Alveolar Macrophages from Patients with AIDS and AIDS-related Complex Constitutively Synthesize and Release Tumor Necrosis Factor Alpha 1 - 3
CARLO AGOSTINI, RENATO ZAMBELLO,4 LIVIO TRENTIN,5 SPIRIDIONE GARBISA, PAOLA FRANCIA 01 CELLE, PIETRO BULIAN, MAURIZIO ONISTO, VENERINO POLETTI, LUIGI SPIGA, ENZO RAISE, ROBERT FoA, and GIANPIETRO SEMENZATO
Introduction
Over the last few years evidence has been accumulating that HIV-l mayaffeet the cellular immune system in the lung of patients with AIDS (1, 2). Major immunologic abnormalities reported to date include an alveolitis, mainly characterized by cytotoxic/suppressor CD8+ lymphocytes and natural killer cells, a severe depletion of pulmonary helperrelated CD4+lymphocytes and, in some cases, an increase of polymorphonuclear cells and alveolar macrophages (AMs) (3-8). Because it is well known that AMs represent an important target for HIV-l (9-11),the study of their functional capabilities in vitro might help in understanding the pathogenesis of HIV-l-related pulmonary complications. In response to tissue damage and/or inflammation, AMs secrete a series of soluble molecules, termed cytokines, that promote activation of the lymphocyte population and mediate the killing of microbes and tumoral cells (12). Among these, great interest has been recently paid to tumor necrosis factor alpha/cachectin (TNFa), a molecule produced by cells of the host immune system, mainly by monocyte-macrophages, under conditions that resemble host invasion (13). The release of TNFa has been implicated in the pathogenesis of several human diseases (14-17),including AIDS (18).In particular, high levelsof TNFa have been found in the sera and supernatants of circulating monocytes and tissue macrophages from patients with HIV-l infection (19-24), even if data are not always consistent (25-27). The current study was designed to determine whether in patients with AIDS and ARC, AMs, i.e., the representatives of the monocytic-macrophagic lineage in the lower respiratory tract, are capable of synthesizing and releasing detectable
SUMMARY To verify the hypothesis that alveolar macrophages (AMs) from patients infected with HIV-1 could synthesize and release TNFa, AMs recovered from the BAL fluid of 11 patients with seropositive HIV-1 (six with AIDS and five with ARC) were tested in vitro for their ability to destroy TNFa-susceptible targets. Furthermore, the presence of TNFa was assessed in AM-conditioned supernatants on the basis of their cytotoxic activity and by using an immunoenzymatic test and immunoblotting. Transcription of the TNFa gene in AMs was also studied by means of the Northern blot analysis. AMs freshly recoveredfrom patients infected with HIV-1 exhibited high levels of cell-mediated cytotoxicity against U937 targets, and the addition of a polyclonal anti-TNFa antibody resulted in a significant inhibition of the target lysis. Cell-free supernatants conditioned by unstimulated AMs exerted high levels of cytotoxic activity against TNFa-sensitive targets, whereas duplicate, neutralization experiments performed in the presence of an anti-TNFa antibody proved that the observed cytotoxic activity was mostly mediated by TNFa. The presence of high amounts of TNFa in the conditioned media was confirmed by the immunoenzymatic test. In addition, the immunoblot analysis showed that the TNFa released by AMs has a Mr 17,000 band, identical to a standard preparation of recombinant TNFa. The Northern blot demonstrated that unstimulated AMs express detectable levels of mRNA transcripts for TNFa. Takentogether, our data support the concept that AMs from patients with HIV-1 infection constitutively release TNFa. Because normal human AMs do not express detectable levels of mRNA for TNFa or release this cytokine, the hypothesis is formulated that in the lung of patients with AIDS, AMs, either by a direct effect of HIV-1 infection or as a consequence of the pulmonary opportunistic infections, are triggered to synthesize TNFa. This in situ overproduction might playa role in the pathogenesis of AIDS-related pulmonary complications. AM REV RESPIR· DIS 1991; 144:195-201
amounts of TNFa. For this purpose, AMs recovered from the bronchoalveolar lavage (BAL) fluid of 11 patients with seropositive HIV-l were tested for their ability to destroy TNFa-susceptible cell lines, and the presence of TNFa was determined in the conditioned supernatants by cytotoxicity experiments, by an enzyme-linked immunosorbent assay, and by immunoblotting. The transcription of the TNFa gene in AMs was also studied by means of the Northern blot analysis. Methods Study Population Eleven patients with seropositive HIV-l (average age, 28 ± 8 yr; seven men and four women) were studied. All patients belonged to the IV group of the CDC classification (28), which includes patients with clinical symptoms and signs of HIV-l infection other than,
(Received in original form July 31, 1990 and in revised form February 26, 1991) 1 From the Padua University School of Medicine, Department of Clinical Medicine, 15t Medical Clinic, Clinical Immunology Branch, and Department of Histology; Department of Biomedical Sciences and Human Oncology, 15t Medical Clinic, University of Torino; Department ofPneumology,Bellaria Hospital Bologna and Department of Infectious Diseases, Maggiore Hospital Bologna, Padua, Italy. 2 Supported by grants from the Ministero della Sanita, Istituto Superiore di Sanita, Progetto AIDS 1989, 1990 (Rome, Italy), by Progetto Finalizzato CNR "Oncologia" 88.00680.44, and by the Italian Association for Cancer Research (AIRC, Milan). 3 Correspondence and requests for reprints should be addressed to Gianpietro Semenzato, M.D., Istituto di Medicina Clinica dell'Universita di Padova, Clinica Medica 1, Via Giustiniani 2, 35128 Padova, Italy. 4 Recipient of a Fellowship from the Ministero della Sanita, Istituto Superiore di Sanita (Rome). 5 Recipient of a Fellowship from CNR.
195
196
AGOSTINI, ZAMBELLO, TRENTIN, GARBISA, FRANCIA 01 CELLE, BULIAN, ONISTO, POLETTI, SPIGA, RAISE, FoA, AND SEMENZATO
or in addition to, lymphoadenopathy. According to the CDC criteria, five patients had ARC (IV-A) and six had AIDS (IV-Cl). BAL was performed to obtain a specific diagnosis of opportunistic pulmonary infection and/or AIDS-related tumors, according to our study protocol of patients with HIV-l and pulmonary involvement, which includes the use of BAL and of transbronchiallung biopsy (TLB) during the. fiberoptic bronchoscopy examination. According to the declaration of Helsinki, this invasiveprocedure is performed only for diagnostic purposes when a pulmonary complication is suspected; for this reason asymptomatic patients infected with HIV-l but without signs of pulmonary disease were not included-in the study. Two of the 11 patients weresmokers. The risk factors for AIDS included intravenous drug abuse (n = 9) and homosexuality (n = 2). Pneumocystis carinii infection was detected in the BAL or TLB of all six patients with AIDS. Two patients with ARC showed a histologic pattern consistent with aspecific chronic pneumonia on TLB, whereas the three remaining patients with ARC did not show detectable signs of pulmonary infections. None of the patients had a pulmonary localization of Kaposi's sarcoma. Five normal volunteers (two smokers and three nonsmokers; four men and one woman; average age, 29 ± 6 yr) were accepted as control subjects. They had a normal physical examination, chest radiograph, and lung function tests. As stated above, we were not allowed to perform BAL analysis on control subjects at risk for AIDS and HIV seronegative.
Preparation of Cell Suspensions and Purification of AMs BAL was performed after local anesthesia according to the method previously described (5). The percentages of lavage recovery were 51.5 ± 6.90/0 and 54.1 ± 4.70/0 of the injected fluid in patients with HIV-l and control subjects, respectively. The mean numbers of cells per milliliter of fluid recovered were 263 x 103 ± 70 and 141 x 103 ± 41 in patients with seropositive HIV-l and normal subjects, respectively. A differential count made by morphologic criteria on cytocentrifuged smears showed that 80.7 ± 5.1, 15.3 ± 4.7, 3.9 ± 3.1,and 0.1 ± 0.1% of BAL cells from patients with HIV-l were macrophages, lymphocytes, neutrophils, and eosinophils, whereas the corresponding values in normal subjects were 91.1 ± 3.1, 7.3 ± 1.0, 0.9 ± 0.2, and 0.7 ± 0.50/0, respectively. Cells recovered from the BAL were centrifuged on a Ficoll-Hypaque density gradient, washed, and then resuspended in RPMI 1640culture medium (Grand Island Biological Co., Paisley, Scotland, UK) supplemented with 2 IJM glutamine, 100 U/ml penicillin, 100 ug/ml streptomycin, and 200/0 FCS (Flow Laboratories, Irvine, Scotland, UK). Adherent cells were removed and isolated from the entire mononuclear cell suspension
by two sequential incubations at 37° C in a humidified atmosphere of 5 % CO 2 and 950/0 air for 60 min in plastic petri dishes, as previously reported (29). To avoid the presence of nonadherent cells on the populations under study, in two patients who showed high intensity lymphocytic alveolitis, CD3+cellswere removed using magnetic microspheres coated with goat anti-mouse IgG (Dynabeads; Dynal, Oslo, Norway), as already described (29); immediately after the removal of CD3+cells, the adherence procedure was performed as above. The population resulting from the adherence procedure was composed of more than 95 % macrophages, as determined by nonspecific esterase staining, and by less than 1% CD3+cells. More than 900/0 of the isolated macrophages were viable, as judged by trypan blue exclusion.
Cell-mediated Cytotoxic Assay Cell-mediated cytotoxicity was assessed by lysis of s1Cr-labeled target cells represented by the TNFa-susceptible U937 targets, as described elsewhere (30). Evaluation was performed using AMs depleted of lymphocytes and immediately used as effector cells. BrieflY,1 x 106 targets werelabeled for 4 h at 37° C in 50/0 CO 2 with 100 flCi Na (51Cr)04 (CEA IRE Sorin; Biomedica, Saluggia, Italy) and extensively washed before use. Then 100 IJI of the target cells suspension (x lOS Iml) were suspended in each well of a V-shaped plate (Tirtetek; Falcon, Becton Dickinson, Sunnyvale, CA), and graded concentrations of effector cells (5:1, 10:1, 20:1, and 40:1) were added to wells in triplicate and incubated at 37° C in 50/0 CO 2 for 18h. Supernatants were then harvested and counted in a gamma counter. The mean value of triplicate assays was used to calculate the percentage of cytotoxicity according to the following formula: (cpm effector cells - cpm spontaneous release) x 100/(cpm maximum release - cpm spontaneous release). Spontaneous release from the U937 targets was always less than 80/0. Because U937 targets are NK-sensitive cells, we determined by flow cytometry whether cells bearing the NK-related MAb CD16 and CD56 were present among the effector cell population. In every experiment less than 0.1% of the effector cells were CDI6+ or CD56+. To verify whether AMs cytotoxicity was mediated by TNFa, inhibition experiments were carried out, in duplicate tests, by adding 1/100 diluted polyclonal rabbit antihuman TNFa antibody (IP-300, a mixture of primarily IgG and IgM; Genzyme Co, Boston, MA) to the effector cells and U937 targets. In preliminary experiments we demonstrated that a 1/100 dilution of IP-3oo antibody inactivates approximately 20 ng of purified human TNFa. Neutralization experiments of cell-mediated cytotoxicity were performed at the 40:1 E:T ratio. Preliminaryexperiments showed that IP-3oo antibody alone had no effect on S1Cr release by targets.
Determination of Cytotoxic Activity Exerted by Supernatants Obtained from Unstimulated and Stimulated AMs Unstimulated AMs (5 x lOs/ml) resuspended in serum-free Iscove's modified Dulbecco's medium (IMDM) (Flow Laboratories) were cultured in 24-wellplates at 37° C in 50/0 CO 2 for 48 h. In separate experiments AMs were stimulated with IFN-y (100 U/ml; Behring Inst., Marburg, FRG), PMA (5 ng/ml), and LPS (10ug/ml; Difco Lab., Detroit, MI). After the incubation period, supernatants were harvested, filtered through a 0.45-flm Millipore filter (Millipore Corp., Bedford, MA), and immediately assayed for cytotoxicity or stored at - 80° C for further use. The lytic activity exerted by supernatants was assessed by measuring cytotoxicityagainst s1Cr-labeled U937 targets. In particular, 50 ul of target cells (1 x 106/ml) and 100 ul of cell-free supernatants containing the putative cytotoxic factor were combined in triplicate in 96-wellU-bottomed sterile microtiter plates (Falcon), and the final volume was adjusted to 200 ul with IMDM. After 20 h of incubation at 37° C in 50/0 CO 2 , 100 fll from each wellwere harvested, and the radioactivity was counted in a gamma counter. The percentage of cytotoxicity was calculated as follows: (cpm test - cpm spontaneous) x 100/(cpm maximum - cpm spontaneous). In all experiments the spontaneous release was less than 80/0. In duplicate neutralization experiments, an antihuman TNFa antibody (Genzyme Co.) was used to evaluate whether the cytotoxic activity exerted by supernatants was due to the presence of TNFa. The inhibitory activity of the anti-TNFa antibody was tested by mixing 1/100 of the antibody with the supernatants containing the putative TNFa, and then the assay was carried out as above. The characteristics of IP-300anti-TNFa antibody have been reported above.
Detection of TNFa in the Conditioned Supernatants TNFa levels were determined in the supernatants obtained from AMs cultured as above using a sandwich enzyme immunoassay (Biokine TNF; T Cell Science, Inc., Cambridge, MA). Briefly, TNFa available in the samples or standards binds to polystyrene microtiter wellspreviously incubated with an anti-TNFa coating antibody (lOP ul/well of a solution of 0.1 ml of anti-TNFa in 10 ml of PBScoating buffer). A horseradish-peroxidaseconjugated anti-TNFu antibody binds to the TNFa captured by the first/antibody and completes the sandwich. After washing to remove the unbound enzyme-conjugated anti-TNFa antibody, a substrate solution is added to the wells. The reaction is then stopped, and the adsorbance is determined at 490 nm. A standard curve was prepared by using different reference preparations containing known concentrations of TNFa. Results were expressed as nglml of conditioned supernatants. According to the manufacturer, the minimum
197i'·.v
TNFa PRODUCTION BY ALVEOLAR MACROPHAGES IN HIV-1 INFECTION
dose of TNFa detectable by the immunoassay is 4.8 pg/ml.
Polyacrylamide Gel Electrophoresis and Immunoblotting The TNFa present in conditioned supernatants obtained from 18-h-cultured AMs was further characterized by SDS-PAGE and immunoblotting. Conditioned media (500 Ill) from 24-well plates, obtained as reported above, and a recombinant standard TNFa preparation (Knoll AG, Ludwigshafen, Germany) were incubated with rabbit antihuman TNFa antibody (Genzyme Co) at final concentration of 1:200; 50 III of antirabbit IgG (H and L chain-specific) chromatography gel (Cappel Laboratories, Cochranville, PA) were then added and incubated under rolling for 1 h at room temperature. After centrifugation for 5 min at 200 g, the precipitate was resuspended in sample buffer, and the SDSPAGE was performed using a TRIS-glycine buffer system with 0.5 M urea and 12070 separating and 5070 stacking gels. Transfer of proteins from polyacrylamide gels to Hybond-c extra (Amersham Co, Buckinghamshire, UK) was performed using 12 mM TRIS and 100 mM glycine buffer containing 10070 methanol with a Bio-Rad Trans-blot Cell (Bio-Rad Laboratories, Richmond, CA) for 2 h at 0.2 ampere. The transfer efficiency was verified by staining the gels after transfer with Coommassie blue. The Hybond-c extra was then saturated with 3070 gelatin solution, and treated with polyclonal rabbit antihuman TNFa antibody 1/200 (Genzyme Co) and a second antibody, as in the Bio-Rad immunoblot goat antirabbit-horseradish peroxidase protocol.
50
·w 40 ~ a>
a.
* *
0
a>
10
o ·0
•
50
a. CJ)
40
'0 E Q)
30
o
'-
P = n.s.
•
P = n.s.
LPS
PMA
20
Q)
a. 5:1
10:1
20:1
E:T
ratio
10
P = n.s.
I ... ~•
40:1
Fig. 1. Cytotoxic activity against TNFa-sensitive U937 targets by unstimulated alveolar macrophages recovered from 11 patients with HIV-1 infection in an 18-h 51Crrelease assay. Each E:T point and the corresponding vertical bar represent the mean value and the SEM of the 11 determinations, respectively. The lower shaded areas represent cytotoxic activity exerted by resting alveolar macrophages isolated from five normal subjects ± SO. Asterisks indicate significance at p < 0.001.
Statistical Analysis Data are expressed as mean ± SD of the mean, and comparisons between values were made using the Cochran-Cox analysis (33). A p value < 0.05 wasconsidered as significant. Results
Evaluation of mRNA· Transcriptsfor TNFa in Alveolar Macrophages The expression of the TNFa gene by AMs was evaluated by Northern analysis according to Thomas's method (31).Total cellular RNA was isolated from AMs (stored at - 80° C at the time they wererecoveredfrom the BAL) by a guanidine isothiocyanate method and quantified by absorbance at 260 nm. Twenty micrograms of each sample were denatured at 60° C for 10 min in an electrophoresis buffer (20 mM morpholinopropane sulfonic acid, 15070 formaldehyde, 50070 formamide, 0.05 mg/ml ethidium bromide), size-fractionated electrophoresing through 1.5070 agarose gel containing 15070 formaldehyde then transferred to nitrocellulose filters. Filters were dried, baked at 80° C in a vacuum oven for 2 h, treated at 42° C for 6 h with a prehybridization solution (50070 formamide, 5X Denhardt's solution, 0.1070 SDS, 100 ug/ul denatured salmon sperm DNA), and hybridized at 42° C for 15 h in the same solution containing the 32p nick-translated probe. TNFa message wasdetected using a l.1-kb Pst-l fragment cDNA probe (32) (a kind gift from Dr. A. Rambaldi, Bergamo, Italy). Filters were washed twice in 2 X and twice in 0.1 X SSC {Vith 0.1070 SDS at 60° C. Filters were exposed for 1to 5 days at - 80° C to Kodak X-OMAT XAR-5 film.
60
Q)
L..
a.
~
:;::::
en
a>
P< 0.001
·00
30
'0 C 20
70
en
*
~
·0
8
*
en
The results of cell-mediated cytotoxicity of freshly recovered AMs are shown in figure 1. When AMs from patients with AIDS and ARC were tested for their lytic activity against TNFa-sensitive targets, high levels of spontaneous killing were observed at every tested E:T ratio (mean values, 10.9 ± 3.1, 17.5 ± 4.1, 22.3 ± 11.1, and 29.6 ± 11.4070 at the 5:1, 10:1, 20:1, and 40:1 E:T ratio, respectively). By contrast, unstimulated AMs from normal subjects showed no cytotoxic activity against TNFa-sensitive targets (p < 0.001). The results of the cytotoxic activity exerted by supernatants conditioned by AMs cultured for 48 h in medium alone or after stimulation with LPS, PMA, or IFN-y are shown in figure 2. Supernatants conditioned by unstimulated AMs recovered from HIV-l+ patients showed high levels of lytic activity against U937 targets (38.9 ± 14.4070 ).In vitro stimulation of HIV-l + AMs failed to increase their lytic activity' (36.9 ± 8.9, 27.1 ± 9.5, and 28.4 ± 10.9070 after LPS, PMA, and IFN-y stimulation, respectively). As for normal AMs, when cultured in me-
alone
IFNy
stimulus Fig. 2. Cytotoxic activity for TN Fa-sensitive U937 targets released by alveolar macrophages cultured for 48 h in medium alone or after stimulation with LPS, PMA, or IFN-y. Horizontal and vertical bars represent the patient's mean values and the corresponding SEM, respectively. Shaded areas represent cytotoxic activity exerted by supernatants obtained from 48-h-cultured alveolar macrophages isolated from five normal subjects ± SO.
dia alone, they were unable to release detectable amounts of cytotoxic factors (0.3 ± 0.2070; p < 0.001 compared with the corresponding value observed in seropositive patients). By contrast, after activation of normal AMs with LPS, PMA, and IFN-y, a definite release of cytotoxic factor was detected (37.8 ± 8.9, 32.3 ± 9.2, and 37.1 ± 8.1070, respectively); these values did not differ from the corresponding values found in patients with HIV-l +(p = NS). The cytotoxic activities exerted by supernatants after activation of AMs from patients with HIV-l + with a combination of LPS and IFNy were similar to that after activation with the individual agents (39.1 ± 8.9 after the combination of LPS and IFNy; p not significant with respect to the corresponding value obtained after stimulation with LPS alone or IFNy alone). To demonstrate that the above reported cytotoxic activities were mediated by TNFa, in duplicate experiments, a 1/100 diluted anti-TNFa antibody was added to both freshly recovered AMs and to the supernatants obtained from unstimulated AMs at the beginning of each test (figure 3). The addition of the anti-TNFa antibody to AMs recovered from patients with seropositive HIV-l (figure 3A) inhibited the cell-mediated lysis of U937 targets with an efficiency ranging between 100· and 24070 (mean, 75.9 ± 33.1070). Furthermore, the anti-TNFa an-
198
AGOSTINI, ZAMBELLO, TRENTIN, GARBISA, FRANCIA 01 CELLE, BULlAN, ONISTO, POLETTI, SPIGA, RAISE, FoA, AND SEMENZATO
E:T RATIO 40:1
I/)
50
50
40
40
Fig. 3. Inhibitory activity of anti-TNFa antibody,added at the beginning of each test, on the cytotoxic activity exerted by unstimulated alveolar macrophages (A) and by conditioned supernatants (B) from patients with seropositive HIV-1.
~
i o
'0
30
30
20
20
c: Q)
CARLO AGOSTINI, RENATO ZAMBELLO,4 LIVIO TRENTIN,5 SPIRIDIONE GARBISA, PAOLA FRANCIA 01 CELLE, PIETRO BULIAN, MAURIZIO ONISTO, VENERINO POLETTI, LUIGI SPIGA, ENZO RAISE, ROBERT FoA, and GIANPIETRO SEMENZATO
Introduction
Over the last few years evidence has been accumulating that HIV-l mayaffeet the cellular immune system in the lung of patients with AIDS (1, 2). Major immunologic abnormalities reported to date include an alveolitis, mainly characterized by cytotoxic/suppressor CD8+ lymphocytes and natural killer cells, a severe depletion of pulmonary helperrelated CD4+lymphocytes and, in some cases, an increase of polymorphonuclear cells and alveolar macrophages (AMs) (3-8). Because it is well known that AMs represent an important target for HIV-l (9-11),the study of their functional capabilities in vitro might help in understanding the pathogenesis of HIV-l-related pulmonary complications. In response to tissue damage and/or inflammation, AMs secrete a series of soluble molecules, termed cytokines, that promote activation of the lymphocyte population and mediate the killing of microbes and tumoral cells (12). Among these, great interest has been recently paid to tumor necrosis factor alpha/cachectin (TNFa), a molecule produced by cells of the host immune system, mainly by monocyte-macrophages, under conditions that resemble host invasion (13). The release of TNFa has been implicated in the pathogenesis of several human diseases (14-17),including AIDS (18).In particular, high levelsof TNFa have been found in the sera and supernatants of circulating monocytes and tissue macrophages from patients with HIV-l infection (19-24), even if data are not always consistent (25-27). The current study was designed to determine whether in patients with AIDS and ARC, AMs, i.e., the representatives of the monocytic-macrophagic lineage in the lower respiratory tract, are capable of synthesizing and releasing detectable
SUMMARY To verify the hypothesis that alveolar macrophages (AMs) from patients infected with HIV-1 could synthesize and release TNFa, AMs recovered from the BAL fluid of 11 patients with seropositive HIV-1 (six with AIDS and five with ARC) were tested in vitro for their ability to destroy TNFa-susceptible targets. Furthermore, the presence of TNFa was assessed in AM-conditioned supernatants on the basis of their cytotoxic activity and by using an immunoenzymatic test and immunoblotting. Transcription of the TNFa gene in AMs was also studied by means of the Northern blot analysis. AMs freshly recoveredfrom patients infected with HIV-1 exhibited high levels of cell-mediated cytotoxicity against U937 targets, and the addition of a polyclonal anti-TNFa antibody resulted in a significant inhibition of the target lysis. Cell-free supernatants conditioned by unstimulated AMs exerted high levels of cytotoxic activity against TNFa-sensitive targets, whereas duplicate, neutralization experiments performed in the presence of an anti-TNFa antibody proved that the observed cytotoxic activity was mostly mediated by TNFa. The presence of high amounts of TNFa in the conditioned media was confirmed by the immunoenzymatic test. In addition, the immunoblot analysis showed that the TNFa released by AMs has a Mr 17,000 band, identical to a standard preparation of recombinant TNFa. The Northern blot demonstrated that unstimulated AMs express detectable levels of mRNA transcripts for TNFa. Takentogether, our data support the concept that AMs from patients with HIV-1 infection constitutively release TNFa. Because normal human AMs do not express detectable levels of mRNA for TNFa or release this cytokine, the hypothesis is formulated that in the lung of patients with AIDS, AMs, either by a direct effect of HIV-1 infection or as a consequence of the pulmonary opportunistic infections, are triggered to synthesize TNFa. This in situ overproduction might playa role in the pathogenesis of AIDS-related pulmonary complications. AM REV RESPIR· DIS 1991; 144:195-201
amounts of TNFa. For this purpose, AMs recovered from the bronchoalveolar lavage (BAL) fluid of 11 patients with seropositive HIV-l were tested for their ability to destroy TNFa-susceptible cell lines, and the presence of TNFa was determined in the conditioned supernatants by cytotoxicity experiments, by an enzyme-linked immunosorbent assay, and by immunoblotting. The transcription of the TNFa gene in AMs was also studied by means of the Northern blot analysis. Methods Study Population Eleven patients with seropositive HIV-l (average age, 28 ± 8 yr; seven men and four women) were studied. All patients belonged to the IV group of the CDC classification (28), which includes patients with clinical symptoms and signs of HIV-l infection other than,
(Received in original form July 31, 1990 and in revised form February 26, 1991) 1 From the Padua University School of Medicine, Department of Clinical Medicine, 15t Medical Clinic, Clinical Immunology Branch, and Department of Histology; Department of Biomedical Sciences and Human Oncology, 15t Medical Clinic, University of Torino; Department ofPneumology,Bellaria Hospital Bologna and Department of Infectious Diseases, Maggiore Hospital Bologna, Padua, Italy. 2 Supported by grants from the Ministero della Sanita, Istituto Superiore di Sanita, Progetto AIDS 1989, 1990 (Rome, Italy), by Progetto Finalizzato CNR "Oncologia" 88.00680.44, and by the Italian Association for Cancer Research (AIRC, Milan). 3 Correspondence and requests for reprints should be addressed to Gianpietro Semenzato, M.D., Istituto di Medicina Clinica dell'Universita di Padova, Clinica Medica 1, Via Giustiniani 2, 35128 Padova, Italy. 4 Recipient of a Fellowship from the Ministero della Sanita, Istituto Superiore di Sanita (Rome). 5 Recipient of a Fellowship from CNR.
195
196
AGOSTINI, ZAMBELLO, TRENTIN, GARBISA, FRANCIA 01 CELLE, BULIAN, ONISTO, POLETTI, SPIGA, RAISE, FoA, AND SEMENZATO
or in addition to, lymphoadenopathy. According to the CDC criteria, five patients had ARC (IV-A) and six had AIDS (IV-Cl). BAL was performed to obtain a specific diagnosis of opportunistic pulmonary infection and/or AIDS-related tumors, according to our study protocol of patients with HIV-l and pulmonary involvement, which includes the use of BAL and of transbronchiallung biopsy (TLB) during the. fiberoptic bronchoscopy examination. According to the declaration of Helsinki, this invasiveprocedure is performed only for diagnostic purposes when a pulmonary complication is suspected; for this reason asymptomatic patients infected with HIV-l but without signs of pulmonary disease were not included-in the study. Two of the 11 patients weresmokers. The risk factors for AIDS included intravenous drug abuse (n = 9) and homosexuality (n = 2). Pneumocystis carinii infection was detected in the BAL or TLB of all six patients with AIDS. Two patients with ARC showed a histologic pattern consistent with aspecific chronic pneumonia on TLB, whereas the three remaining patients with ARC did not show detectable signs of pulmonary infections. None of the patients had a pulmonary localization of Kaposi's sarcoma. Five normal volunteers (two smokers and three nonsmokers; four men and one woman; average age, 29 ± 6 yr) were accepted as control subjects. They had a normal physical examination, chest radiograph, and lung function tests. As stated above, we were not allowed to perform BAL analysis on control subjects at risk for AIDS and HIV seronegative.
Preparation of Cell Suspensions and Purification of AMs BAL was performed after local anesthesia according to the method previously described (5). The percentages of lavage recovery were 51.5 ± 6.90/0 and 54.1 ± 4.70/0 of the injected fluid in patients with HIV-l and control subjects, respectively. The mean numbers of cells per milliliter of fluid recovered were 263 x 103 ± 70 and 141 x 103 ± 41 in patients with seropositive HIV-l and normal subjects, respectively. A differential count made by morphologic criteria on cytocentrifuged smears showed that 80.7 ± 5.1, 15.3 ± 4.7, 3.9 ± 3.1,and 0.1 ± 0.1% of BAL cells from patients with HIV-l were macrophages, lymphocytes, neutrophils, and eosinophils, whereas the corresponding values in normal subjects were 91.1 ± 3.1, 7.3 ± 1.0, 0.9 ± 0.2, and 0.7 ± 0.50/0, respectively. Cells recovered from the BAL were centrifuged on a Ficoll-Hypaque density gradient, washed, and then resuspended in RPMI 1640culture medium (Grand Island Biological Co., Paisley, Scotland, UK) supplemented with 2 IJM glutamine, 100 U/ml penicillin, 100 ug/ml streptomycin, and 200/0 FCS (Flow Laboratories, Irvine, Scotland, UK). Adherent cells were removed and isolated from the entire mononuclear cell suspension
by two sequential incubations at 37° C in a humidified atmosphere of 5 % CO 2 and 950/0 air for 60 min in plastic petri dishes, as previously reported (29). To avoid the presence of nonadherent cells on the populations under study, in two patients who showed high intensity lymphocytic alveolitis, CD3+cellswere removed using magnetic microspheres coated with goat anti-mouse IgG (Dynabeads; Dynal, Oslo, Norway), as already described (29); immediately after the removal of CD3+cells, the adherence procedure was performed as above. The population resulting from the adherence procedure was composed of more than 95 % macrophages, as determined by nonspecific esterase staining, and by less than 1% CD3+cells. More than 900/0 of the isolated macrophages were viable, as judged by trypan blue exclusion.
Cell-mediated Cytotoxic Assay Cell-mediated cytotoxicity was assessed by lysis of s1Cr-labeled target cells represented by the TNFa-susceptible U937 targets, as described elsewhere (30). Evaluation was performed using AMs depleted of lymphocytes and immediately used as effector cells. BrieflY,1 x 106 targets werelabeled for 4 h at 37° C in 50/0 CO 2 with 100 flCi Na (51Cr)04 (CEA IRE Sorin; Biomedica, Saluggia, Italy) and extensively washed before use. Then 100 IJI of the target cells suspension (x lOS Iml) were suspended in each well of a V-shaped plate (Tirtetek; Falcon, Becton Dickinson, Sunnyvale, CA), and graded concentrations of effector cells (5:1, 10:1, 20:1, and 40:1) were added to wells in triplicate and incubated at 37° C in 50/0 CO 2 for 18h. Supernatants were then harvested and counted in a gamma counter. The mean value of triplicate assays was used to calculate the percentage of cytotoxicity according to the following formula: (cpm effector cells - cpm spontaneous release) x 100/(cpm maximum release - cpm spontaneous release). Spontaneous release from the U937 targets was always less than 80/0. Because U937 targets are NK-sensitive cells, we determined by flow cytometry whether cells bearing the NK-related MAb CD16 and CD56 were present among the effector cell population. In every experiment less than 0.1% of the effector cells were CDI6+ or CD56+. To verify whether AMs cytotoxicity was mediated by TNFa, inhibition experiments were carried out, in duplicate tests, by adding 1/100 diluted polyclonal rabbit antihuman TNFa antibody (IP-300, a mixture of primarily IgG and IgM; Genzyme Co, Boston, MA) to the effector cells and U937 targets. In preliminary experiments we demonstrated that a 1/100 dilution of IP-3oo antibody inactivates approximately 20 ng of purified human TNFa. Neutralization experiments of cell-mediated cytotoxicity were performed at the 40:1 E:T ratio. Preliminaryexperiments showed that IP-3oo antibody alone had no effect on S1Cr release by targets.
Determination of Cytotoxic Activity Exerted by Supernatants Obtained from Unstimulated and Stimulated AMs Unstimulated AMs (5 x lOs/ml) resuspended in serum-free Iscove's modified Dulbecco's medium (IMDM) (Flow Laboratories) were cultured in 24-wellplates at 37° C in 50/0 CO 2 for 48 h. In separate experiments AMs were stimulated with IFN-y (100 U/ml; Behring Inst., Marburg, FRG), PMA (5 ng/ml), and LPS (10ug/ml; Difco Lab., Detroit, MI). After the incubation period, supernatants were harvested, filtered through a 0.45-flm Millipore filter (Millipore Corp., Bedford, MA), and immediately assayed for cytotoxicity or stored at - 80° C for further use. The lytic activity exerted by supernatants was assessed by measuring cytotoxicityagainst s1Cr-labeled U937 targets. In particular, 50 ul of target cells (1 x 106/ml) and 100 ul of cell-free supernatants containing the putative cytotoxic factor were combined in triplicate in 96-wellU-bottomed sterile microtiter plates (Falcon), and the final volume was adjusted to 200 ul with IMDM. After 20 h of incubation at 37° C in 50/0 CO 2 , 100 fll from each wellwere harvested, and the radioactivity was counted in a gamma counter. The percentage of cytotoxicity was calculated as follows: (cpm test - cpm spontaneous) x 100/(cpm maximum - cpm spontaneous). In all experiments the spontaneous release was less than 80/0. In duplicate neutralization experiments, an antihuman TNFa antibody (Genzyme Co.) was used to evaluate whether the cytotoxic activity exerted by supernatants was due to the presence of TNFa. The inhibitory activity of the anti-TNFa antibody was tested by mixing 1/100 of the antibody with the supernatants containing the putative TNFa, and then the assay was carried out as above. The characteristics of IP-300anti-TNFa antibody have been reported above.
Detection of TNFa in the Conditioned Supernatants TNFa levels were determined in the supernatants obtained from AMs cultured as above using a sandwich enzyme immunoassay (Biokine TNF; T Cell Science, Inc., Cambridge, MA). Briefly, TNFa available in the samples or standards binds to polystyrene microtiter wellspreviously incubated with an anti-TNFa coating antibody (lOP ul/well of a solution of 0.1 ml of anti-TNFa in 10 ml of PBScoating buffer). A horseradish-peroxidaseconjugated anti-TNFu antibody binds to the TNFa captured by the first/antibody and completes the sandwich. After washing to remove the unbound enzyme-conjugated anti-TNFa antibody, a substrate solution is added to the wells. The reaction is then stopped, and the adsorbance is determined at 490 nm. A standard curve was prepared by using different reference preparations containing known concentrations of TNFa. Results were expressed as nglml of conditioned supernatants. According to the manufacturer, the minimum
197i'·.v
TNFa PRODUCTION BY ALVEOLAR MACROPHAGES IN HIV-1 INFECTION
dose of TNFa detectable by the immunoassay is 4.8 pg/ml.
Polyacrylamide Gel Electrophoresis and Immunoblotting The TNFa present in conditioned supernatants obtained from 18-h-cultured AMs was further characterized by SDS-PAGE and immunoblotting. Conditioned media (500 Ill) from 24-well plates, obtained as reported above, and a recombinant standard TNFa preparation (Knoll AG, Ludwigshafen, Germany) were incubated with rabbit antihuman TNFa antibody (Genzyme Co) at final concentration of 1:200; 50 III of antirabbit IgG (H and L chain-specific) chromatography gel (Cappel Laboratories, Cochranville, PA) were then added and incubated under rolling for 1 h at room temperature. After centrifugation for 5 min at 200 g, the precipitate was resuspended in sample buffer, and the SDSPAGE was performed using a TRIS-glycine buffer system with 0.5 M urea and 12070 separating and 5070 stacking gels. Transfer of proteins from polyacrylamide gels to Hybond-c extra (Amersham Co, Buckinghamshire, UK) was performed using 12 mM TRIS and 100 mM glycine buffer containing 10070 methanol with a Bio-Rad Trans-blot Cell (Bio-Rad Laboratories, Richmond, CA) for 2 h at 0.2 ampere. The transfer efficiency was verified by staining the gels after transfer with Coommassie blue. The Hybond-c extra was then saturated with 3070 gelatin solution, and treated with polyclonal rabbit antihuman TNFa antibody 1/200 (Genzyme Co) and a second antibody, as in the Bio-Rad immunoblot goat antirabbit-horseradish peroxidase protocol.
50
·w 40 ~ a>
a.
* *
0
a>
10
o ·0
•
50
a. CJ)
40
'0 E Q)
30
o
'-
P = n.s.
•
P = n.s.
LPS
PMA
20
Q)
a. 5:1
10:1
20:1
E:T
ratio
10
P = n.s.
I ... ~•
40:1
Fig. 1. Cytotoxic activity against TNFa-sensitive U937 targets by unstimulated alveolar macrophages recovered from 11 patients with HIV-1 infection in an 18-h 51Crrelease assay. Each E:T point and the corresponding vertical bar represent the mean value and the SEM of the 11 determinations, respectively. The lower shaded areas represent cytotoxic activity exerted by resting alveolar macrophages isolated from five normal subjects ± SO. Asterisks indicate significance at p < 0.001.
Statistical Analysis Data are expressed as mean ± SD of the mean, and comparisons between values were made using the Cochran-Cox analysis (33). A p value < 0.05 wasconsidered as significant. Results
Evaluation of mRNA· Transcriptsfor TNFa in Alveolar Macrophages The expression of the TNFa gene by AMs was evaluated by Northern analysis according to Thomas's method (31).Total cellular RNA was isolated from AMs (stored at - 80° C at the time they wererecoveredfrom the BAL) by a guanidine isothiocyanate method and quantified by absorbance at 260 nm. Twenty micrograms of each sample were denatured at 60° C for 10 min in an electrophoresis buffer (20 mM morpholinopropane sulfonic acid, 15070 formaldehyde, 50070 formamide, 0.05 mg/ml ethidium bromide), size-fractionated electrophoresing through 1.5070 agarose gel containing 15070 formaldehyde then transferred to nitrocellulose filters. Filters were dried, baked at 80° C in a vacuum oven for 2 h, treated at 42° C for 6 h with a prehybridization solution (50070 formamide, 5X Denhardt's solution, 0.1070 SDS, 100 ug/ul denatured salmon sperm DNA), and hybridized at 42° C for 15 h in the same solution containing the 32p nick-translated probe. TNFa message wasdetected using a l.1-kb Pst-l fragment cDNA probe (32) (a kind gift from Dr. A. Rambaldi, Bergamo, Italy). Filters were washed twice in 2 X and twice in 0.1 X SSC {Vith 0.1070 SDS at 60° C. Filters were exposed for 1to 5 days at - 80° C to Kodak X-OMAT XAR-5 film.
60
Q)
L..
a.
~
:;::::
en
a>
P< 0.001
·00
30
'0 C 20
70
en
*
~
·0
8
*
en
The results of cell-mediated cytotoxicity of freshly recovered AMs are shown in figure 1. When AMs from patients with AIDS and ARC were tested for their lytic activity against TNFa-sensitive targets, high levels of spontaneous killing were observed at every tested E:T ratio (mean values, 10.9 ± 3.1, 17.5 ± 4.1, 22.3 ± 11.1, and 29.6 ± 11.4070 at the 5:1, 10:1, 20:1, and 40:1 E:T ratio, respectively). By contrast, unstimulated AMs from normal subjects showed no cytotoxic activity against TNFa-sensitive targets (p < 0.001). The results of the cytotoxic activity exerted by supernatants conditioned by AMs cultured for 48 h in medium alone or after stimulation with LPS, PMA, or IFN-y are shown in figure 2. Supernatants conditioned by unstimulated AMs recovered from HIV-l+ patients showed high levels of lytic activity against U937 targets (38.9 ± 14.4070 ).In vitro stimulation of HIV-l + AMs failed to increase their lytic activity' (36.9 ± 8.9, 27.1 ± 9.5, and 28.4 ± 10.9070 after LPS, PMA, and IFN-y stimulation, respectively). As for normal AMs, when cultured in me-
alone
IFNy
stimulus Fig. 2. Cytotoxic activity for TN Fa-sensitive U937 targets released by alveolar macrophages cultured for 48 h in medium alone or after stimulation with LPS, PMA, or IFN-y. Horizontal and vertical bars represent the patient's mean values and the corresponding SEM, respectively. Shaded areas represent cytotoxic activity exerted by supernatants obtained from 48-h-cultured alveolar macrophages isolated from five normal subjects ± SO.
dia alone, they were unable to release detectable amounts of cytotoxic factors (0.3 ± 0.2070; p < 0.001 compared with the corresponding value observed in seropositive patients). By contrast, after activation of normal AMs with LPS, PMA, and IFN-y, a definite release of cytotoxic factor was detected (37.8 ± 8.9, 32.3 ± 9.2, and 37.1 ± 8.1070, respectively); these values did not differ from the corresponding values found in patients with HIV-l +(p = NS). The cytotoxic activities exerted by supernatants after activation of AMs from patients with HIV-l + with a combination of LPS and IFNy were similar to that after activation with the individual agents (39.1 ± 8.9 after the combination of LPS and IFNy; p not significant with respect to the corresponding value obtained after stimulation with LPS alone or IFNy alone). To demonstrate that the above reported cytotoxic activities were mediated by TNFa, in duplicate experiments, a 1/100 diluted anti-TNFa antibody was added to both freshly recovered AMs and to the supernatants obtained from unstimulated AMs at the beginning of each test (figure 3). The addition of the anti-TNFa antibody to AMs recovered from patients with seropositive HIV-l (figure 3A) inhibited the cell-mediated lysis of U937 targets with an efficiency ranging between 100· and 24070 (mean, 75.9 ± 33.1070). Furthermore, the anti-TNFa an-
198
AGOSTINI, ZAMBELLO, TRENTIN, GARBISA, FRANCIA 01 CELLE, BULlAN, ONISTO, POLETTI, SPIGA, RAISE, FoA, AND SEMENZATO
E:T RATIO 40:1
I/)
50
50
40
40
Fig. 3. Inhibitory activity of anti-TNFa antibody,added at the beginning of each test, on the cytotoxic activity exerted by unstimulated alveolar macrophages (A) and by conditioned supernatants (B) from patients with seropositive HIV-1.
~
i o
'0
30
30
20
20
c: Q)
