Background
ABCA3 is a member of the
ATP-
binding
cassette (ABC) family of transporters which utilize the energy of ATP hydrolyses to drive the transport of a variety of substrates across biological membranes [
1]. The
ABCA3 gene is highly expressed in alveolar epithelial type II cells (AECII) of the lung [
2,
3]. In AECII ABCA3 protein localizes to the limiting membrane of lamellar bodies as lipid-rich organelles for production, storage and secretion of pulmonary surfactant [
4,
5]. Surfactant is a complex mixture of 90% lipids (mostly phospholipids) and 10% surfactant-specific proteins (e.g. small hydrophobic proteins SP-B and SP-C) produced by AECII which reduces surface tension on the air-liquid interface and prevents alveolar collapse at the end of expiration. ABCA3 is a lipid transporter which transports surfactant phospholipids into lamellar bodies where surfactant is assembled. It is essential for lamellar body biogenesis [
6‐
9] so patients with
ABCA3 mutations and
Abca3 knock-out mouse have distinctive dense inclusions within immature lamellar bodies and disturbed composition of surfactant phospholipids [
7,
8,
10‐
13]. Since the last steps in SP-B and SP-C processing occur inside of functional lamellar bodies, ABCA3 deficiency in human and mouse leads to accumulation of SP-B and SP-C precursors [
7,
8,
14,
15].
In 2004 mutations of the
ABCA3 gene were recognized as a cause of lung diseases in full-term neonates with fatal pulmonary surfactant deficiency [
10]. Today
ABCA3 mutations are known to cause also chronic interstitial lung disease (ILD) in children and older patients [
12,
16,
17]. With more than 100 identified mutations, ABCA3 is the most frequent known cause of genetic ILD (own unpublished data), [
12,
16,
17]. Similar to SP-C deficiency, ABCA3-related ILD is complex and heterogeneous in histopathology and symptom severity. The disease onset varies from directly after birth, early in infancy or later in childhood [
10,
12,
16,
18,
19], sometimes following the exposure to environmental stressors such as cigarette smoke [
12,
16].
ABCA3 mutations classify either as functional defects of properly localized proteins or trafficking/folding defects where misfolded proteins accumulate in the ER [
6,
20]. Folding of newly synthesized proteins is a highly controlled process happening in the ER lumen with assistance of molecular chaperones. Proteins which fail to fold properly are harmful for the cell and retained inside the ER by the ER quality control. ER accumulation of misfolded proteins causes ER stress and activates cytoprotective mechanisms named unfolded protein response (UPR). UPR promotes the ER protein folding capacity by increasing the production of molecular chaperones and attenuates general protein translation to decrease the misfolded protein load in the ER [
21]. If UPR fails to resolve ER stress and restore cell homeostasis, the cell will be eliminated by initiation of tightly controlled apoptotic cell-death pathways [
22].
Recent data show that ER stress and apoptosis of AECII play a role in lung disease, especially in pathogenesis of idiopathic pulmonary fibrosis (IPF) and genetic SP-C-associated pulmonary fibrosis [
23‐
26]. Moreover, cultured lung epithelial A549 cells expressing SP-C mutations, which cause misfolding and aggregation of the SP-C pre-protein, increase ER stress, activate UPR and initiate apoptotic cell-death [
26‐
28]. Fibrosis is one of the hallmarks documented in ABCA3-associated ILD [
12,
16,
17] and knowing that
ABCA3 mutations can cause ER retention of the mutated transporter [
6,
20], we investigated the influence of three
ABCA3 mutations, R43L, R280C and L101P, found in children with surfactant deficiency and chronic ILD [
10,
14,
19], on ER stress and apoptosis induction in lung epithelial A549 cells.
Methods
Plasmid Vectors
Wild type (WT) full length human hABCA3 cDNA without stop codon was cloned into EcoRI/AgeI sites of pEYFP-N1 plasmid (Clontech, Mountain View, CA) to obtain pEYFP-N1/WT vector for expression of C-terminal ABCA3-YFP protein fusions. pUB6-HA/WT vector for expression of C-terminal fusions of ABCA3 with hemagglutinin tag (HA-tag) was produced by modification of pUB6/V5-His vector (Invitrogen, Karlsruhe, Germany). His-tag was put out of frame and WT hABCA3 cDNA with 3' HA-tag sequence (5'-TAC CCA TAC GAT GTT CCA GAT TAC GCT-3') was cloned into KpnI/XhoI restriction sites. Three hABCA3 point mutations R43L, R280C and L101P were introduced in the WT ABCA3 in both vector types by PCR-based site-directed mutagenesis (QuickChange Site-Directed Mutagenesis, Stratagene, La Jolla, CA). Mutagenesis primers were as follows: R43L-For 5'-CAT CTG GCT CCTCTT GAA GAT TC-3', R43L-Rev 5'-GAA TCT TCA AGAGGA GCC AGA TG-3', L101P-For 5'-CAG TGC GCA GGG CAC CTG TGA TCA AC-3', L101P-Rev 5'-GTT GAT CAC AGG TGC CCT GCG CAC TG-3', R280C-For 5'-CAT TGC CTG TGC TGT CGT G-3', R280C-Rev 5'-CAC GAC AGC ACAGGC AAT G-3'. Successful mutagenesis in pEYFP-N1/ABCA3 and pUB6-HA/ABCA3 (ABCA3 denotes WT or one of the three mutations) was confirmed by sequencing.
Cell Culture and Transfection
Human lung carcinoma epithelial cell line A549 (ACC 107) was obtained from the German Collection of Microorganisms and Cell Cultures (DSMZ, Braunschweig, Germany). Cells were grown in RPMI 1640 medium with 10% FBS. At 70% confluence cells were transiently transfected with pEYFP-N1/ABCA3 or pUB6-HA/ABCA3 vectors using ExGen 500 (Fermentas, Burlington, ON) according to the manufacturer's protocol. 48 h after transfection samples were collected for experiments. Transfection efficiency was confirmed in all samples through YFP signal or immunofluorescence of HA-tag (see bellow). ABCA3 construct expression was similar in all transfected cells as assessed by ABCA3 mRNA and protein expression levels. ABCA3 mRNA expression after transfection strongly increased (>30-fold) compared to non-transfected A549 cells (data not shown). When indicated transfected cells were incubated 14 h with 10 μg/ml of tunicamycin (Sigma, St. Louis, MO) or 16 h with 25 ng/ml of TNFα (Invitrogen) prior to sample collection.
Immunofluorescence
Cells grown on coverslips were transfected with pEYFP-N1/ABCA3 or pUB6-HA/ABCA3 vectors, washed in PBS buffer, fixed in 4% paraformaldehyde, permeabilised and incubated 1 h at room temperature with the following primary antibodies: monoclonal mouse anti-LAMP3/CD63 (1:200, Chemicon, Tamecula, CA), polyclonal goat anti-calnexin (1:200, Santa Cruz Biotechnology, Santa Cruz, CA) and to detect HA-tagged ABCA3 monoclonal rat anti-HA-tag (1:200, Roche, Manheim, Germany). Signals were visualized with Alexa Fluor 555 anti-goat or Alexa Fluor 555 anti-mouse IgG secondary antibodies (1:200, Invitrogen). Alexa Fluor 555 and YFP fluorescence of mounted samples was examined with Axiovert 135 fluorescent microscope and evaluated with AxioVision 4.7.1 software (Carl Zeiss, Jena, Germany).
Immunoblotting
Cells were collected and resuspended in lysis buffer (0.15 M NaCl, 1% Triton-X100, 0.5% sodium deoxycholate, 50 mM Tris, 5 mM EDTA) supplied with protease inhibitor cocktail (Complete; Roche). 30 μg of total protein was transferred to a PVDF membrane and immunoblotted with the following primary antibodies: monoclonal mouse anti-GFP (1:500, Clontech), monoclonal rabbit anti-BiP/Grp78 (1:1000, Cell Signaling, Frankfurt, Germany), monoclonal mouse anti-caspase 4 (1:500, Stressgene), and monoclonal mouse anti-β-actin HRP conjugate (1:10000, Santa Cruz). Signal was detected using chemiluminiscent labeling with Amersham ECL Detection Reagents (GE Healthcare, Buckinghamshire, UK).
Membrane Isolation and Deglycosylation Assay
For crude membrane preparation cells transfected with pEYFP-N1/ABCA3 vectors were collected in PBS supplemented with 1 mM EDTA and protease inhibitor (Complete, Roche). Cells were broken in a Potter-Elvehjem homogenizer and subsequently sonicated (Branson Digital Sonifier S450D) while keeping the samples on ice. Samples were centrifuged at 1000 × g, 10 min, 4 °C and obtained postnuclear supernatants were centrifuged at 100 000 × g, 1 h, 4 °C. Membrane pellets were resuspended in 25 mM Hepes/NaOH, pH 7.0 with protease inhibitor and stored at -80°C. For the deglycosylation assay 5 μg of membranes were incubated 1 h at 37 °C with PNGaseF or EndoH (New England Biolabs, Ipswich, MA). The samples were separated on 3-8% SDS-PAGE gels (NuPAGE; Invitrogen) and immunoblotted with monoclonal mouse anti-GFP antibody (1:500, Clontech).
Liposome Preparation and NBD-Lipid Uptake
NBD-labeled lipids phosphatidylcholine (C
12-NBD-PC) and phosphatidylethanolamine (C
12-NBD-PE) were purchased from Avanti Polar Lipids (Alabaster, AL), other lipids from Sigma. Liposome preparation and NBD-lipid uptake were performed following previously published protocols [
6]. Liposomes (~100 nm) containing C
12-NBD-PC or C
12-NBD-PE were prepared by mixing lipids in chloroform in the following molar ratios: L-αDPPC: C
12-NBD-PC: egg PC: egg PG: cholesterol = 5:5:5:3:2 and L-αDPPC: egg PC: egg PG: cholesterol: C
12-NBD-PE = 10:5:3:2:2. Cells grown on coverslips were transfected with pUB6-HA vectors expressing ABCA3 WT or one of the three mutations. 48 h later cells were incubated for 2 h with liposomes so that the final concentration of C
12-NBD-PC or C
12-NBD-PE in the medium was 150 μM. Cells were washed three times with PBS and prepared for immunofluorescence with primary monoclonal rat anti-HA-tag antibody (1:200, Roche) and secondary anti-rat IgG Alexa Fluor 555 antibody (1:200, Invitrogen). NBD-lipid uptake was examined with an Olympus FluoView FV 1000 confocal microscope.
RT-PCR and XBP1 Splicing
A549 cells were transfected with YFP vectors and where indicated treated 14 h with 10 μg/ml of tunicamycin (Sigma). 48 h after transfection total RNA was isolated (High Pure RNA Isolation Kit, Roche) and cDNA was synthesized (SuperScript III First-Strand Synthesis System, Invitrogen). For every RNA sample a control reaction without reverse transcriptase was performed to exclude genomic DNA contamination. cDNA was further used as a template for PCR with XBP1 and 18S rRNA primers as published in [
29]. For easier evaluation half of the PCR product was digested with
PstI endonuclease cutting only inside of the 26 nt intron removed from the
XBP1 mRNA by splicing. Cut and uncut PCR products were analyzed on a 3% agarose gel. Results were calculated as the ratio of the spliced (s) and unspliced (u) band (s/u; without
PstI digestion) and additionally confirmed by calculating the ratio of the spliced band and the sum of the two
PstI digest bands (s/(u1+u2), with
PstI digestion). Hybrid band (h) was considered as equally contributing to unspliced and spliced bands.
FACS Analyses
Cells transfected with pEYFP-N1/ABCA3 vectors were gated in the FL-1 channel (YFP-positive population of cells) and apoptosis was determined using different approaches. Early apoptosis was assayed by 1) measuring intracellular glutathione (GSH) levels with the cell-permeable monochlorobimane (Sigma) method [
30] and 2) by annexin V
+/propidium iodide (PI)
- staining (Cy5-conjugated anti-annexin V and PI; BD Bioscience, Heidelberg, Germany), and late apoptosis was determined via intracellular active caspase 3 levels (PE-conjugated anti-active-caspase 3; BD Bioscience) in cells permeabilized with the IntraPrep kit (Beckman Coulter, Krefeld, Germany) according to the manufacturer's protocol. To exclude unspecific binding isotype controls for caspase 3 and negative controls for glutathione and annexin V/PI staining were applied. BD FACSCanto II Flow Cytometer was used for the assay and FACSDiva v6.1.3 for data analyses (BD Bioscience).
Statistical Analyses
Statistical analysis was performed by one-way ANOVA and Bonferroni's test using GraphPad Prism version 4.0 (GrapPad Software Inc., San Diego, CA). All results were presented as means ± SEM of minimum four experiments and p-values < 0.05 were considered significant.
Discussion
ABCA3 mutations cause surfactant deficiency and fatal respiratory distress syndrome in full-term neonates [
10] and chronic ILD of children [
12,
16,
17]. The cellular pathomechanisms of ABCA3-related chronic ILD are probably complex including influence of
ABCA3 mutations on surfactant homeostasis but possibly also on the fitness and function of AECII as cells for surfactant production, lung repair and immunological defense [
38].
In this study we investigated the influence of three
ABCA3 mutations, R43L, R280C and L101P, on intracellular stress and induction of apoptosis in cultured lung epithelial A549 cells. All three mutations were found in children with ABCA3-associated lung disease being either fatal neonatal respiratory distress syndrome (L101P and R43L [
10,
14]) or chronic ILD (R280C; own unpublished data, [
19]). While cell biology of R43L and R280C mutations was studied here for the first time, L101P mutation was used as a known example of the trafficking/folding defect leading to the ER retention of ABCA3 with no information on ER stress [
6,
20].
Initial characterization in A549 cells demonstrated that each of the mutations affected the ABCA3 transporter in a different way. We showed correct localization of WT and R43L proteins in LAMP3
+ vesicles and dual localization of R280C protein in LAMP3
+ vesicles and calnexin
+ ER compartment (Figure
1A, B and
1C), indicating less efficient but not abolished R280C trafficking. Similar dual localization is known in the case of other ABCA3 mutants as G122S [
6]. R43L and R280C proteins showed WT-processing with two protein bands (Figure
2A) and presence of complex oligosaccharides (Figure
2B) confirming their ability to proceed from the ER to the Golgi. L101P protein remained in the ER, having therefore no complex sugars, and no smaller 180 kDa protein form (Figure
1B and
2A, B) [
6,
20].
While ER retention of L101P excludes ABCA3 function, the function of R43L and R280C transporters was studied additionally. Since ABCA3 is involved in lamellar body biogenesis [
7], in A549 cells, which normally have a low number of compact LAMP3
+ vesicles, expression of ABCA3-WT induced biogenesis of LAMP3
+ vesicles by increasing their number and size (Figure
3D) [
9]. In addition, ABCA3-WT signal was observed as ring-like structures consistent with the ABCA3 presence in the limiting membrane of lamellar bodies (Figure
1 and
3) [
5]. In contrast to the WT, expression of
ABCA3 mutations, especially L101P, impaired biogenesis of LAMP3
+ vesicles by reducing their number and size (Figure
3D). This can be a consequence of the inability of these mutated transporters to effectively load lipids into the nascent lamellar bodies.
The uptake assay of NBD-labeled phospholipids PC and PE into ABCA3-HA-positive vesicles demonstrated frequent overlap of the NBD signal, both PC and PE coupled, with the ring-like ABCA3-WT fluorescence and its accumulation in the inner space of the WT-ABCA3 vesicles (Figure
3A, B and
3C). In the case of R43L and R280C mutations such colocalization was rarely observed suggesting functional impairment of R43L and R280C proteins. Similar uptake experiments have been previously described [
6]. In contrast to those data, which show diffuse distribution of NBD-PC and NBD-PE fluorescence throughout the cytoplasm, we detected numerous distinct NBD-positive vesicles (Figure
3A, B). The uptake and number of NBD-vesicles observed in the cytoplasm was similar for both phospholipids in A549 cells and in transfected A549, and was also independent of the
ABCA3 mutation, including L101P. This confirms that for the liposome/lipid uptake through the plasma membrane into the cytoplasm ABCA3 function is dispensable and ABCA3 is solely an intracellular vesicular transporter. Furthermore, we observed an impact of
ABCA3 mutations on the uptake of both PC and PE. This parallels data published in Cheong
et al. (2007) [
7] which demonstrated decreased incorporation of radiolabeled PC and PE into the lamellar bodies and surfactant of the ABCA3 heterozygote mouse, consistent with a role of ABCA3 in the transport of both phospholipid species. However, this is also in contrast to Cheong
et al. (2006) [
6] showing no uptake of NBD-PE into the lysosome-like vesicles in A549 cells. Therefore, we must point out that it is questionable if the PE transport through functional ABCA3 actually happens
in vivo or only in an
in vitro system after exposing cultured A549 cells to NBD lipids.
Continuous overload of the ER compartment with misfolded proteins is harmful for the cell and can trigger apoptotic cell death [
22]. L101P protein which accumulates in the ER, caused significant increase of the ER stress and early and late apoptosis markers in the A549 cells (Figure
4,
5 and
6). The ER stress caused by R280C mutation was slightly lower and surface staining with annexin V, as an early apoptosis sign, was the only apoptotic marker detected in R280C cells. This might be a consequence of a dual nature of this protein which could be less harmful for the cell than complete ER retention of ABCA3 and/or possible dependency of the magnitude of its intracellular effects on small unmanageable inconsistencies of an experimental system. The connection between ER stress and apoptosis induction was established through the upregulation of caspase 4 in the case of both L101P and R280C mutations (Figure
7). Correctly localized R43L mutation, despite its influence on the ABCA3 function and lamellar body biogenesis, had almost no impact on stress and apoptosis under any conditions above the range of the WT values (Figure
4,
5,
6 and
7).
ER stress-dependent signaling, examined via XBP1 splicing, was enhanced by exposure of transfected cells to tunicamycin, an ER stressor (Figure
5B, E). In this way the stress pressure imposed on the cells was doubled: 1) genetic background of the
ABCA3 mutations and 2) exposure to the stress-causing agent. The cells with mutations R280C and L101P, which impair ABCA3 trafficking, were more prone to further XBP1 splicing than the WT or A549. This is interesting if known that viral infections (e.g. RSV, herpes virus) or cigarette smoke are common outside factors which, as tunicamycin, elevate ER stress [
39,
40]. Doan
et al. (2007) [
12] observed onset of ABCA3-associated ILD in children following exposure to cigarette smoke and Young
et al. (2008) [
16] described a teenage patient with a late onset of ABCA3-related disease with fibrosis following the beginning of cigarette consumption. If
ABCA3 mutations can raise susceptibility of AECII to external stress, additional exposure to outside stressors as respiratory viral infections or smoke might contribute to or even trigger genetic ILD.
Misfolding and ER retention of other lipid ABC transporters of the A subfamily, as ABCA1 and ABCA4, cause Tangier disease and Stargardt macular dystrophy, respectively [
41,
42], but their influence on ER stress and apoptosis is unknown. The most common mutation ∆F508 of another ABC transporter of the C subfamily, CFTR/ABCC7, leading to cystic fibrosis, results in CFTR misfolding and retention in the ER, and can raise stress and activate UPR [
43]. The concept that ER stress and apoptosis lead to lung disease has been explored recently when it was demonstrated that ER stress and apoptosis of AECII are involved in the injury of lung epithelium in idiopathic pulmonary fibrosis and SP-C deficiency [
25,
26]. Also, expression of SP-C mutations resulting in proSP-C misfolding and aggregation increases ER stress, activates UPR and induces apoptosis in A549 or HEK293 cells [
27,
28]. SP-C and ABCA3 are both AECII-expressed proteins essential for surfactant homeostasis and both lead to genetic ILD equally variable in the age of onset, severity and pathology [
44]. Therefore, it is interesting to see that the common mechanisms underlying both types of genetic ILD must exist and our data show that they probably encompass the ER stress and apoptosis of AECII.
Competing interests
The authors declare that they have no competing interests.
Authors' contributions
SK, NW, MG and AHec conceived and designed the study. NW, SK, EK, AHec, AS and MW performed the experiments and analyzed the data, which were interpreted by SK, AHec, MG and NW. AHolz contributed the pEYFP/ABCA3-WT plasmid. SK wrote the manuscript which was edited by MG, NW, EK and AHec. All authors have read and approved the final manuscript.