Introduction

Top Abstract Introduction Materials and Methods Results Discussion References

A number of inflammatory lung diseases are presumed to be T-cell mediated, and are characterized by the accumulation of T cells in the alveolar space or in the interstitium (1). The role(s) that CD8⁺ T cells play in the pathogenesis of interstitial lung disease is unclear, as is the nature of the effector function(s) expressed at the sites of inflammatory activity. CD8⁺ T cells express potent cytotoxic activity, which appears to function primarily in the clearance of virus infection (5). In vitro analysis of antiviral activity of CD8⁺ T cells strongly suggests that cytolytic activity is both necessary and sufficient to effect the lysis of virus- infected target cells (12). Several groups have shown that CD8⁺ T cells utilize two primary molecular mechanisms to induce target cell death: the perforin/granzyme-mediated pathway of cytolysis and the Fas ligand (FasL)- Fas (CD95) pathway of induction of target-cell apoptosis. In the absence of both of these mechanisms, there is no cytolysis of conventional target cells, which are generally of hematopoetic origin (12, 13, 15). The molecular basis of T-cell-mediated cytolysis of respiratory cells has not been directly addressed, but the issue is an important one, since the respiratory tract is a frequent target of virus infection.

Initial interest in Fas focused on its role in maintenance of the lymphoid compartment (20). However, over the past several years, there has been intensified interest in extralymphoid expression of Fas and its potential participation in tissue inflammation and injury. In particular, numerous reports have suggested a role for Fas in chronic inflammatory liver disease (26), thyroiditis (34, 35), experimental diabetes (36), and lung disease. The literature on participation of Fas in lung disease is replete with demonstrations of expression of Fas in clinical (37) and experimental specimens (41, 42). In addition, several intriguing studies have involved the experimental induction of pneumonitis (43) and type II cell apoptosis by ligation of Fas by agonist antibodies (44, 45). There is so far no direct evidence that Fas-mediated apoptosis plays a role in lung injury mediated by FasL-expressing T cells. Several lines of evidence have, however, suggested that tumor necrosis factor (TNF)- has a primary or secondary role in pulmonary inflammation (46), though not specifically in injury mediated directly by T cells. TNF- is a minor contributor to T-cell effector function, and has not been shown to participate in acute cytolysis of conventional target cells (49). In the present study, we took advantage of a novel in vitro assay to present evidence that alveolar epithelial-derived cells are susceptible to CD8⁺ T-cell-triggered apoptosis induced by TNF-, but are resistant to apoptotic signals triggered by FasL. We also showed that perforin-independent, T-cell-mediated cytotoxicity of alveolar epithelial-derived cells is mediated entirely and exclusively by TNF-. Furthermore, using a transgenic model of T-cell-mediated lung injury (50, 51), we showed in vivo that Fas-deficient animals are not protected against injury by activated CD8⁺ T cells, even when the T cells are derived from perforin-deficient animals, whereas pretreatment of T cells with antibody to TNF- abolishes lung injury induced by perforin-deficient T cells.

	Materials and Methods

Top Abstract Introduction Materials and Methods Results Discussion References

Alveolar Epithelial-Derived Cells

MLE-15 cells (52) were transfected with the class I major histocompatibility complex (MHC) molecule, K^d (construct provided by Dr. Michael Bevan, University of Washington, Seattle, WA). The transfectants were stained with SF1.1 (anti-K^d; American Type Culture Collection [ATCC], Rockville, MD) and TIB95 (anti-K^q, the native haplotype of the parental line; ATCC), and sorted by flow cytometry according to their level of K^d expression with a FACScan instrument using CellQuest software (Becton-Dickinson, Palo Alto, CA). The cells were then cloned by limiting dilution, and were passaged twice a week.

T-Cell Cytotoxicity Assay

The CD8⁺ T-cell clones used in our experiments were generated by limiting dilution as previously described (53), from either wild-type or perforin-deficient mice (18). The clones were restimulated weekly in vitro with irradiated syngeneic splenocytes that were infected with A/Japan/57 influenza virus, and were placed into fresh Iscove's complete medium supplemented with 10 IU/ml of interleukin (IL)-2. On Day 5 after in vitro stimulation, T cells were tested with a ⁵¹Cr release assay for cytolytic activity against target cells infected with A/Japan/57 or against target cells loaded with 10⁹ M synthetic peptide representing the 210-219 epitope of the A/Japan/57 hemagglutinin (HA) (50). The target cells used were MLE-K^d cells and L121⁺ cells, the latter of which are transfected with a Fas- overexpression construct (54). In some experiments, the anti-Fas monoclonal antibody MFL3 was added to wells at a final concentration of 5 µg/ml, or anti-TNF- antiserum (IP400, Genzyme, Boston, MA) was added at a final dilution of 1:100. Cytotoxicity assays were conducted for 6 h in 96-well plates with a final volume of 0.2 ml/well of target and effector cells in culture medium, after which 0.1 ml was harvested and counted on a gamma counter (Isomedic; ICN Biomedicals, Inc., Costa Mesa, CA). Percent specific ⁵¹Cr release was calculated according to the formula: (test cpm  spontaneous release cpm)/(total cpm  spontaneous cpm) × 100. Spontaneous release from targets incubated with medium alone was always less than 10%. Specific lysis values represent the mean percent specific ⁵¹Cr release from four replicate wells.

DNA Fragmentation Assay

MLE-K^d cells were labeled with tritiated thymidine for 6 h and were then incubated for 24 h with 0.01 µg/ml Jo-2 antibody (PharMingen, San Diego, CA) or with an isotype- matched control hamster antibody (MFL3), with or without 20 µg/ml protein G (Sigma, St. Louis, MO). The cells were then lysed and harvested. L1210⁺ and L1210 cells, which are transfected with a Fas-overexpression or antisense construct, respectively (54), were similarly treated and were used as controls. In some experiments, soluble TNF- (Genzyme, Boston, MA) was added to MLE-K^d cells at varying concentrations for 6 h prior to harvesting (5 × 10⁷ U/mg). Percent DNA fragmentation was calculated as described (55) according to the formula: (total cpm [immediately after labeling]  test cpm)/total cpm × 100. All values represent the mean of four replicate wells.

Primary Type II Pneumocyte DNA Fragmentation

Primary alveolar type II cells were prepared according to a modification of a previously published method (56). Briefly, BALB/c mice were anesthetized and exsanguinated by severing the inferior vena cava and left renal artery. The tracheae were exposed and cannulated, and the lungs were perfused with 10-20 ml sterile saline via the pulmonary artery until visually free of blood. Dispase (Collaborative Research, Inc., Bedford, MA) was instilled into the lungs via the tracheal catheter, and was followed by 1% low melting agarose, warmed to 45°C. The lungs were immediately covered with ice and incubated for 2 min to gel the agarose. Lungs were then dissected free of the animal, put in a culture tube containing an additional 1 ml of Dispase, and incubated for 45 min at room temperature. Lungs were then transferred to a culture dish containing deoxyribonuclease (DNAse) I (Sigma, St. Louis, MO) and the tissue was gently teased away from the airways. The cell suspension was successively filtered and then pelleted. Crude cell suspensions were added to culture dishes coated with anti-CD45 and anti-CD32 antibodies (PharMingen) and incubated for 1-2 h. Plates were removed from the incubator and gently "panned" to free settled type II cells, which were resuspended in Dulbecco's modified Eagle's medium with 10% fetal bovine serum. Purity of the type II cell preparations used for these studies was 92 ± 1% according to morphologic criteria (56), and cell viability determined by Trypan blue exclusion after overnight incubation was > 60%. (Purity for more than 100 separate preparations was consistently > 90%.) Cells were plated in 96-well plates and labeled overnight with tritiated thymidine. The wells were washed to remove unincorporated label and nonadherent cells, and the remaining cells were then incubated for 6 h with TNF-. The percent DNA fragmentation was calculated as described (55), and represents the mean of four replicate wells.

Adoptive Transfer

SPC-HA transgenic mice (H-2^d) expressing the A/Japan/ 305/57 influenza HA under the transcriptional control of the surfactant protein-C (SP-C) promoter were used in these studies (50). Transgene-negative littermates were used as control recipients. Animals subjected to adoptive transfer were used at 10-12 wk of age (18-22 g). Fas-deficient (lpr) animals in the H-2^d haplotype (generously provided by Dr. Philip Cohen of the University of North Carolina, Chapel Hill, NC) were bred with the SPC-HA transgenic mice for use in some experiments. The offspring were screened for the HA transgene as described (50), and for the lpr mutation with a three-primer polymerase chain reaction (PCR) method as described (57). The CD8⁺ cytolytic T-cell clone 40-2 was restimulated in vitro with irradiated syngeneic splenocytes that were infected with A/Japan/57 influenza virus. On Day 5 after stimulation, clones were separated from stimulators by density gradient centrifugation and were injected via the tail vein into appropriate recipient animals. In some instances, T cells were preincubated with anti-TNF- antibody (1:100, IP400; Genzyme) for 30 min at 4°C and were then washed twice before adoptive transfer. In order to quantitatively assess the physiologic dysfunction induced by T-cell transfer, we measured carbon monoxide (CO) uptake and inspiratory capacity (IC) 4 d after cell transfer, as previously described (51). Briefly, after tracheal intubation, the IC of each animal was estimated by inflation with air to a pressure equivalent to 20 cm H₂O (58, 59). The lungs were then inflated with a volume equal to the IC of a gas mixture containing 0.3% CO, 0.3% methane, 20% oxygen, and the balance consisting of nitrogen. The gas was allowed to dwell for a 10 s breathhold, after which one half of the sample was drawn off and discarded as dead space and the other half was drawn off as an alveolar sample (60). Three measurements were made for each animal, and the CO uptake was determined as the ratio of CO loss to methane loss (59, 61).

Statistical Analysis

Significant differences were determined with Student's unpaired t test or through one-way analysis of variance (ANOVA).

	Results

Top Abstract Introduction Materials and Methods Results Discussion References

Alveolar Epithelial-Derived Cells Are Insensitive to FasL Expressed on CD8⁺ T Cells

In order to explore the molecular basis of CD8⁺ T-cell- mediated injury to alveolar epithelium, we developed an in vitro system for measuring T-cell cytolysis of alveolar epithelial-derived cells. The MLE-15 cell line, which maintains phenotypic characteristics of type II alveolar epithelial cells (52), was transfected with the class I MHC molecule K^d (the haplotype of the transgenic animals and the T-cell clones used in these studies). The level of expression of K^d was comparable to the expression level of the endogenous MHC class I molecule, K^q, as determined by flow cytometry (not shown). We have previously shown that CD8⁺ T cells recognize and lyse transfectant MLE-K^d cells if the target cell is infected with the appropriate strain of influenza virus or if exogenous peptide loading of K^d occurs with the appropriate K^d-restricted epitope, but that no such recognition of untransfected MLE-15 cells occurs (51). As shown in Figure 1, the recognition and cytolysis of peptide-loaded MLE-K^d cells by the CD8⁺ cytolytic T-cell clone 40-2 was completely unaffected by the anti-FasL antibody MFL3 (at 5 µg/ml). In order to demonstrate that this antibody (5 µg/ml) was capable of completely blocking FasL-dependent lysis by clone 40-2, we used the paired target cells L1210⁺ and L1210, which are transfected with sense and antisense Fas overexpression constructs, respectively. These cells are (not surprisingly) either very sensitive or insensitive, respectively, to FasL-triggered apoptosis (19, 54). Cao and colleagues, using the identical T-cell clone 40-2, have previously shown that T-cell recognition of foreign antigen triggers both the perforin and the FasL/ Fas pathways of cytotoxicity. In contrast, recognition of a "self antigen" triggers cytolysis mediated specifically and exclusively by FasL, with no utilization of the otherwise intact perforin-mediated pathway of cytotoxicity (54). Therefore, this target cell pair used in our study provides a very useful readout of perforin-independent cytotoxicity. As shown in Figure 2, there was minimal inhibition by the anti-FasL antibody MFL3 of recognition and cytolysis of target cells (L1210⁺ or L1210) by 40-2 cells in the presence of the HA210-219 peptide, indicating significant residual perforin-mediated lysis. However, in the absence of peptide antigen (i.e., in the presence of self antigen only), there was no cytolysis of L1210 cells, and yet there was significant lysis of L1210⁺ cells, indicating perforin-independent (FasL-mediated) cytolysis. Furthermore, the cytotoxicity of 40-2 cells toward L1210⁺ cells in the absence of peptide was completely inhibited by the MFL3 antibody at 5 µg/ml. Therefore, this concentration of the antibody was adequate to completely inhibit the FasL-mediated cytotoxic activity of clone 40-2. Importantly, as shown in Figure 1, blockade of FasL-mediated activity by antibody did not inhibit the lysis of MLE-K^d cells, indicating the insensitivity of these alveolar type II cells to Fas-dependent lysis.

View larger version (12K): [in this window] [in a new window]   Figure 1.   Results of CD8+ T-cell cytotoxicity assay, using clone 40-2 as the effector cells and MLE-Kd cells as targets. T cells were tested with a 51Cr release assay for cytolytic activity against target cells infected with influenza virus A/Japan/57 or against target cells loaded with 109 M synthetic peptide representing the HA210-219 epitope. Clone 40-2 efficiently recognized and lysed A/Japan/57-infected, as well as peptide-pulsed, MLE-Kd cells (P < 0.01). There was no cytotoxicity in the absence of peptide. Anti-FasL antibody (MFL3; 5 µg/ml), had no effect, though this concentration completely inhibited FasL activity on L1210+ cells, which are exquisitely sensitive to FasL (see Figure 2). Cells in the assay were incubated for 6 h prior to harvesting (effector-to-target-cell ratio of 20:1). Specific lysis values represent the mean percent specific 51Cr release from four replicate wells (SEM depicted by error bars). Data shown are representative of three experiments with similar results.

View larger version (26K): [in this window] [in a new window]   Figure 2.   Results of CD8+ T-cell cytotoxicity assay, using clone 40-2 as the effector cells and L1210+ cells as targets. T cells were tested with a 51Cr release assay for cytolytic activity against target cells loaded with 109 M synthetic peptide representing the HA210-219 epitope. Clone 40-2 efficiently recognized and lysed L1210+ activity cells in the presence of antigenic peptide, and no inhibition was demonstrable upon treatment with anti-FasL antibody, thus indicating perforin-mediated cytolysis (the dominant effect). In the absence of "foreign" antigen (i.e., in the presence of self antigen alone), 40-2 cells still recognized and lysed L1210+ cells, albeit to a lesser degree than in the presence of foreign antigen (indicating partial expression of effector function; i.e., FasL without perforin). In the presence of anti-FasL antibody (MFL3; 5 µg/ml), there was complete inhibition of cytotoxic activity in the absence of foreign antigen (P < 0.01), confirming that this effector activity was entirely FasL-mediated, and that this concentration of antibody completely inhibits FasL activity in the presence of an exquisitely FasL-sensitive cell. Cells in the assay were incubated for 5 h prior to harvesting (effector-to-target-cell ratio of 20:1). Specific lysis values represent the mean percent specific 51Cr release from four replicate wells (SEM depicted by error bars). Data shown are representative of three experiments with similar results.

Perforin-Independent T-Cell Cytotoxicity of Alveolar Epithelial-Derived Cells Is Mediated Entirely by TNF-

In order to further explore the mechanisms involved in the T-cell-mediated cytotoxicity of alveolar epithelial-derived cells, we used an alternative strategy, utilizing CD8⁺ T-cell clones from perforin-deficient animals (18). Bulk HA-specific CD8⁺ T cells were generated from perforin-deficient animals (bred into the H-2^d haplotype) that were immunized with A/Japan/57 influenza virus. We then produced individual CD8⁺ T-cell clones from the bulk cultures by limiting dilution (53). These T cells exhibited no cytolysis of Fas-deficient L1210 cells, but did exhibit lysis of peptide-loaded L1210⁺ cells, and this perforin-independent cytolytic activity was completely inhibited by antibody to FasL (MFL3; not shown). We therefore tested the sensitivity of MLE-K^d cells to perforin-independent cytolysis, using perforin-deficient T cells. As shown in Figure 3, the CD8⁺ T-cell clone PKO-GV effected cytotoxicity at levels similar to those of wild-type T cells in the presence of antigenic peptide, but showed no effect of MFL3 antibody. Interestingly, cytolysis of MLE-K^d cells by PKO-GV cells was completely inhibited by antibody to TNF-, which has been shown to be expressed on CD8⁺ T cells, predominantly as a membrane-bound molecule (49). Several other, independently evaluated perforin-deficient CD8⁺ T-cell clones exhibited the same phenotype (not shown). These data indicate that FasL-insensitive MLE-K^d cells are unusually sensitive to TNF--mediated cytotoxicity by CD8⁺ T cells. The sensitivity of MLE-K^d cells to TNF- is a very uncommon phenotype: it is very difficult to demonstrate this phenomenon in other target cells without the addition of inhibitors of transcription or protein synthesis (62) and/or without testing for cytotoxicity over a considerably longer assay period, usually 24 h or more (19, 49). Furthermore, the inhibition of perforin-independent cytolysis by anti-TNF- antibody confirms that MLE-K^d cells are resistant to apoptotic signals delivered by FasL expressed on CD8⁺ T cells.

View larger version (12K): [in this window] [in a new window]   Figure 3.   Results of CD8+ T-cell cytotoxicity assay, using perforin-deficient clone PKO-GV as the effector cells, and MLE-Kd cells as targets. T cells were tested with a 51Cr release assay for cytolytic activity against target cells treated with 109 M synthetic peptide representing the HA210-219 epitope. Clone PKO-GV efficiently recognized and lysed peptide-pulsed, MLE-Kd cells, demonstrating the efficacy of perforin-independent cytotoxic activity against alveolar epithelial-derived cells. There was no cytotoxicity in the absence of peptide. Anti-FasL antibody (5 µg/ml) had no effect, though this concentration completely inhibited FasL activity on L1210+ cells, which are exquisitely sensitive to FasL (see Figure 2). However, a polyclonal anti-TNF- antibody (1:100 final dilution) completely blocked all cytotoxicity of PKO-GV cells against MLE-Kd cells (P < 0.05), indicating that the perforin-independent cytolysis affected by CD8+ T cells was entirely TNF--mediated, and that MLE-Kd cells exhibit unusual sensitivity to TNF-. Cells in the assay were incubated for 6 h prior to harvesting (effector-to-target-cell ratio of 20:1). Specific lysis values represent the mean percent specific 51Cr release from four replicate wells (SEM depicted by error bars). Data shown are representative of three experiments with similar results.

In order to rule out the possibility that FasL-induced apoptosis of alveolar epithelial-derived cells plays a role in "bystander" injury, as has been suggested with other cell types (65, 66), we investigated the relationship between antigen recognition of and cytotoxicity to neighboring cells, using a transfectant cell line (BHA) as antigen-presenting cells (P815 cells expressing full length A/Japan/57 HA). Like most cells that have been studied with regard to apoptosis (62), P815 (and BHA) cells are resistant to TNF-mediated apoptosis in the absence of protein synthesis inhibitors, and anti-FasL antibody completely blocks all cytolysis of these target cells by perforin-deficient T cells (not shown). Figure 4 shows the results of a representative experiment in which BHA cells (unlabeled) were used as the antigen-presenting cells and ⁵¹Cr-labeled MLE-K^d cells (without virus infection or peptide loading) were used as bystander cells. Recognition of "cold" BHA cells by PKO-GV cells resulted in cytolysis of "hot" MLE-K^d cells, but in minimal lysis in the absence of the BHA cells. The level of cytolysis was almost identical to that produced by peptide treatment of the MLE-K^d cells (in the absence of BHA cells). This phenomenon is clearly the result of a specific recognition event, since untransfected (parental) P815 cells do not activate the T cells to effect bystander cell lysis. These results also clearly indicate that the capacity to induce bystander-cell cytotoxicity is not a constitutive phenomenon (i.e., after in vitro activation), since specific antigen recognition in the assay is required. Importantly, the cytolysis of bystander MLE-K^d cells induced by BHA-triggered PKO-GV cells was completely inhibited by anti-TNF- antibody.

View larger version (12K): [in this window] [in a new window]   Figure 4.   Bystander-cell cytotoxicity affected by perforin-deficient CD8+ T cell clone PKO-GV. MLE-Kd cells were 51Cr- labeled and then either treated with 109 M synthetic peptide or mixed in a 1:1 ratio with unlabeled BHA cells (P815 cells transfected with A/Japan/57 influenza virus HA). The T cells were triggered by recognition of the HA210-219 epitope expressed by the unlabeled (transfected) BHA cells, and bystander cytolysis of the labeled MLE-Kd cells (without peptide) occurred. No such triggering occurred upon addition of unlabeled parental (untransfected) P815 cells (P < 0.05). In the absence of either peptide or BHA cells, no lysis of MLE-Kd cells was evident. Importantly, all bystander cytolysis was inhibited by anti-TNF- antibody (P < 0.05). Specific lysis values represent the mean percent specific 51Cr release from four replicate wells (SEM depicted by error bars). Data shown are representative of three experiments with similar results.

We further investigated TNF-mediated bystander injury with the wild-type CD8⁺ T-cell clone 40-2 as the effector cells and with A/Japan/57 influenza-infected MLE-K^d cells as the triggering cells. In these experiments, the bystander targets were either ⁵¹Cr-labeled MLE-K^d cells or ⁵¹Cr-labeled parental MLE cells (both without virus infection or peptide loading). As shown in Figure 5, there was no significant cytolysis when 40-2 cells were incubated with either the labeled uninfected MLE-K^d cells or with labeled uninfected parental (untransfected) MLE-15 cells, and there was no effect of anti-TNF antibody in either circumstance. However, the addition of unlabeled influenza-virus-infected MLE-K^d cells (at a 1:1 ratio to labeled uninfected bystanders cells) resulted in significant bystander cytotoxicity, which was independent of K^d expression by the bystander cell. As in the previous experiment, bystander-cell cytolysis was completely inhibited by anti-TNF antibody. These results suggest that this effect represents true bystander injury (i.e., cytotoxicity unrelated to peptide transfer and cross-presentation). These data also suggest that perforin does not participate in bystander injury to alveolar epithelial-derived cells, in contrast with the primary role played by perforin in direct cytolysis of virus-infected MLE-K^d cells by wild-type T cells (data not shown).

View larger version (25K): [in this window] [in a new window]   Figure 5.   Bystander-cell cytotoxicity affected by wild-type CD8+ T cell clone 40-2. MLE-Kd cells or parental (untransfected) MLE-15 cells were 51Cr-labeled and plated with 40-2 cells at an effector-to-target-cell ratio of 10:1. Addition of unlabelled A/Japan/57 influenza-virus-infected MLE-Kd cells (1:1; unlabeled: labeled) triggered significant bystander-cell cytotoxicity of uninfected (labeled) MLE-Kd cells and parental (untransfected) MLE cells (P < 0.05). The addition of anti-TNF antibody completely ablated the bystander cytolysis (P < 0.05). Specific lysis values represent the mean percent specific 51Cr release from four replicate wells (SEM depicted by error bars). Data shown are representative of multiple experiments with similar results.

Alveolar Epithelial-Derived Cells Express Functional Fas

Although others have demonstrated Fas expression by the parental (MLE-15) cell line (45, 67, 68), our results as described earlier could be explained by the loss of Fas expression that might have occurred during transfection, selection, and cloning of the MLE-K^d cell. Other investigators have reported induction of apoptosis of alveolar epithelial cells, including MLE-15 cells, with the agonistic anti-Fas antibody Jo-2 (44, 45). In order to confirm that MLE-K^d cells express functional Fas, we investigated whether Jo-2 could induce apoptosis in MLE-K^d cells. We were unable to detect cytolysis of MLE-K^d cells induced by Jo-2 antibody in standard 6-h ⁵¹Cr-release assays, although this was not surprising in view of our observations with other cell lines (even those that are functionally Fas-sensitive) in standard ⁵¹Cr-release assays (unpublished results). Therefore we took advantage of a more sensitive DNA fragmentation assay (55) that has been used previously to measure Fas-mediated cell death (54). This assay has the advantage of detecting earlier events in apoptosis, occurring before the membrane disruption necessary for ⁵¹Cr-release (55). As shown in Figure 6, MLE-K^d cells exhibited significant DNA fragmentation by 6 h after crosslinking with the Jo-2 antibody. This effect in MLE-K^d cells was not enhanced by the addition of protein G, even though significant enhancement of DNA fragmentation occurred in L1210⁺ cells upon addition of protein G. Spontaneous apoptosis (medium alone) was less than 5% by 6 h. An isotype-matched control hamster antibody (MFL3) had no effect, at a variety of concentrations, with or without protein G (not shown). Similarly, protein G alone had no effect. These data strongly suggest that Fas is present and capable of triggering the apoptotic cascade in MLE-K^d cells, although the functional significance of this remains unclear.

View larger version (33K): [in this window] [in a new window]   Figure 6.   DNA fragmentation assay done with MLE-Kd, L1210+, and L1210 cells. Cells were labeled with tritiated thymidine and treated with Jo-2 antibody with or without protein G. The cells were incubated for 24 h, lysed, and harvested. Jo-2 efficiently triggered apoptosis in MLE-Kd cells (P < 0.01), with little additional effect of protein G. L1210+ cells were also triggered to undergo apoptosis upon treatment with Jo-2, and this effect was enhanced by the addition of protein G (P < 0.05). There was no effect of Jo-2 on L1210 cells, indicating that the proapoptotic effect was specifically Fas-dependent. Spontaneous apoptosis (medium alone) was less than 5% by 6 h. An isotype-matched control hamster antibody (MFL3) had no effect at a variety of concentrations, with or without protein G. Similarly, protein G alone had no effect. Data shown are representative of three experiments with similar results.

MLE-K^d Cells and Primary Type II Alveolar Epithelial Cells Are Susceptible to Apoptosis Induced by Soluble TNF-

Although anti-TNF antibody completely inhibited the cytolysis of MLE-K^d cells by perforin-deficient CD8⁺ T cells, the data do not reveal whether this was due to membrane-bound or secreted TNF-. To investigate whether MLE-K^d cells are susceptible to apoptosis triggered by soluble TNF-, we performed DNA fragmentation assays after 6 h of treatment with varying doses of TNF-. As shown in Figure 7A, soluble TNF- induced significant DNA fragmentation in MLE-K^d cells as compared with media controls, and the effect reached a plateau at 1,000 U/ml. In order to confirm that the susceptibility of the MLE-K^d cells to cytotoxicity mediated by soluble TNF is not unique to this type II alveolar cell line, but rather reflects an intrinsic property of untransformed lung epithelial cells, we isolated primary murine type II pneumocytes for use in correlative studies. As shown in Figure 7B, soluble TNF- (1,000 U/ml) induced significant apoptosis in primary type II cells as compared with controls (P < 0.05). This suggests that the TNF-sensitivity of MLE-K^d cells is not a unique feature of this particular transformed cell line.

View larger version (14K): [in this window] [in a new window]   View larger version (11K): [in this window] [in a new window]   Figure 7.   (A) Effect of soluble TNF- on DNA fragmentation in MLE-Kd cells. The cells were labeled overnight with tritiated thymidine, and were then washed and incubated for 6 h with varying doses of soluble TNF-. Significant DNA fragmentation was induced at all TNF- concentrations examined (P < 0.05), although a plateau effect occurred at 1,000 U/ml. Data shown are representative of three experiments with similar results. (B) Effect of soluble TNF- on DNA fragmentation in primary type II pneumocytes. After isolation, cells were labeled overnight with tritiated thymidine, and excess label and nonadherent cells were washed out. The adherent cells were then incubated for 6 h with soluble TNF- (1,000 U/ml). Significant DNA fragmentation was induced by soluble TNF-, as compared with control levels (P < 0.05). Data shown are representative of two experiments with similar results.

Perforin-Deficient CD8⁺ T Cells Induce Significant Lung Injury In Vivo That Is Independent of Fas and Dependent upon TNF-

We have previously reported that wild-type, influenza-virus- specific CD8⁺ T cells induce significant lung injury upon being adoptively transferred into transgenic mice expressing influenza HA in their type II alveolar epithelial cells (50, 51). This injury is characterized by significant loss of alveolar epithelial cells, as well as by restrictive physiology, significant impairment of gas exchange, and death (51). We also demonstrated that the impairment in gas exchange, as determined by CO uptake, correlated especially strongly with the histologic and clinical abnormalities observed in these mice (51). In order to determine the relative contribution of the perforin-independent mediators FasL and TNF- to T-cell cytotoxicity and injury in vivo, we performed adoptive transfer of the perforin-deficient (FasL⁺, TNF-⁺), HA-specific CD8⁺ T-cell clone PKO-GV into three recipient groups: wild-type HA⁺ mice, Fas-deficient/HA⁺ mice, and HA control littermates. As shown in Figure 8, there was a significant decrement in CO uptake in HA⁺ transgenic recipients as compared with their HA littermates (which was similar to that measured in animals that had not received cells; data not shown). Interestingly, there was no abrogation of injury by PKO-GV T cells in the absence of Fas expression in the Fas-deficient recipients (lpr/HA⁺ mice). However, preincubation of PKO-GV T cells with anti-TNF antibody resulted in complete abrogation of lung injury, further suggesting the Fas-independent nature of this (perforin-independent) CD8⁺ T-cell-mediated injury process. Histologically, the lung sections harvested at Day 4 from lpr/HA⁺ animals that received anti-TNF-antibody-treated T cells were indistinguishable from the lungs of HA recipients (data not shown); there was minimal perivascular and peribronchiolar cuffing without significant alveolar infiltration, resembling that previously described (50, 51). No difference was observed histologically between the HA⁺ and the lpr/HA⁺ recipients of PKO-GV T cells, and both showed significant infiltration of the alveolar walls and spaces, resembling that previously described (50, 51). Prior to T-cell transfer, lpr/HA⁺ mice were indistinguishable from normal (HA) mice both histologically and physiologically (data not shown). The 4-d time point was chosen on the basis of previous data (51) as well as from the tempo of progression of clinical illness observed in these animals.

View larger version (13K): [in this window] [in a new window]   Figure 8.   CO uptake in HA+ transgenic mice (expressing influenza HA in type II epithelial cells) after adoptive transfer of HA-specific, perforin-deficient CD8+ T cells. Significant diminution in CO uptake was evident by Day 4 after transfer into HA+ animals, as compared with HA control littermates. HA+ Fas-deficient mice (lpr/HA+) showed an identical effect after transfer. The effect was abolished by preincubation of the T cells with anti-TNF antibody (and washing) prior to transfer. Data shown are representative of three experiments with similar results.

	Discussion

Top Abstract Introduction Materials and Methods Results Discussion References

Many interstitial lung diseases are in part characterized by varying degrees of CD8⁺ T-cell (in addition to other inflammatory cell) infiltration of the alveolar septae, spaces, or both. We have previously shown that antigen-specific CD8⁺ T-cell recognition by a single clonal population and of a single epithelial "autoantigen," is in and of itself sufficient to trigger an inflammatory cascade that results both histologically and physiologically in the clinical picture of interstitial pneumonia (50, 51). Adoptive transfer of 10⁷ T cells resulted in a complete loss of alveolar epithelial cells, resembling the epithelial denuding seen in disease states associated pathologically with diffuse alveolar damage (such as acute interstitial pneumonia). T cells express a range of cytotoxic and noncytotoxic effector functions, which have been intensively investigated in the context of viral infection. CD8⁺ T cells are particularly associated with cytotoxic activity when they recognize class I MHC-restricted peptide epitopes, such as viral antigens. The major mechanisms involved in CD8⁺ T cell cytotoxicity have been elucidated in vitro, and several groups have clearly shown that the perforin/granzyme pathway and the Fas- FasL pathway account for all of these cells' in vitro cytolytic effect as observed in cytotoxicity assays using conventional target cells, which are generally of hematopoetic origin (12, 13, 15). The role of Fas in T-cell suicide and fratricide is well described (20, 69), although Fas expression has recently been reported in a number of nonhematopoetic tissues, such as liver (26) and lung (37). The role of Fas in the expression of immune effector activity (or injury) in these tissues is unclear, although it has been suggested that Fas may play a role in immunity to virus infection (37). Several studies have suggested that Fas may participate in immune injury in chronic inflammation of the liver (28) and lung (38, 40). There is so far no direct evidence that Fas-mediated apoptosis plays a role in lung injury caused specifically by FasL-expressing T cells.

Experimentally, alveolar epithelial-derived cells may be triggered to undergo apoptosis upon ligation of Fas with Jo-2 antibody (44, 45), and intratracheal administration of this antibody in mice results in significant lung injury (43). We have confirmed Fas expression in the lungs of HA⁺ transgenic mouse lines by reverse transcription- PCR, and in the alveolar epithelium of these animals by in situ hybridization (unpublished observations). It was therefore somewhat surprising to observe that the absence of Fas expression in lpr/HA⁺ transgenic mice did not abrogate the injury produced by T-cell transfer, even in the absence of perforin. It was also surprising to observe that MLE-K^d cells were totally insensitive to the effects of FasL expressed by activated T cells, whereas neutralizing antibody to FasL completely inhibited cytotoxicity to Fas-sensitive L1210⁺ target cells by wild-type CD8⁺ T cells of clone 40-2 (without peptide) and by T-cells of the perforin-deficient clone PKO-GV (with peptide). In this experiment, anti-FasL antibody completely eliminated all FasL-mediated cytotoxicity by both 40-2 cells and PKO-GV cells against L1210⁺ cells, but had no effect on the cytolysis of MLE-K^d cells. Interestingly, neutralizing antibody to TNF- completely inhibited all cytotoxicity to MLE-K^d cells by the perforin-deficient T-cell clone PKO-GV. The lung injury observed in vivo in lpr/HA⁺ mice after transfer of perforin-deficient T cells, and its abrogation by pretreatment of such T cells with antibody to TNF-, further support the hypothesis that alveolar epithelial cells may be insensitive to FasL-triggered apoptosis. This in vivo result is consistent with our in vitro observation that perforin-independent type II cell cytotoxicity is mediated exclusively by TNF- (though we cannot formally exclude the possibility that the carryover of a small amount of anti-TNF may also have blocked TNF- secreted in vivo from non-T-cell [host] sources despite the washing of T cells after preincubation). Nevertheless the effect of pretreatment with this antibody, as well as the significant injury observed in Fas-deficient lpr/HA⁺ mice, argue against an important effect of FasL on the injury process.

MLE-K^d cells were also susceptible to cytotoxicity triggered by soluble TNF-, although our data do not address the quantitative importance of this contribution. This observation nevertheless reflects a qualitatively unusual phenotype, which may have important implications in the lung. TNF receptor engagement simultaneously triggers several parallel signaling pathways, including the proapoptotic activation of the caspase cascade, but also the induction of transcription of antiapoptotic gene products mediated primarily by nuclear factor-B (70). In nearly all cells examined, induction of apoptosis by TNF requires treatment with actinomycin D or cycloheximide in order to inhibit the transcription or translation of antiapoptotic gene products, especially in short-term assays (62). In order to confirm this phenotype in untransformed type II pneumocytes, we tested the sensitivity of primary alveolar type II cell isolates for sensitivity to apoptosis triggered by soluble TNF-, and observed a similar effect. As noted earlier, our data do not address the quantitative physiologic significance of this effect, since we used a saturating dose of TNF-. However, the qualitative effect was quite evident. Not unexpectedly, there was a moderate degree of spontaneous apoptosis in primary type II cells that occurred over a period of time after they were removed from their normal anchorage to basement membrane.

It has been reported for other cell types that the level of Fas expression does not correlate with biologic responsiveness as determined by induction of apoptosis with Jo-2 antibody (71). This strongly suggests that mechanisms may exist to tightly regulate programmed cell death in most cell types. Though MLE-K^d cells, upon exposure to FasL, are relatively resistant to Fas-mediated apoptosis, these cells clearly express functional Fas, as indicated by the induction of DNA fragmentation by the Jo-2 antibody in a more sensitive assay. These unknown regulatory mechanisms that tightly control programmed cell death may have important ramifications with regard to inflammatory lung disease, as may the unusual sensitivity of alveolar epithelial-derived cells to TNF- expressed by CD8⁺ T cells. Numerous other factors may contribute to the regulation of alveolar epithelial-cell susceptibility to Fas-mediated apoptosis. For example, other cell types, such as T cells, appear to be resistant to Fas-induced apoptosis during the S phase of the cell cycle (72), and this resistance may correlate with the inability to recruit caspase-8 to the death-inducing signal complex (DISC) early after T cell stimulation (73). There are also suggestions in other cell types that Fas may participate in transducing signals for nonapoptotic cellular functions, such as proliferation in keratinocytes (74) and angiogenesis (75), as is the case with other members of the TNF receptor family (76).

The findings in this study lead to two conclusions: (1) that MLE-K^d cells are relatively resistant to FasL-mediated Fas-dependent cytotoxicity affected by CD8⁺ T cells, in contrast to the effect of TNF-; and (2) that TNF- expressed by CD8⁺ T cells accounts entirely for the perforin-independent cytotoxic effect of MLE-K^d cells. We have also shown that alveolar epithelial-derived cells, as well as primary type II pneumocytes, are sensitive to the apoptotic signal generated by soluble TNF-. Data obtained after adoptive transfer of perforin-deficient CD8⁺ T cells into Fas-deficient (lpr/HA⁺) mice expressing HA exclusively in type II cells in vivo (50) are also consistent with these observations, suggesting that the phenotype of MLE-K^d cells may reflect that of alveolar type II cells in general. The sensitivity of alveolar cells to TNF expressed (or produced) by CD8⁺ T cells may have greater importance in bystander than in direct cytotoxicity, since we have shown that bystander-cell lysis is perforin-independent (and therefore TNF-dependent). The mechanisms underlying the unusual sensitivity of alveolar epithelial cells to TNF remain to be elucidated, but the observations described earlier raise the possibility that although perforin-mediated apoptosis may represent a primary antiviral effector mechanism of T cells, TNF- may play a role in the lung injury that occurs during (and possibly after) a viral respiratory infection.

There is a further potentially important consequence of the preferential sensitivity of alveolar epithelial cells to TNF- as compared with FasL. In contrast to ligation of Fas, ligation of TNF receptors results in activation of proinflammatory signaling, leading to transcription of a host of cytokines and chemokines (62). For example, human alveolar epithelial cell lines have been reported to express IL-8, a potent neutrophil chemoattractant, upon exposure to TNF- (48). This raises the possibility that alveolar epithelial cells may be predisposed to nonspecific amplification of inflammatory processes mediated by CD8⁺ T cells, as we have observed in vivo in our model of T-cell-mediated lung injury (51). We previously demonstrated that from 4 to 5 d after the initial T-cell-mediated insult to alveolar cells, the inflammatory infiltrate consists primarily of host macrophages. These cells are also capable of producing TNF- (as well as other cytokines), which may serve to further amplify the inflammatory cascade. Detailed investigation will be required to dissect the mechanisms controlling the amplification of inflammation in the lung, which may be critical in understanding the pathogenesis of interstitial lung diseases.