of cells (figure 1 B and 1 C)

View larger version:

• In this window
• In a new window

• Download as PowerPoint Slide

Figure 1

Staurosporine (STS)-induced apoptosis in cultured human microvascular endothelial cells (HMECs). A MTT assay to determine cell viability after 24 h of treatment of HMECs with STS. Cultured HMECs displayed dose-dependent sensitivity to STS with an IC₅₀ value of 275 nmol/L. B and C Analysis of phosphatidylserine asymmetry in HMECs by flow cytometry. HMECs were either left untreated (Con) or subjected to treatment with 275 nmol/L STS for 24 h. Annexin V staining was performed using TACS Annexin V-FITC apoptosis detection kit (R&D Systems). For all experimental conditions, ⩾10,000 cells were analyzed on a FACSCalibur instrument (Becton Dickinson). The percentage of annexin V–positive cells for each experimental condition is indicated

Response to staurosporine and expression of pro- and anti-apoptotic genes in endothelial cells infected with R. rickettsiiTo determine whether R. rickettsii infection protects against staurosporine-induced apoptosis in HMECs, cells infected with R. rickettsii for 24 or 48 h were subjected to 275 nmol/L staurosporine for an additional 24 h. As expected, treatment of uninfected HMECs with staurosporine at both 24 h and 48 h led to significant increases in annexin V staining (P=.028 and P=.053, respectively). However, R. rickettsii infection for similar durations did not have a significant effect on the levels of apoptosis. Interestingly, R. rickettsii–infected HMECs were able to resist staurosporine-mediated induction of apoptosis when challenged at 24 and 48 h after infection (P=.002 and P=.019, respectively) (figure 2 A and 2 B)

View larger version:

• In this window
• In a new window

• Download as PowerPoint Slide

Figure 2

Infection with Rickettsia rickettsii protects human microvascular endothelial cells (HMECs) from staurosporine (STS)–induced apoptosis. A Cells infected with viable R. rickettsii (∼5×10⁴ pfu/cm²) for 24 h and corresponding uninfected controls were subjected to STS challenge (275 nmol/L for 24 h). After washing, cells collected by trypsinization were processed for annexin V–FITC staining and analyzed on a FACSCalibur instrument (Becton Dickinson). The experimental conditions shown are as follows: C, uninfected, untreated control HMECs; STS, uninfected, STS-treated HMECs; Rr infected but not treated HMECs; and Rr + STS, infected, STS-treated HMECs. Data represent mean values (±SE) from a minimum of 3 independent experiments. B Similar analysis performed with HMECs infected for 48 h prior to STS treatment for 24 h. *P<.05, compared with uninfected, untreated control. ^#Statistically significant decrease in the percentage of annexin V–positive cells, compared with uninfected, STS-treated HMECs

To understand the mechanism(s) underlying infection-induced resistance to apoptosis, we performed DNA microarray analysis to screen the expression profiles of 112 pro- and anti-apoptotic genes. Employing a threshold of 2.5-fold or higher for changes in gene expression, we identified several key genes that had altered levels of expression at 48 h. R. rickettsii infection increased expression of 3 genes (TRAF3, AS1 and TNFRFS6B) involved in the protection against apoptosis and 2 genes (AS1R2 and TP53) implicated in the initiation of apoptosis. No genes were detected that had down-regulated expression in response to infection (data not shown). Several genes were affected by infection (48 h) followed by staurosporine treatment for 24 h (figure 3 and table 1). Although a predominance of anti-apoptosis genes was found to be significantly up-regulated, the expression of several pro-apoptotis genes (BOK, BCL2L13, TP53, ABL1 and BAK1) was also increased. The known anti-apoptosis genes that had their expression induced in response to R. rickettsii infection included TRAF1, BNIP2, BCL2L1 (Bcl-X_L), TRAF3, BIRC2 (cIAP1), BIRC3 (cIAP2), BIRC5 (survivin), BNIP3L, BAG3 and AKT1 (figure 3)

View larger version:

• In this window
• In a new window

• Download as PowerPoint Slide

Figure 3

Infection of human microvascular endothelial cells (HMECs) with Rickettsia rickettsii alters gene expression. HMECs were either left uninfected or infected with R. rickettsii for 48 h and were then challenged with staurosporine for 24 h. The results are presented as the relative induction of each gene, normalized to the levels of β₂-microglobulin and 90-kDa heat-shock protein 1β mRNA, with a threshold setting of a 2.5-fold change in gene expression. The scatter plot displays the fold-difference in the relative expression of genes in staurosporine-treated, uninfected HMECs (STS) versus staurosporine-treated, R. rickettsii–infected HMECs (Rr + STS)

R. rickettsii infection of HMECs and cIAP2 mRNA and protein expression cIAP2 is an important cellular anti-apoptotic protein that has been implicated in protection against apoptosis during Chlamydia infection [18, 19]. Because our array data indicated increased cIAP2 expression during R. rickettsii infection of HMECs, we quantitatively measured the steady-state levels of cIAP2 mRNA and protein in R. rickettsii–infected HMECs. Detailed analysis by semiquantitative reverse-transcriptase PCR and real-time quantitative PCR revealed a very low baseline level of cIAP2 transcription in uninfected HMECs but significantly increased levels cIAP2 mRNA in HMECs infected with R. rickettsii for 24 and 48 h (figure 4 A and 4 B). As expected, uninfected HMECs expressed very low levels of cIAP2 protein, but significant increases in cellular cIAP2 protein levels were observed at both 24 and 48 h after infection. Interestingly, cIAP2 protein levels remained elevated in HMECs infected with R. rickettsii for either 24 or 48 h and subsequently challenged with staurosporine, whereas staurosporine treatment alone had no significant effect (figure 4 C)

View larger version:

• In this window
• In a new window

• Download as PowerPoint Slide

Figure 4

Increased expression of cIAP2 mRNA and protein during Rickettsia rickettsii infection of human microvascular endothelial cells (HMECs). A RNA samples from uninfected HMECs (C) or HMECs infected with 2 independent stock preparations of R. rickettsii (Rr) for 24 and 48 h were subjected to reverse-transcriptase polymerase chain reaction (PCR) for analysis of cIAP2 mRNA expression. A representative gel with cIAP2 and glyceraldehyde 3-phosphate dehydrogenase (GAPDH) amplicons of 342 and 172 bp, respectively, is shown. B Quantitative PCR analysis of cIAP2 expression in R. rickettsii–infected HMECs (Rr) at 24 and 48 h after infection and simultaneously processed uninfected controls (C). PCR reactions for each RNA sample were run in triplicate and the level of cIAP2 expression was normalized to that for the housekeeping gene GAPDH. For comparison, the ratio of cIAP2 and GAPDH expression in RNA preparations from uninfected HMECs was assigned a value of 1. C The levels of cIAP2 protein (70 kDa) in lysates of uninfected, untreated HMECs (C); uninfected, staurosporine-treated HMECs (STS); R. rickettsii–infected, untreated HMECs (Rr); and R. rickettsii–infected, staurosporine-challenged HMECs (Rr + STS) were determined by immunoblotting (at least 3 experiments). To ensure equal protein loading of different samples, blots were stripped and probed with an antibody against human α-tubulin (50 kDa). A representative result depicting relative changes in cIAP2 level attributable to R. rickettsii infection for 24 and 48 h and after 24 h treatment with staurosporine is shown

siRNA-mediated cIAP2 knockdown and infection-induced protection against apoptosis To evaluate the potential role of cIAP2 in protection against staurosporine-induced apoptosis, we employed the siRNA-mediated knockdown approach. In initial experiments, a green fluorescence protein–expressing plasmid was introduced into HMECs via nucleofection, and the procedure’s efficiency was analyzed by flow cytometry. Approximately 56% and 67% of cells expressed green fluorescence protein at 24 and 48 h after nucleofection, respectively, indicating uptake, expression, and stability of DNA. As expected, <5% of cells expressed green fluorescence protein in mock-transfected HMECs. Next, we confirmed the efficacy of cIAP2 siRNA-mediated knockdown in diminishing the protein level in HMECs. Western blot analysis revealed that after nucleofection, cIAP2 protein levels were significantly reduced at 24 and 48 h, as evidenced by the presence of 50% less target protein in HMECs transfected with cIAP2 siRNA, compared with cells transfected with scrambled, control siRNA (n=2; data not shown). At both 24 and 48 h, flow cytometric analysis after annexin V staining of HMECs with diminished cIAP2 levels revealed baseline levels of apoptosis significantly greater than those of cells nucleofected with non-targeting siRNA. Staurosporine treatment of HMECs with decreased cIAP2 levels (as a result of siRNA-mediated knockdown) did not increase apoptosis, compared with baseline levels of apoptosis or with levels of apoptosis in cells transfected with control siRNA. Following 24 h of infection with R. rickettsii annexin V positivity in cIAP2 siRNA–transfected HMECs was significantly lower than that in corresponding untreated control cells and only slightly higher than the baseline level for HMECs containing control siRNA. Cells infected for 24 h and subsequently challenged with staurosporine had a significant decrease in phosphatidylserine exposure, compared with those treated with staurosporine alone (P=.013) (figure 5). At 48 h, cIAP2 siRNA-transfected HMECs that were either untreated or challenged with staurosporine had similar levels of annexin V binding (∼19% and ∼17%, respectively), which indicated that introduction of cIAP2 siRNA did not augment susceptibility to staurosporine in HMECs. R. rickettsii infection for 48 h resulted in a slight reduction below the baseline level of annexin V binding in cIAP2 siRNA-transfected cells, whereas infected cells challenged with staurosporine exhibited even lower levels of apoptosis (figure 5)

View larger version:

• In this window
• In a new window

• Download as PowerPoint Slide

Figure 5

siRNA-mediated cIAP2 knockdown has no effect on protection against staurosporine-induced apoptosis in Rickettsia rickettsii–infected human microvascular endothelial cells. Cells were subjected to nucleofection with a green fluorescence protein–expressing plasmid (pmaxGFP; 2 μg DNA) in combination with either BIRC3 siRNA or scrambled siRNA (200 nmol/L). The scrambled siRNA (a control for transfection with BIRC3-targeting siRNA) consisted of 4 pooled nucleotide sequences with at least 4 mismatches to any human gene. Cells were allowed to recover for 4 h prior to infection with R. rickettsii and were challenged with staurosporine 24 and 48 h after infection. Flow cytometric analysis after annexin V binding was then conducted to measure the extent of apoptosis. Data are presented as the mean (±SE) from 3 independent observations. The SE for the experimental condition represented by the last column in the 24 h portion of the graph is negligible

View larger version:

• In this window
• In a new window

• Download as PowerPoint Slide

Table 1

List of pro- and anti-apoptotis genes that display regulation of expression in Rickettsia rickettsii–infected versus R. rickettsii–uninfected human microvascular endothelial cells after staurosporine challenge

Discussion

Activation or prevention of host cell death is an important determinant for the outcome of an infection. Manipulation of host cell responses, however, has significant additional relevance for α-proteobacteria belonging to the genus Rickettsia which includes obligate intracellular pathogenic species with tropism for vascular endothelium of their hosts. Because of their unique intracellular niche, rickettsiae are dependent on a viable host cell for their growth and replication. Evidence suggesting that rickettsiae are capable of manipulating host cell apoptosis emerged from findings that inhibition of infection-induced activation of nuclear factor–κB in host cells resulted in apoptotic death [20]. Subsequent studies further revealed that nuclear factor–κB activation is essential for maintaining mitochondrial integrity, ensuring the balance between pro- and anti-apoptotic proteins, and preventing activation of the caspase cascade during R. rickettsii infection of human endothelial cells in vitro [21, 22]. Together, these findings led us to hypothesize that host cells infected with R. rickettsii tend to develop the capacity to resist the effects of inducers of apoptosis, at least early in the establishment of infection

Staurosporine, a known inducer of apoptosis in a variety of human cell types, has been employed to define apoptosis resistance in cells infected with various pathogenic bacteria, including Chlamydia, Coxiella and Neisseria species [6, 17, 23, 24]. Our initial experiments confirmed the susceptibility of HMECs to the apoptosis-inducing effect of staurosporine (figure 3 and 4). The level of annexin V positivity (an early marker of apoptosis) in HMECs infected with R. rickettsii for 24 and 48 h and subsequently challenged with staurosporine was significantly lower than that in the corresponding uninfected controls, indicating that R. rickettsii infection protects HMECs against staurosporine-induced apoptosis

Because analysis of the transcriptome represents a highly dynamic approach that yields critical information about cellular changes in gene expression profiles in response to infection, we performed a screening of pro- and anti-apoptotic genes that were differentially expressed in HMECs during R. rickettsii infection and after staurosporine challenge. A total of 14 genes were significantly up-regulated, of which 8 (TRAF1, BNIP2, BCL2L1, TRAF3, BIRC2, BNIP3L, AKT1 and BIRC5) are known apoptosis suppressors, whereas the other 6 are known to promote apoptosis (BOK, BCL2L13, DAPK2, TP53, ABL1 and BAK1). Among these, BOK and DAPK2 had high levels of transcription, whereas the 4 remaining genes has only slight increases in transcription over the 2.5-fold change threshold. R. rickettsii–induced upregulation of TRAF1, BNIP2, BCL2L1, TRAF3, BIRC2, BIRC3, BNIP3L, AKT1 and BIRC5 transcription may play an important role in the suppression of apoptosis, although modulation of the host cell gene expression patterns for both pro- and anti-apoptotic proteins is clearly evident. The anti-apoptotic activity of R. rickettsii–infected host cells against staurosporine appears to be sufficient to overcome the combined effects of the pro-apoptotic signals generated both during infection and as a result of treatment with an apoptosis inducer. This tendency to resist apoptosis is consistent with similar cellular responses observed during host cell infection with other obligate intracellular infectious agents, such as several viruses and Chlamydia species). Interestingly, other pathogens that are known to elicit an anti-apoptotic phenotype have also been demonstrated to induce the expression of various pro-apoptotic genes [9, 17, 25, 26]

The inhibitor-of-apoptosis family of proteins, an important group of anti-apoptotic proteins, includes endogenous repressors of the terminal caspase cascade. Among the well-characterized inhibitor-of-apoptosis proteins, cIAP1 and cIAP2 are known to be regulated by nuclear factor–κB [27]. Our results yield cumulative evidence for the increased expression of cIAP2 mRNA and protein in R. rickettsii–infected HMECs. Further, higher levels of cIAP2 are maintained in infected cells that are undergoing staurosporine challenge, implicating cIAP2 as potentially involved in infection-induced cytoprotection. Intriguingly, siRNA-mediated downregulation of cIAP2 had no effect on protection against staurosporine-induced apoptosis at 24 and 48 h after infection (figure 5 B). Thus, cIAP2 downregulation by siRNA is not able to revert resistance to apoptosis in cells acutely infected with R. rickettsii. Host cells infected with obligately intracellular Chlamydia trachomatis and Chlamydophila pneumoniae have been shown to display profound resistance to an anti-Fas antibody, granzyme B–perforin, tumor necrosis factor α, and stress-mediated apoptosis [28, 29]. In C. trachomatis–infected HeLa cells, upregulation of cIAP2 and recruitment and stabilization of other inhibitor-of-apoptosis proteins, such as cIAP1 and xIAP, led to apoptosis resistance [19]. In contrast, a similar study using mouse embryonic fibroblasts deficient in different inhibitor-of-apoptosis proteins suggests that, although cIAP1, cIAP2, and xIAP are dispensable for apoptosis resistance that is afforded by chlamydial infection, proteolytic degradation of pro-apoptotic BH3 proteins is of critical importance [30]. A potential caveat in the interpretation of the intracellular effects of inhibitor-of-apoptosis proteins is that only xIAP has caspase-inhibitory activity, whereas cIAP1 and cIAP2 may not function as effective inhibitors of effector caspases as predicted [31]

In conclusion, although our data clearly demonstrate the upregulation of cIAP2 during R. rickettsii infection of HMECs, the resistance of infected cells to staurosporine-induced apoptosis is apparently not dependent on the anti-apoptotic effects of cIAP2 and is likely mediated by complex, multi-threaded mechanisms. In this context, it is important to consider that cytotoxic T lymphocyte–mediated clearance of rickettsial infection depends on the granule exocytosis pathway, which involves the action of granzymes and perforin [32]. Because protection against apoptotic stimuli likely represents a means of immune evasion, which in turn may influence the progression of disease, further studies with use of organ and/or vascular bed–specific microvascular endothelium and potent immunologic signals such as tumor necrosis factor α, Fas ligand–binding protein, and granzyme B–perforin as inducers of apoptosis are warranted