Introduction

Top Abstract Introduction Materials and Methods Results Discussion References

Sarcoidosis is a multisystem disorder of unknown etiology characterized by mononuclear phagocyte and T-cell infiltration of involved organs with granuloma formation (1). The lung is the most commonly affected organ, and the analysis of bronchoalveolar lavage (BAL) specimens has significantly contributed to our current understanding of this disease, with activated alveolar macrophages (AM) and alveolar T cells (AT) considered the major effector cells driving disease progression (2). The current model of sarcoid pathogenesis envisages that an unknown major histocompatibility complex class II bound antigen is presented to the T-cell receptor (6) and that following antigenic triggering AT produce T helper (Th)1-type lymphokines interleukin (IL)-2 and interferon (IFN)- (7, 8). AM are also activated during this process, producing cytokines (e.g., tumor necrosis factor [TNF]-, IL-1, and IL-6) and chemokines that amplify the inflammatory process and recruit monocytes from the periphery (9). Studies demonstrating higher levels of TNF- in patients who have progressive lung fibrosis suggest that this cytokine is pivotal in the pathogenesis of sarcoidosis (10, 14). Similarly, the failure of IFN- gene knockout mice to form granulomas after exposure to mycobacteria highlights the absolute requirement for IFN- in granulomatous inflammation (15).

The importance of TNF- and IFN- in granulomatous inflammation makes defining the molecular mechanisms controlling the expression of these cytokines an important step in understanding disease pathogenesis. To date, little is known about the molecular regulation of these cytokines, and what is known suggests multiple signaling pathways may be involved (16). In addition, recent studies from this laboratory indicate that cytokine gene regulation in transformed cell lines may not be applicable to disease situations (unpublished observations). The complexity of cytokine gene regulation is highlighted by studies showing that in different cell types (e.g., macrophages versus T cells), either nuclear factor (NF)-B- or NF-AT-dependent pathways may regulate TNF- gene expression (17, 21). Even within a given cell type, the mechanism of cell activation can determine which signaling pathway primarily regulates cytokine production. We have recently demonstrated that lipopolysaccharide (LPS)-stimulated TNF- production by primary human monocytes is NF-B dependent, whereas TNF- production after CD45 ligation and zymosan stimulation is phosphatidylinositol-3- kinase and NF-B independent (16, 18). Binding sites for other transcription factors, including activator protein-1 and NF-IL-6, have also been identified in the promoter region of the TNF- gene and may be important in other forms of cell activation (22). The molecular regulation of IFN- is similarly complex, with evidence that both NF-AT and NF-B are involved in the production of this cytokine (19). Although the importance of NF-AT-dependent signaling in IFN- gene transcription is well established, the role of NF-B is less certain, as specific binding sites have yet to be identified in the gene promoter region.

Based on these observations regarding the molecular regulation of cytokine genes, the factors influencing the production of proinflammatory molecules in diseases where the nature of cell stimulation is unknown are most likely to be predicted from the direct investigation of primary cells involved in the pathologic process. Recently, we reported the successful use of adenovirus (Adv) vectors to investigate the molecular mechanisms of cytokine gene regulation in primary human macrophages. Using an Adv vector encoding IB (AdvIB), we demonstrated the importance of NF-B in TNF- production by rheumatoid synovial macrophages (17, 25). Despite a previous report that AM from normal volunteers are refractory to Adv infection (26), we sought to determine if diseased AM were similarly difficult to infect with Adv. We confirmed that AM from normal volunteers were not permissive to Adv infection but found in contrast that sarcoid AM and AT could be efficiently infected. Upregulation of the putative Adv receptors coxsackie/adenovirus receptor (CAR), v3, and v5 integrins (27, 28) was also observed on sarcoid AM and AT relative to normal AM and peripheral blood T cells. Although we had previously demonstrated Adv infection of rheumatoid synovial T cells (25), it is generally accepted that primary T cells are refractory to infection (29, 30). Having established that sarcoid AM and AT could be efficiently infected with Adv, we then used the AdvIB vector to investigate the role of NF-B in the regulation of proinflammatory cytokines that earlier studies had determined were important in the pathogenesis of sarcoidosis (10, 14, 15). We observed that TNF- and IL-6 production by sarcoid AM and IFN- production by sarcoid AT is NF-B dependent, whereas IL-4 production by sarcoid AT is NF-B independent. The successful transfer of Adv transgenes to sarcoid AM and AT provided important information about cytokine gene regulation in sarcoidosis and should be a useful tool for future studies involving the analysis of cytokine gene regulation in inflammatory lung disease.

	Materials and Methods

Top Abstract Introduction Materials and Methods Results Discussion References

Cells

BAL cells were obtained from normal volunteers and patients undergoing diagnostic bronchoscopy for suspected pulmonary sarcoidosis as previously described (31). The diagnosis of sarcoidosis was based on established clinical criteria (1), with consistent histologic changes in at least one affected organ required for inclusion in analysis. The BAL fluid was centrifuged at 1,500 rpm, and cells were resuspended in serum-free RPMI 1640 at 1 × 10⁶ cells/ ml. For each patient, duplicate cytopreparations were prepared, air-dried, and fixed with formaldehyde. A May-Grunwald-Giemsa stain was performed and a cellular differential was calculated. Sarcoid AT were positively selected from the BAL preparation with magnetic polystyrene beads coated with mouse primary monoclonal antibody specific for CD2 (Dynal CD2 CELLection kit) and resuspended in serum-free RPMI. Peripheral blood specimens were obtained from the same sarcoidosis patients and normal volunteers. Peripheral blood mononuclear cells were separated using a lymphoprep gradient and then suspended in serum-free RPMI 1640 medium at 1 × 10⁶ cells/ml. All cell cultures involving infection with Adv vectors were performed in flat-well culture plates (Nunc Life Technologies Ltd., Paisley, UK) at 1 × 10⁶ cells/ml.

Adenoviral Vectors

Replication-deficient, recombinant Adv vectors encoding Escherichia coli -galactosidase (Advgal), porcine IB with a cytomegalovirus promoter and nuclear localization sequence (AdvIB), a green fluorescent protein (GFP) reporter with a cytomegalovirus promoter (AdvGFP), and another with no insert (Adv0) were kindly provided by Drs. Wood and Byne (Oxford, UK), Dr. de Martin (Vienna, Austria), and Dr. T. Mahon (Kennedy Institute, London, UK), respectively. Porcine IB has > 95% homology with the human molecule and effectively inhibits human NF-B- dependent signaling pathways (16, 17). Viruses were propagated in the 293 human embryonic kidney cell line and were purified by ultracentrifugation through two cesium chloride gradients. The titer of viral stocks was determined through a plaque assay on 293 cells as described (32).

Infection Techniques and Assessment of Transgene Expression

Cells were infected at the stated multiplicity of infection (MOI) while suspended in serum-free RPMI 1640. After 2 h, the medium containing virus and nonadherent cells was removed and replaced with complete medium (RPMI 1640 with 5% heat-inactivated fetal calf serum, 25 mM N-2-hydroxyethylpiperazine-N'-ethane sulfonic acid, 2 mM L-glutamine, and 100 U/ml penicillin/ streptomycin). The supernatants containing nonadherent cells were centrifuged, and the cell pellets were resuspended in fresh medium before being reintroduced into the appropriate culture well. In studies involving M-CSF treatment, sarcoid AM were treated with 100 ng/ml M-CSF for 48 h prior to infection with Advgal or Adv0. Similarly, prepared cells were cultured without macrophage colony-stimulating factor (M-CSF) before infection at 48 h and used as a positive control. GFP expression was assessed by direct cellular fluorescence at 48 h using ultraviolet fluorescence microscopy. -galactosidase expression was analyzed 48 h after infection using a fluorescence-activated cell sorter (FACS) as previously described (25, 33).

Analysis of Cell Surface Expression of CAR, v3 and v5 Integrins

The surface expression of CAR, v3, and v5 integrins was analyzed by FACS, using monoclonal antibodies (mAb) to v3 (LM 609, provided by IXSYS, San Diego, CA), v5 (P5H9-E11, W. Smith and J. Gamble, Hanson Centre, Adelaide, Australia), and CAR (kindly provided by R. Finberg, Dana-Farber Cancer Institute, Boston, MA). Expression of v3 and v5 integrins was assayed using an isotype-matched control mAb, OX14 (100 µg/ml), as previously described (17). FACS analysis was performed on cells gated for the correct forward and side-scatter characteristics (size and granularity), as well as surface markers, as previously described (25).

Western Blotting and Electrophoretic Mobility Shift Assay

For the analysis of IB and NF-B cytosolic/nuclear expression 8 × 10⁶, sarcoid BAL cells were prepared in serum-free RPMI medium on a 6-well plate and left uninfected or infected with either Adv0 or AdvIB at an MOI of 150:1. After 48 h, cells were removed from the culture plate using a nonenzymatic "cell dissociation solution" (Sigma, St. Louis, MO) and firm pipetting. Cytosolic and nuclear extracts were prepared as described by Whiteside and coworkers (34). Protein was quantified by Bradford assay, and 100 µg was loaded to each track for separation by sodium dodecyl sulfate polyacrylamide gel electrophoresis on a 10% (wt/vol) polyacrylamide gel before electrotransfer onto nitrocellulose membranes. p42/44 mitogen-activated protein kinase (p42/44 MAPK) expression was analyzed by immunoblotting as a loading control. The anti-IB antibody and anti-p42/ 44 MAPK antibody were purchased from Santa Cruz Biotechnology (Santa Cruz, CA). The secondary antibody was a horseradish peroxidase-conjugated donkey antirabbit antibody (Amersham International, Amersham, UK). For the electrophoretic mobility shift assay, nuclear extracts were prepared and protein quantified by Bradford assay. A total of 20 µg of protein was run on a 5% TBE gel with a ³²P-labeled NF-B consensus oligonucleotide (Promega, Madison, WI) and analyzed as previously described (35).

Analysis of Cytokines

Supernatants from uninfected and AdvIB (or Adv0) infected cells were aspirated at 24 h, centrifuged to remove nonadherent cells, and later analyzed. Enzyme-linked immunosorbent assays for the detection of IL-4 and IFN- were performed using antibody pairs (Pharmingen, Sorrentino, CA) according to the manufacturer's recommended procedures (36, 37). Analysis of TNF- and IL-6 was performed as previously described (38). Cytokine analysis was performed in triplicate with a mean value calculated for each patient specimen. The mean percentage cytokine production by cells infected with Adv0 and AdvIB was calculated relative to uninfected cells from the same specimen (standard error of the mean [SEM] represented by error bars). Analysis was performed on unstimulated cells and those treated with 10 ng/ml LPS (Salmonella typhimurium; Sigma).

Statistical Methods

All statistical testing was performed using a paired comparison, one-sided paired Student's t test.

	Results

Top Abstract Introduction Materials and Methods Results Discussion References

Patient Data

BAL was performed before transbronchial biopsy. The diagnosis of pulmonary sarcoidosis was verified by transbronchial biopsy demonstrating noncaseating granulomas. Cellular analysis of each BAL specimen showed < 5% respiratory epithelial cells. The cellular differential for each specimen was calculated by counting > 300 cells and found to consist of AM (55 to 72%) and AT (28 to 45%), which is in keeping with previously published data (2, 39). The BAL specimens from normal volunteers contained > 94% AM.

Adenovirus Infection of AM and AT

Previous studies had suggested that normal AM were refractory to Adv infection (26). Using our Adv gene delivery system, we sought to verify this result and determine if sarcoid AM and AT were also refractory to infection. We observed that sarcoid cells, but not normal BAL cells, were readily infected with Advgal and that the optimal MOI was 150:1, with > 95% of cells expressing -galactosidase (Figure 1A). Double gating for cell surface markers (CD3 for AT and CD14 for AM) and cell size/granularity determined that both sarcoid AM and AT were equally infected (Figure 1B). A second AdvGFP confirmed that there was efficient infection of sarcoid BAL specimens (Figure 2A). GFP expression was not detected in AM from normal volunteers, confirming the results of Kaner and colleagues (26) (Figure 2B). As significant numbers of AT are not present in normal BAL specimens (< 5%), we analyzed peripheral blood T cells and monocytes form sarcoid patients to determine if increased susceptibility to Adv infection was a property of all hemopoetic cells in sarcoidosis. We observed that peripheral blood T cells and monocytes from sarcoid patients were refractory to Adv infection (results not shown). These observations were in agreement with previous studies showing that freshly prepared peripheral blood monocytes and T cells from normal subjects could not be infected with Adv (17, 29). The data suggest that enhanced susceptibility to Adv infection is a property of T cells and macrophages present at sites of inflammation.

View larger version (38K): [in this window] [in a new window]   Figure 1.   -galactosidase expression in sarcoid AM and AT. Adv transgene expression by sarcoid AM and AT was analyzed by FACS. The fluorescence of cells infected with Advgal was analyzed at 48 h and compared with cells infected with Adv0. (A) Histograms representing the FL1 fluorescence of the total cell population after infection with Advgal (shaded) and Adv0 (open) at an increasing MOI: (i) 50:1, (ii) 150:1, and (iii) 300:1. At an MOI of 150:1, > 95% of cells express -galactosidase. (B) Cells were gated for size/granularity and surface markers (CD14+/CD3 for AM and CD14/CD3+ for AT) to confirm that both AM and AT were infected with Adv at an MOI of 150:1. FACS data in this figure are representative of five patient samples.

View larger version (102K): [in this window] [in a new window]   Figure 2.   GFP expression in sarcoid AM and AT. Adv transgene expression in BAL specimens obtained from patients with sarcoidosis and normal volunteers. Cells were infected with AdvGFP at an MOI of 150:1, and transgene expression was assessed by ultraviolet (UV) light microscopy after 48 h. White light microscopy confirmed that equivalent cell numbers were present in each specimen. (A) Efficient GFP transgene expression was detected in the sarcoid BAL specimen. (B) Little GFP expression could be detected in the normal BAL specimen after infection with AdvGFP at the same MOI. The occasional fluorescent cell seen in the normal BAL preparation may represent bronchial epithelial cells. The data represented in this figure are representative of six BAL specimens.

Efficient Infection of Sarcoid AM and AT Is Associated with Increased Expression of CAR, v3, and v5 Integrins

Adv entry into cells has been associated with the surface expression of CAR and v3 and v5 integrins (27, 28). With FACS analysis, we observed low or undetectable levels of CAR, v3, and v5 integrins on the surface of normal AM and peripheral blood T cells that were refractory to infection (Figure 3A and 3C). Furthermore, these Adv receptors were also undetectable on the surface of peripheral blood monocytes form normal volunteers and patients with sarcoidosis (results not shown). In contrast, sarcoid AM that were permissive to Adv infection consistently expressed all three of these surface antigens and sarcoid AT expressed CAR and v5 integrins, but not v3 integrins (Figures 3B and 3D). These results indicate that enhanced susceptibility to Adv infection correlates with upregulation of CAR, v3, and v5 integrins.

View larger version (41K): [in this window] [in a new window]   Figure 3.   Adenovirus receptor expression. The expression of the CAR, v3, and v5 integrins on (A) normal AM, (B) sarcoid AM, (C) sarcoid peripheral blood T cells, and (D) sarcoid AT was analyzed by FACS. Cells were gated for size/granularity and the appropriate cell surface marker (CD14+/CD3 for AM and CD14/CD3+ for AT). The FL1 fluorescence of cells incubated with mAb to (i) v3, (ii) v5, and (iii) CAR, respectively (shaded histogram), was compared with that of cells incubated with an isotype control mAb. The results in (B to D) were obtained from a single patient with sarcoidosis. The data presented are respresentative of four consecutive sarcoid patients and three normal volunteers.

M-CSF Downregulates the CAR, v3, and v5 Integrins on Sarcoid AM and Prevents Adenovirus Infection

Previously, we demonstrated that treating primary monocytes with M-CSF for 48 h upregulates v3 and v5 integrins and enhances the efficiency of Adv infection (17). However, further treatment of monocytes with M-CSF for 5 d leads to a paradoxical downregulation of v3 and v5 integrins and reduced susceptibility to Adv infection (Dr. A. Foey, personal communication). We questioned whether freshly prepared sarcoid AM and AT were similar to 2-d M-CSF-treated monocytes and whether further M-CSF exposure would result in reduced Adv infection. We found that M-CSF caused downregulation of CAR, v3, and v5 integrins and a dramatic reduction in Adv transgene expression (Figures 4A and 4B). Sarcoid AM not treated with M-CSF demonstrated no change in surface antigen expression or efficiency of Adv infection after 48 h of in vitro culture.

View larger version (27K): [in this window] [in a new window]   Figure 4.   Effect of M-CSF on adenovirus transgene and receptor expression. The effect of M-CSF treatment on the infection of sarcoid AM by Adv was analyzed by FACS. (A) Sarcoid AM were cultured in complete medium for 2 d with and without 100 ng/ml M-CSF. Cells were then infected with Adv-gal at an MOI of 150:1 and analyzed by FACS after a further 48 h. The cellular fluorescence of M-CSF-treated cells (shaded histogram) was compared with cells not treated with M-CSF (open histogram). Two-day cells treated with M-CSF infected with the Adv0 control virus demonstrated similar fluorescence to the Advgal-infected cells treated with M-CSF (results not shown). (B) Comparison of cell surface v3, v5, and CAR levels on M-CSF-treated (shaded histograms) and non-M-CSF-treated (open histograms) sarcoid AM from the same specimen. The results in this figure are representative of three experiments.

Infection with AdvIB Results in IB Overexpression in Sarcoid BAL Cells

Having demonstrated efficient Adv infection of sarcoid AM and AT, we investigated whether the delivery of transgenes could be used to modify intracellular signaling pathways. Cytosolic and nuclear overexpression of IB was observed in sarcoid BAL cells infected with an AdvIB, but not with an empty Adv vector (Adv0) (Figure 5A). Reprobing of the immunoblot with a p42/44 MAPK antibody confirmed that equivalent amounts of protein had been loaded in each lane. An EMSA for NF-B DNA binding activity was performed on the corresponding nuclear extracts from the Adv0- and AdvIB-infected cells. The significant constitutive nuclear NF-B activity detected in Adv0-infected sarcoid BAL cells reflected the activated nature of these cells (Figure 5B).

View larger version (82K): [in this window] [in a new window]   Figure 5.   Inhibition of NF-B-dependent signaling. Cytosolic IB and NF-B nuclear binding activity in sarcoid BAL specimens infected with either AdvIB or Adv0 (MOI, 150:1). After 48 h, cells from each virally infected preparation were either left untreated (unstim) or stimulated with 10 ng/ml LPS for 45 min. The cells were then lysed, and cytosolic and nuclear extracts were prepared. (A) Immunoblot of cytosolic and nuclear IB expression. Equivalent amounts of protein were loaded on each track (100 µg), and reprobing with a p42/44 MAPK antibody was performed to confirm this. (B) An EMSA for NF-B DNA binding activity performed on 20 µg of nuclear extract. The experiment illustrated in this figure is representative of four studies performed on four consecutive sarcoid BAL samples.

As expected, LPS stimulation resulted in cytosolic degradation of IB in the Adv0-infected cells. There was a corresponding increase in NF-B DNA binding activity in the nuclear extracts of the LPS-treated, Adv0-infected cells. In contrast, in four consecutive studies there was no significant reduction in the expression of IB in the cytosolic extracts of AdvIB-infected cells after activation with LPS. In AdvIB-infected cells overexpressing IB, there was a significant decrease in constitutive NF-B DNA binding activity relative to Adv0-infected cells (Figure 5B, comparison of lane 1 with lane 3). Furthermore, there was a dramatic reduction in the LPS-mediated augmentation of NF-B DNA binding activity in the nuclear extracts of AdvIB-infected cells relative to Adv0-infected cells (Figure 5B, comparison of lane 2 with lane 4). The experiments illustrated in Figures 5A and 5B are representative of four consecutive studies performed on sarcoid BAL specimens. In each of the four studies, there was consistent inhibition of constitutive and LPS-stimulated NF-B activation in AdvIB-infected sarcoid BAL cells overexpressing IB.

IB Overexpression Inhibits TNF- and IL-6 Production by AM and IFN- by AT, but Not IL-4 by AT

Having demonstrated that infection of sarcoid BAL cells with AdvIB resulted in overexpression of IB and inhibition of nuclear NF-B DNA binding activity, we investigated whether this would influence the production of proinflammatory cytokines that are dependent on NF-B- dependent signaling pathways. Due to variation between patients, results are expressed as a percentage relative to uninfected cells from the same BAL specimen. At 24 h, the constitutive TNF- production by AdvIB-infected AM was reduced to 25.9 ± 11.6% (P < 0.001, n = 7) and IL-6 to 31.3 + 9.7% (P < 0.002, n = 7) (Figures 6A and 6B). The enhanced TNF- and IL-6 production after LPS stimulation was also inhibited by IB overexpression, with TNF- production by AdvIB-infected cells reduced to 11.4 ± 3.9% (P < 0.0001, n = 7) and IL-6 to 7.5 ± 2.8% (P < 0.0002, n = 7) (Figures 6C and 6D). Our results indicate that both constitutive and LPS-stimulated TNF- and IL-6 productions by sarcoid AM require nuclear translocation of NF-B. We also examined the effect of AdvIB infection on IFN- production by sarcoid AT. Constitutive IFN- production in AdvIB-infected cells was reduced to 27.9 ± 9.1% (P < 0.0002, n = 7) (Figure 7A). In contrast, IL-4 production by AdvIB-infected cells was 91.89 ± 10.1% and not significantly different in uninfected cells (P = 0.65, n = 5) (Figure 7B). There were no significant differences in production of the measured cytokines by uninfected compared with Adv0-infected cells, indicating that virus infection per se did not influence cytokine gene regulation (Figures 6 and 7). The percentage cytokine production by Adv0-infected AM relative to uninfected AM was 106.8 ± 25.0, 114.4 ± 17.1, 97.6 ± 12.3, and 104.6 ± 13.9% for constitutive and LPS-stimulated TNF- production, and constitutive and LPS-stimulated IL-6 production, respectively. Similarly, the percentage of IFN- and IL-4 production by Adv0-infected AT relative to uninfected AT was 126.3 ± 28.0 and 95.2 ± 21.7%, respectively.

View larger version (44K): [in this window] [in a new window]   Figure 6.   Inhibition of alveolar macrophage cytokines by IB overexpression. Constitutive and LPS-stimulated TNF- and IL-6 production at 24 h by sarcoid AM was analyzed by ELISA. TNF- and IL-6 production by AdvIB- and Adv0-infected cells (MOI, 150:1) is expressed as a percentage relative to uninfected cells. AdvIB infection reduced constitutive production of (A) TNF- to 25.9 ± 11.6% (P < 0.001) and (B) IL-6 to 31.3 ± 9.7% (P < 0.002). Range for constitutive TNF-: uninfected, 805.8 pg/ ml; AdvIB, 278.0 to 893.2 pg/ml. Range for constitutive IL-6: uninfected, 2,592 to 10,328 pg/ml; AdvIB, 516.1 to 3738 pg/ml. After LPS treatment (10 ng/ml), AdvIB infection reduced production of (C ) TNF- to 11.4 ± 3.9% (P < 0.0001) and (D) IL-6 to 7.5 ± 2.8% (P < 0.0002). Range for LPS-induced TNF-: uninfected, 9,620 to 64,443 pg/ml; AdvIB, 915.2 to 11,251 pg/ml. Range for LPS-induced IL-6: uninfected, 12,490 to 86,128 pg/ml; AdvIB, 492.3 to 5,784 pg/ml. The differences in constitutive and LPS-induced TNF- and IL-6 production by Adv0 compared with uninfected cells were not statistically significant. Error bars indicate SEM (n = 7).

View larger version (27K): [in this window] [in a new window]   Figure 7.   Inhibition of alveolar T-cell cytokines by IB overexpression. Constitutive IFN- and IL-4 production by sarcoid AT at 24 h was analyzed by ELISA. IFN- and IL-4 production by AdvIB- and Adv0-infected cells (MOI, 150:1) is expressed as a percentage relative to uninfected cells. (A) AdvIB infection reduced IFN- production to 27.9 ± 9.1% (P < 0.0002). Range: uninfected, 45.4 to 1,254 pg/ml; AdvIB, 15.7 to 290.1 pg/ml. (B) In contrast, IL-4 production by AdvIB-infected AT was 91.89 ± 10.1% and not significantly different to uninfected cells (P = 0.65). Range: uninfected, 43.9 to 82.9 pg/ml; AdvIB, 39.7 to 89.9 pg/ml. Error bars indicate SEM (n = 5 to 7).

Inhibition of IFN- Production Is Due to Direct Inhibition of NF-B Activity in Sarcoid AT

Assuming that mutual activation of AM and AT is responsible for cytokine production in sarcoidosis, it was possible that the inhibition of IFN- was due to AdvIB infection of AM, resulting in reduced AT activation. To test if the reduction in IFN- production was the direct result of NF-B inhibition in sarcoid AT, sarcoid AT were isolated (> 95% pure), infected independently with AdvIB, and then reintroduced to the uninfected AM. Under these conditions, IFN- produced by the AdvIB-infected AT at 24 h was reduced to 46.2 ± 7.7% (P < 0.006, n = 4) (Figure 8). The percentage of IFN- production by Adv0-infected AT relative to uninfected AT was 109.0 ± 27.8%. The inhibition of IFN- was approximately two-thirds that seen when both the AM and AT populations were infected with AdvIB. The data suggest that in the experiments involving coinfection of AT and AM, reduced NF-B nuclear binding in sarcoid AT is the major reason that IFN- production is inhibited but that it is likely that reduced activation of AM also contributes to reduced IFN- production.

View larger version (36K): [in this window] [in a new window]   Figure 8.   Evidence of inhibition of alveolar T-cell NF-B signaling pathways. IFN- production at 24 h by sarcoid AT selectively infected with AdvIB (MOI, 150:1) was analyzed by ELISA. Sarcoid AT were positively selected from BAL specimens by magnetic beads coated with anti-CD2 mAb. AT were then infected with either AdvIB or Adv0 before being reintroduced to the T cell-depleted cultures containing uninfected AM. The positively selected AT were determined to be > 94% pure by FACS analysis. The levels of IFN- produced by AdvIB- and Adv0-infected cells are expressed as a percentage relative to uninfected cells. Selective AdvIB infection of sarcoid AT reduced IFN- production to 46.2 ± 7.7% (P < 0.006). Range: uninfected, 291.3 to 534.6 pg/ml; AdvIB, 132.3 to 231.8 pg/ml. Error bars indicate SEM (n = 4).

	Discussion

Top Abstract Introduction Materials and Methods Results Discussion References

This study has established that Adv vectors are an effective way of investigating cytokine gene regulation in sarcoid AM and AT. In addition to demonstrating that AM and AT from patients with pulmonary sarcoidosis efficiently expressed Adv transgenes, we confirmed an earlier report that normal AM are refractory to Adv infection (26). We observed that sarcoid AM and AT that were permissive to Adv infection expressed CAR, v3, and v5 integrins, but normal AM and peripheral blood T cells did not, supporting earlier reports that these surface antigens are involved in Adv infection of primary human macrophages and T cells (17, 26, 29). By infecting sarcoid BAL cells with AdvIB, we established that NF-B-dependent signaling is necessary for TNF- and IL-6 production by AM and IFN- production by AT. IL-4 production by AT was not inhibited by IB overexpression, indicating that NF-B is not required for the expression of this cytokine.

Although hemopoetic cells are not natural targets for Adv, under certain conditions significant levels of infection can be achieved. Previous studies by ourselves and others have demonstrated that partial differentiation of monocytes with M-CSF for 48 h upregulates v3 and v5 integrins and increases the efficiency of Adv infection (17, 29). In addition, we have also observed that macrophages obtained from rheumatoid joints can also be efficiently infected with Adv, without pretreatment with growth factors (25). Interestingly, more prolonged differentiation of peripheral blood monocytes with M-CSF (7 d) results in a paradoxical decrease in infection and reduced v3 and v5 integrin expression (Dr. A. Foey, personal communication). Until now, it has been unclear whether these differences in susceptibility to Adv infection reflected monocytic differentiation into macrophages or contact with a chronic inflammatory environment. Our results suggest that it is the latter, as normal AM do not express CAR, v3, and v5 integrins, and cannot be infected with Adv. Further information regarding the mechanisms that control Adv infection of monocytic cells was provided by experiments involving M-CSF-treated sarcoid AM. Sarcoid AM, like 2-d peripheral blood monocytes treated with granulocyte macrophage colony-stimulating factor (GM-CSF) or M-CSF, are readily infected with Adv and express v3 and v5 integrins (17). In contrast, normal AM are refractory to Adv infection, not expressing CAR, v3, or v5 integrins and in this respect, resemble 48-h sarcoid AM treated with M-CSF and 7-d peripheral blood monocytes treated with M-CSF. The data suggest that sarcoid AM are immature, and as they are driven toward a more mature macrophage phenotype, they become refractory to Adv infection. An earlier report that sarcoid AM express monocytic surface markers not present on AM obtained from normal volunteers supports the hypothesis that sarcoid AM are phenotypically immature (40). It is possible that the Adv infection is not only a property of immature macrophages, but also a reflection of a more proinflammatory cell phenotype. We, like others, have observed that sarcoid AM spontaneously produce a number of proinflammatory cytokines, including TNF-, IL-1, and IL-6, but little of the immunoregulatory cytokine IL-10 (9, 10, 41, 42). This correlates with earlier work from our laboratory showing that the ratio of TNF- to IL-10 production is greater in 2-d compared with 7-d monocytes treated with M-CSF (Dr. A. Foey, personal communication). It is possible that further differentiation of sarcoid AM could also alter the balance of cytokine production toward a more anti-inflammatory profile.

Freshly prepared peripheral blood T cells are as refractory to Adv infection as are monocytes (29, 30). Unlike the situation with monocytes, however, there is no known simple in vitro strategy (e.g., GM-CSF treatment) that results in efficient Adv infection. Even modification of virus tropism has been only moderately successful in promoting Adv entry into T cells (30). The environmental factors permitting efficient Adv infection of sarcoid AT are obscure. The similarities with rheumatoid synovial T cells suggest that infection may be influenced by exposure to a chronic inflammatory environment, although attempts to replicate these conditions ex vivo (e.g., growth factors, antigenic stimulation) have been unsuccessful in reproducing the levels of infection observed in our study (29, 30). In a situation analogous to monocytes, CAR and v5 integrins are also expressed on sarcoid AT, but not on peripheral blood T cells. The data suggest that these surface antigens are also involved in T-cell infection by Adv.

A major goal of this study was to establish if Adv transgenes could be used to investigate the disease-specific role of NF-B in cytokine gene regulation in pulmonary sarcoidosis. Although it has been previously demonstrated that NF-B activity is elevated in inflammatory lung disease (43), a direct correlation with TNF-, IL-6, or IFN- production has not previously been established. This is the first occasion that the selective inhibition of the molecular pathway that controls the nuclear translocation of NF-B has allowed investigation of the regulation of cytokine gene transcription in primary AM and AT. We observed that TNF- and IL-6 production by sarcoid AM and IFN- production by sarcoid AT are NF-B dependent, but that IL-4 production is NF-B independent. The inhibition of IFN- observed when isolated sarcoid AT were infected with AdvIB and reintroduced to uninfected sarcoid AM provides the strongest evidence that IFN- gene regulation is NF-B dependent. The inhibition of IFN- was, however, greater when AT and AM were coinfected with AdvIB, suggesting that NF-B activity in AM also indirectly contributes to IFN- production by AT. Macrophage functions requiring NF-B that may influence IFN- production by AT include TNF- production and antigenic presentation (46).

The requirement for NF-B in the production of TNF- and IL-6 has previously been established in rheumatoid synovial macrophages and LPS-stimulated peripheral blood monocytes (17, 25). However, these data may not be applicable to sarcoidosis where the mechanism of cell activation is unknown, as we have also previously shown that the requirement for NF-B in TNF- and IL-6 gene expression depends on the mechanism of cell stimulation (16). Although the role of NF-AT-dependent signaling in IFN- gene transcription has been well established (47, 48), the importance of other transcription factors, including NF-B, is less certain. Despite the failure to identify a canonical NF-B binding site within the promoter region of the IFN- gene, earlier reports have suggested that it is involved in the transcription of this gene, possibly by associating with a site that also binds NF-AT (19, 20). Our study provides the strongest evidence to date that NF-B is required for IFN- gene expression but does not demonstrate direct binding of the transcription factor to the gene promoter. It is possible that NF-B may be an intermediary protein required for the production of another transcription factor that primarily regulates IFN- gene expression. Answering this question would require the production of an IFN- gene construct lacking the putative NF-B binding site within the promoter region. Although pulmonary sarcoidosis is primarily a Th1-mediated disease, AT retain the ability to produce Th2-type lymphokines such as IL-4. The observation that IL-4 production was not inhibited by AdvIB infection is consistent with previous reports that the transcription of this cytokine gene requires NF-AT, but not NF-B, dependent signaling (48).

In this study, we have established the requirement of NF-B in the regulation of important proinflammatory cytokines in pulmonary sarcoidosis. Our data suggest that CAR, v3, and v5 integrins are required for Adv infection of AM and AT, although the in vivo factors influencing the expression of these surface antigens are not fully understood. Apart from being a useful tool for in vitro studies of cytokine gene regulation, the data also suggest that Adv vectors could be used as a vehicle for the in vivo delivery of transgenes to hemopoetic cells involved in inflammatory lung disease. The ability to selectively infect diseased, but not normal, AM and the accessibility of the lower respiratory tract may make Adv-mediated gene therapy a possible treatment option for these conditions in the future.