Leaky lysosomes in lung transplant macrophages: azithromycin prevents oxidative damage

The study protocol was approved by the local Ethical Committee (Linköping, Sweden) according to the guidelines of the Helsinki Declaration.

Following informed consent, bronchoscopies with bronchoalveolar lavage (BAL) were carried out. The control group was 7 healthy, non-smoking subjects [in all cases pulmonary disease was thoroughly ruled out by a chest X-ray, lung function test and bronchoscopy]. The lung transplant recipients were 5 patients without AZM treatment who underwent surveillance post-transplantation BAL with transbronchial lung biopsies at 3 and 6 months and 3 patients with AZM treatment (250 mg three times per week) who were investigated to rule out rejection and/or infection. Transplant recipients with AZM were transplanted 11, 6 and 2 yrs before bronchoscopy and were treated with AZM for 42, 1 and 2 months, respectively. All transplant recipients received conventional triple-drug immunosuppression [with methylprednisolone, a calcineurin inhibitor (cyclosporine A or tacrolimus) and a cytostatic agent (mycophenolate mofetil)], conventional infectious prophylaxis [for cytomegalovirus, Aspergillus spp. and Pneumocystis spp.] and conventional prophylaxis for gastro-esophageal reflux (with the proton pump inhibitor omeprazol). None of the transplant recipients were diagnosed with infection and/or rejection when the bronchoscopy was performed. The demographic details of the subjects and the BALF characteristics are presented in Table 1.

Briefly, BAL was performed by a standardized washing of the middle lobe or lingula of the lung (transplant) for 6–9 times with 20 mL of sterile 0.9% (w/v) saline solution during a fibreoptic bronchoscopy [12]. None of the BAL fluid (BALF) was stained with blood. The BALF samples were centrifuged at 200xg for 10 min. Leucocytes from the BALF were counted and seeded in 35 mm Petri dishes (± cover slips) and the lung macrophages were allowed to attach. Concomitantly, murine J774 macrophages were seeded at the same density. After a thorough rinse in phosphate-buffered saline (PBS), the dishes, which contained approximately 0.4 x 10⁶ attached lung macrophages or J774 cells/dish, were returned to standard culture conditions. Both cell types were cultured in Dulbecco's Modified Eagle’s Medium supplemented with 100 IU/ml penicillin, 100 μg/ml streptomycin, 0.25 μg/ml amphotericin B (all from GIBCO, Paisley, UK) and 10% fetal bovine serum (PAA Laboratories GmbH, Pasching, Austria) under standard culture conditions for 48 h. Before the experiments, all dishes were thoroughly rinsed in PBS. J774 macrophages were pre-treated (or not) as described below and then rinsed again. Both cell types were oxidatively stressed by glucose oxidase (GO; Sigma-Aldrich Inc., St. Louis, MO, USA). GO was added directly to the culture medium for 60 min under standard culture conditions. Under these conditions, a stable concentration of approximately 120 μM H₂O₂ was generated in the culture medium.

By employing J774 macrophages, the protection afforded by AZM on oxidant-challenged lysosomes and cells were further explored. For this purpose we used two reference compounds, both well known to attenuate the reactivity of lysosomal Fe. Thus, the pre-treatments were with AZM at 33–660 μM for 45 min to 4 hrs, H-DFO (which is a strong Fe chelator that exclusively targets the lysosomal compartment) at 2 mM for 4 hrs [30], or NH₄Cl at 10 mM for 4 hrs, which blocked the enzymatic liberation of Fe in lysosomes by raising the pH to reduce the amount of available lysosomal Fe [29]. Ampoules of AZM (Pfizer Inc., Sweden) for intravenous use were reconstituted with distilled water to 100 mg/mL and diluted to the required concentrations with distilled water. The appropriate concentrations and exposure time that were utilized for NH₄Cl and H-DFO were based on previous studies on J774 macrophages, which demonstrated optimal reduction of lysosomal Fe reactivity and maintained viability [29, 30].

To confirm the lysosomotropism of AZM, J774 macrophages were pretreated with 10 mM NH₄Cl for 5 min (which raised the lysosomal pH and largely prevented the lysosomal uptake of AZM) and then exposed to AZM in the continued presence of NH₄Cl. Following its dissociation to NH₃, H⁺ and Cl^-, NH₃ (pKa = 9.2) becomes protonated and trapped in the lysosome as NH₄⁺, which substantially increases the intralysosomal pH [31] and prevents the accumulation of lysosomotropic substances with lower pKa values, such as AZM [27].

The cells grown on cover slips were fixed in paraformaldehyde and stained for ferric Fe (Fe³⁺) by the Prussian blue staining procedure [11]. The cellular content of Fe³⁺ was scored independently by two examiners according to the method reported by Golde et al.[32]. Thus, 200 lung macrophages/cover slip were counted, and each macrophage was graded on a scale of 0–4; 0 = no blue color, 1 = faint blue staining in cytoplasm, 2 = dense blue color in minor portion of cytoplasm or medium color intensity throughout cell, 3 = deep blue staining in most of cytoplasm, 4 = dark blue throughout cytoplasm. A mean score, i.e. the Golde Index, for 100 macrophages was then calculated; zero being the minimum and 400 the maximum score.

The total cellular Fe was estimated using atomic absorption spectrophotometry, equipped with an iron lamp (243.3 nm), with a lower detection limit of 0.65 μg Fe/L [11]. The cellular expression of the ferritin heavy (H-ferritin) chain in cell lysates was determined by an ELISA analysis (MyBioSource, CA, USA) according to the manufacturer's instructions. Reduced glutathione (GSH) was estimated by the Glutathione Assay Kit (BioCat GmbH, Heidelberg, Germany) using a Wallace 1420 Victor Plate Reader (PerkinElmer, Waltham, MA, USA). The Fe, H-ferritin and GSH contents were normalized to the protein concentration of each sample [11].

Acridine orange (AO; Gurr, Poole, Dorset, UK), which is a metachromatic fluorophore and lysosomotropic weak base (pKa of ~ 10), is retained in its charged form (AOH⁺) inside acidic lysosomes. Using flow cytometry, the lysosomal leakage of AO (and protons) in to the cytosol of lung macrophages and J774 cells was assessed by determination of the mean value of AO-green fluorescence (the AO-relocation test) [27]. Upon blue light excitation, the AO that is bound to proteins and DNA in the cytosol and nuclei emits a weak green fluorescence, which increases when the lysosomes leak AO. Alternatively, the AO is highly concentrated in lysosomes and emits an intense red fluorescence upon excitation. The AO-relocation test is sensitive and able to monitor early and minor LMP. The method was thoroughly standardized for lung macrophages using identical settings and J774 macrophages, equally exposed to oxidants in parallel experiments, were used as a reference. Briefly, the cells were rinsed in PBS, stained with AO (2.5 μg/mL) for 15 min at 37°C, rinsed 3 times with complete culture medium, oxidatively stressed in the presence of GO and rinsed with culture medium before the analysis was performed immediately after the 1-h oxidant exposure. To initiate controlled LMP the synthetic lysosomotropic detergent O-methyl-serine dodecylamide hydrochloride (MSDH; kindly provided by Gene M. Dubowchik, Bristol-Myers Squibb, Wallingford, CT, USA) was used. The green (FL1 = 530 nm) fluorescence was recorded on log scale using a BD LSR Flow Cytometer (Becton-Dickinson, Mountain View, CA, USA) that was equipped with a 488-nm exciting argon laser. CellQuest software was used for data acquisition and analyses. Data are expressed as arbitrary units (A.U.).

Micrographs of J774 macrophages exposed to MSDH were also taken. Briefly, cells seeded on cover-slips were incubated with AO (2.5 μg/mL) for 15 min at 37°C, washed with PBS, and placed on the stand of a LMS Zeiss laser scanning confocal microscope. AO was excited using a 488 nm light from a 100-mW diode laser, and leakage of AO (and concomitant loss of the lysosomal proton gradient) during the MSDH exposure was followed by laser scanning micrographs in a channel defined by band-pass filters for 495–555 nm and 630 nm.

The frequency of cells with an apoptotic or necrotic morphology (i.e., cytoplasmic budding/pycnotic fragmented nuclei/apoptotic bodies and membranous rupture/nuclear swelling, respectively) was estimated by phase contrast microscopy in a blinded fashion. 0.5 × 10³ cells/dish were counted in preselected fields. Ten hrs after ended oxidative stress, the fractions of cells (apoptotic or necrotic)/initial numbers of cells within the fields were determined. Detached cells were all necrotic. At the same time point the fraction of fragmented apoptotic DNA was assessed using hypotonic propidium iodide (Sigma Chemical Co., St. Louis, MO, USA) and cytofluorimetric analysis [11]. According to the manufacturer’s instructions, the fluorescence of AMC (7-amino-4-methyl-coumarin), which was liberated from Ac-DEVD-AMC (Becton-Dickinson, Mountain View, CA, USA) by active caspase-3-like caspases, was also analyzed in lung macrophages using a Wallace 1420 Victor Plate Reader (PerkinElmer, Waltham, MA, USA) during the 1-h oxidant challenge and up to 5 h after the end of the oxidative stress. The increase of caspase-3 as a percent of the control was determined, and the peak value was recorded.

The baseline green AO fluorescence values of lung macrophages from lung transplant recipients without AZM were significantly higher than those of the lung macrophages from healthy subjects. This observation might indicate that lysosomes in the lung transplant macrophages were more prone to leakage from the start and/or that the content of AO-binding proteins and DNA was higher than in the macrophages from healthy subjects (Figure 1A). However, following oxidative stress, macrophages from transplant recipients without AZM exhibited significantly higher green AO fluorescence values that corresponded to more lysosomal leakage of AO compared to the macrophages from healthy subjects (Figure 1A). Indeed, upon an oxidant challenge, the increase of green fluorescence was 44 ± 18 A.U. in the transplant macrophages, but only 19 ± 16 A.U. in the lung macrophages from healthy subjects (p < 0.05). In contrast, the macrophages from lung transplant recipients with AZM displayed a slight decrease of AO green fluorescence (−6 ± 6) upon oxidant challenge, which was significantly less than the increase of AO green fluorescence observed in the macrophages from the healthy subjects (p < 0.01) (Figure 1A).

Compared to cultures of lung macrophages from the healthy subjects, cultures of macrophages from transplant recipients without AZM displayed significantly more dead cells, including both apoptotic (p < 0.01) and necrotic cells (p < 0.05) (Figure 1B). In cultures of macrophages from transplant recipients with AZM apoptosis and necrosis were significantly less than in cultures of macrophages from transplant recipients without AZM (p < 0.01 and p < 0.05, respectively) (Figure 1B).

Apoptosis is often associated with increased activity of caspase-3. Correspondingly, oxidant-challenged macrophages from transplant recipients without AZM expressed high peak values of caspase-3 activation (321;% of controls). However, apoptosis may also occur without caspase-3 activation [33]. This possibility and significant necrosis might explain the great variation found (1 SD = 571%). Consequently, the caspase-3 activation observed in oxidatively stressed macrophages from recipients without AZM this did not statistically differ from the caspase-3 activation observed in macrophages from healthy control subjects (45 ± 54; mean value ± 1 SD;% of controls). This analysis was not performed on macrophages from transplant recipients with AZM because of limited sample size.

Cell features that are decisive for lysosomal Fe reactivity were evaluated in the lung macrophages and the results are summarized in Table 2. The mean value for the Golde index, which is based on a cytochemical staining of ferric Fe, of macrophages that were retrieved from transplants recipients without and with AZM were significantly increased compared to that of the healthy control subjects (55.8 ± 37.2 (p < 0.05), 71.7 ± 15.1 (p < 0.05) and 8.2 ± 11.9, respectively). The Golde index normally ranges from 4 to 25 [34]. Figure 2 illustrates representative lung macrophages from healthy subjects and lung transplant recipients. Total cellular Fe assessed by atomic absorption demonstrated significantly more Fe in macrophages from transplant recipients without AZM (p < 0.05) and with AZM (p < 0.05) than in macrophages from healthy subjects. Indeed, there was even significantly more Fe in macrophages from transplant recipients with AZM than in macrophages from transplant recipients without AZM (p < 0.05). The cellular levels of the major antioxidant GSH in the macrophages from transplant recipients without AZM and healthy subjects were similar, while GSH was significantly decreased in macrophages from transplant recipients with AZM (p < 0.05; compared to the macrophages from healthy subjects). Compared to healthy subjects, the level of the cytoprotective H-ferritin in the macrophages from transplant recipients without AZM was slightly but not significantly increased. This assay was not performed on macrophages from transplant recipients with AZM due to limited sample size.

A 4-h exposure to a range of AZM concentrations (33–660 μM) effectively protected lysosomes in J774 macrophages against an oxidant challenge (data not presented). As opposed to 4-h exposures to millimolar concentrations of H-DFO or NH₄Cl, even a short exposure time (45 min) to a low concentration (33 μM) of AZM proved to be equally efficient (Figure 3A). Importantly, oxidant-induced LMP was prevented by micromolar concentrations of AZM, while millimolar doses of H-DFO or NH₄Cl for 4 h were needed to achieve full protection of the lysosomes against oxidant injury (Figure 3A).

In control experiments, we observed that a 1-h exposure to 50 and 100 μM of the lysosomotropic detergent MSDH resulted in a significant increase of AO-green fluorescence (p < 0.05 and p < 0.05, respectively), which was dose-dependent (p < 0.05) (Figure 3A). Figure 3B and 3C are detailed micrographs of J774 macrophages before (Figure 3B) and after exposure to 200 μM MSDH (Figure 3C). Within a few minutes MSDH caused extensive lysosomal disruption resulting in a massive leakage of AO into the cytosol (Figure 3C).

We also observed that AZM, H-DFO and NH₄Cl caused an effective prevention of oxidant-induced cell death (Figure 4). To test whether or not the protection of J774 macrophages by AZM was due to the accumulation of this drug inside the lysosomes in a pH-dependent way, cultures were pretreated with NH₄Cl before being exposed to AZM in the continued presence of NH₄Cl. The normal pH inside the macrophage lysosomes is about 4.5, but upon addition of 10 mM NH₄Cl, the lysosomal pH increases almost instantly. Under such circumstances, the cytoprotective effect by AZM was abolished (Figure 4).

Finally, we employed J774 macrophages to test whether or not AZM, H-DFO or NH₄Cl had an influence on the cell features that are important for lysosomal Fe-reactivity, which are mainly Fe and the GSH and H-ferritin contents of the cells. However, such an interference of significance was thoroughly ruled out, and the results are summarized in Table 3.