View metadata, citation and similar papers at core.ac.uk brought to you by CORE

provided by Elsevier - Publisher Connector

Biochimica et Biophysica Acta 1843 (2014) 797–805

Contents lists available at ScienceDirect

Biochimica et Biophysica Acta

journal homepage: www.elsevier.com/locate/bbamcr

Dissociated presenilin-1 and TACE processing of ErbB4 in lung alveolar type II cell differentiation

Najla Fiaturi a,⁎, Anika Ritzkat b,c, Christiane E.L. Dammann b,c,d, John J. Castellot a,d,1, Heber C. Nielsen b,d,1

a Program in Pharmacology and Experimental Therapeutics, Department of Integrative Physiology and Pathobiology, Sackler School of Graduate Biomedical Studies, Tufts University, Boston, MA 02111, USA b Department of Pediatrics, Tufts Medical Center, Boston, MA 02111 USA c Hannover Medical School, Hannover, Germany d Graduate Program in Cell, Molecular and Developmental Biology, Department of Integrative Physiology and Pathobiology, Sackler School of Graduate Biomedical Studies, Tufts University, Boston, MA 02111, USA

article info abstract

Article history: Neuregulin (NRG) stimulation of ErbB4 signaling is important for type II cell surfactant synthesis. ErbB4 may me- Received 25 October 2013 diate expression via a non-canonical pathway involving enzymatic cleavage releasing its intracellular do- Received in revised form 18 December 2013 main (4ICD) for nuclear trafficking and gene regulation. The accepted model for release of 4ICD is consecutive Accepted 13 January 2014 cleavage by Tumor necrosis factor alpha Converting (TACE) and γ-secretase . Here, we show Available online 24 January 2014 that 4ICD mediates surfactant synthesis and its release by γ-secretase is not dependent on previous TACE cleav- age. We used siRNA to silence Presenilin-1 (PSEN-1) expression in a mouse lung type II epithelial cell line (MLE12 Keywords: fi ErbB receptor cells), and both siRNA knockdown and chemical inhibition of TACE. Knockdown of PSEN-1 signi cantly de- creased baseline and NRG-stimulated surfactant phospholipid synthesis, expression of the surfactant Presenilin-1 SP-B and SP-C, as well as 4ICD levels, with no change in ErbB4 ectodomain shedding. Neither siRNA knockdown Neonatal lung nor chemical inhibition of TACE inhibited 4ICD release or surfactant synthesis. PSEN-1 cleavage of ErbB4 for non- Type II cell canonical signaling through 4ICD release does not require prior cleavage by TACE. Surfactant © 2014 Elsevier B.V. All rights reserved.

1. Introduction expressed by type II cells, play a prominent role in stimulation of type II cell maturation and surfactant synthesis [4,5]. NRG is expressed in Respiratory distress syndrome (RDS), formerly known as hyaline the midtrimester human fetal lung [6] and increases in fetal lung at membrane disease, is a common problem in preterm infants born be- the onset of surfactant synthesis [7]. fore 28 weeks. This disease is caused primarily by deficiency of pulmo- ErbB4 is a member of the ErbB receptor tyrosine kinase family, nary surfactant in immature lungs and is more common the earlier the which also includes the epidermal growth factor receptor, also called infant is born [1]. Despite the beneficial effects of prenatal glucocorti- ErbB1, ErbB2, and ErbB3 [8]. The ErbB receptors are transmembrane ty- coids and postnatal surfactant replacement therapies, RDS remains rosine kinase proteins and act as important regulators of cell prolifera- one of the significant causes of morbidity and mortality in premature in- tion and differentiation during fetal organ development including lung fants [2,1]. is a mixture of surface active phospho- development [9]. ErbB4 signal transduction is a complex process that in- lipids and proteins (termed (SP-B) and SP-C) which volves both canonical and non-canonical signaling pathways. Binding of is produced in alveolar type II epithelial cells [3]. The development of NRG to its extracellular ligand-binding site causes ErbB4 to form homo surfactant synthesis is under multifactorial control, in which paracrine or heterodimers with other ErbB receptors linked by disulfide bonds mesenchyme-type II cell communication mechanisms play a central in the extracellular domain [10]. These receptor dimers then undergo role. We have shown that the growth factor Neuregulin (NRG-1), auto phosphorylation on tyrosine residues within the intracellular do- which is secreted by fibroblasts, and its target receptor ErbB4, which is main. In the canonical signal pathway tyrosine phosphorylation acti- vates signal cascades through specific intracellular signaling pathways such as the phosphatidylinositol-3 (PI3) Kinase/Akt pathway to ulti- Abbreviations: SP-B, Surfactant protein B; SP-C, ; NRG-1, fl Neuregulin-1; TACE, Tumor necrosis factor alpha converting enzyme; DSPC, Disaturated mately in uence [11]. However, within the ErbB fami- phosphatidyl choline ly, ErbB4 is unique in that it may undergo proteolytic processing to ⁎ Corresponding author at: 136 Harrison Ave., Tufts University, Boston, MA 02148, USA. initiate non-canonical signaling [12]. The accepted model for the non- Tel.: +1 781 228 9652. canonical pathway involves two sequential cleavage processes. The fi E-mail addresses: naj [email protected] (N. Fiaturi), [email protected] (A. Ritzkat), fi [email protected] (C.E.L. Dammann), [email protected] rst step in this pathway is performed by a transmembrane (J.J. Castellot), [email protected] (H.C. Nielsen). metalloprotease Tumor necrosis factor alpha Converting Enzyme 1 Co-senior authors. (TACE) which releases the ErbB4 ectodomain by a cleavage that

0167-4889/$ – see front matter © 2014 Elsevier B.V. All rights reserved. http://dx.doi.org/10.1016/j.bbamcr.2014.01.015 798 N. Fiaturi et al. / Biochimica et Biophysica Acta 1843 (2014) 797–805 produces two fragments: a 120 kDa ectodomain fragment which is Millipore (Billerica, MA). Anti-prosurfactant protein C (rabbit polyclonal released into the extracellular space and an 80 kDa membrane- antibody, ab90716), anti-prosurfactant protein B (rabbit polyclonal anti- associated fragment that contains the ErbB4 body ab15011), monoclonal antibody to beta actin (HRP conjugated, and the entire cytoplasmic region, including the tyrosine kinase domain. cat# 20272), Anti-TACE antibody (rabbit polyclonal antibody, cat# Ectodomain cleavage of ErbB4 in cells occurs at a low constitutive or ab2051) were all from Abcam (Cambridge, MA). Peroxidase-conjugated basal level [13] that can be increased by neuregulin or other ErbB4 Affinipure Goat Anti-Rabbit IgG (H + L), (111-035-144) was from Jackson ligands [14]. The ectodomain cleavage of ErbB4 is sensitive to ImmunoResearch. Anti-ERBb4 antibody (cat# sc-283) was obtained from metalloprotease inhibitors [13] and does not occur in cells genetically Santa Cruz (Santa Cruz, CA). Chloroform (spectrophotometric grade) was deficient in TACE. The ErbB4 m80 fragment (membrane associated frag- obtained from Sigma Aldrich (St Louis, MO). Methanol was from Fisher ment) that remains following ectodomain cleavage is further processed Scientific; Silica gel H thin layer chromatography sheets were from by γ-secretase that cleaves at the transmembrane domain to release a Analtech (Newark, DE). Osmium tetroxide 05500-1G, carbon tetrachlo- soluble intracellular s80 fragment (4ICD) into the cytosol from which ride and dipalmitoylphosphatidyl choline (P-5911) were from Sigma it translocates to the nucleus in association with chaperone proteins Aldrich (St Louis, MO). Ultima-Gold scintillation fluid was from Perkin [14,15]. Gamma secretase is an enzyme complex which consists of 4 Elmer (Waltham, MA). For centrifugation we used a Beckman J-6 M. components: presenilin 1 (PSEN-1) or Presenilin 2 (PSEN-2), Nycastrin, anterior pharynx defective-1 (APH-1) and the Presenilin-Enhancer 2. PSEN-1 or PSEN-2 are the active enzymatic component; other compo- 2.2. Cell culture nents behave as scaffolding molecules and essential cofactors [16–19]. Studies with transgenic mice show that PSEN-1 and PSEN-2 have func- MLE12 cells were used as a model for type II alveolar epithelial tionally distinct phenotypes in several organs including the lung cells. MLE12 cells exhibit characteristics of alveolar type II cells, in- [20,21]. cluding the expression of SP-B and SP-C and formation of microvilli Nuclear localization of ErbB4 is the preferred mechanism of ErbB4 and multivesicular bodies. They have a strong response to fetal signaling in several regulatory processes during development [22].The fibroblast-conditioned media (FCM) and NRG with increased DPSC accepted model for γ-secretase activity in ErbB4 processing is that synthesis [7,24]. MLE-12 cells were grown in DMEM containing ectodomain cleavage by TACE is a prerequisite step [12]. In this study 10% FBS, 2% pen/strep and 2% L-glutamine. Media were changed we sought to more specifically define the sequential interactive relation- every second day. ship of PSEN-1 and TACE for ErbB4 processing controlling lung alveolar type II cell surfactant production. Our focus on PSEN-1 was motivated by the more severe developmental phenotype of alveolar maturation in 2.3. Transfection with siRNA the PSEN-1 knockout mouse compared to the PSEN-2 knock out and our previous work showing the importance of PSEN-1 signaling for MLE12 cells were transfected with siRNA using the transfection re- fetal type II cell maturation [22,23]. We studied the effect of PSEN-1 agent Dharmafect 2. To knock down presenilin 1, three pre-designed knockdown and TACE knockdown in MLE12 cells and evaluated the ef- Presenilin-1 siRNA sequences that target three different regions of fects on ErbB4 cleavage in association with the expression of the SP-B Presenilin mRNA were used. To knock down TACE we used a cocktail and SP-C mRNA and protein and synthesis of the major surfactant phos- of 3 siRNAs targeting the TACE gene. The protocol for transfection of pholipid disaturated phosphatidylcholine (DSPC). MLE12 cells was adapted from the manufacturer's guidelines; all steps were done using RNAse-free pipette tips and RNase spray for 2. Methods and materials decontaminating the work area. MLE12 cells were plated into 6 well plates. The transfection process was initiated when the cells were 30– 2.1. Materials 40% confluent. 5 μl siRNA in 95 μl serum free DMEM/well and 1 μltrans- fection reagent Dharmafect 2 in 99 μ serum free DMEM/well) were in- Dulbecco's Modified Eagle's (DME) low glucose medium was pur- cubated for 5 min at room temperature, then mixed and incubated for chased from Sigma-Aldrich (St. Louis, MO). Fetal bovine serum (FBS) 20 min at room temperature. At the end of the incubation, 800 μlofan- was from BD Biosciences (Lot # ANB 18202A), L-glutamine, Pen/Strep tibiotic free media was added to the mixture. A total of 1 mL of transfec- and an siRNA cocktail of three siRNA sequences targeting PSEN-1 tion media was added to each well of cells and allowed to incubate at (Psen 1Mss208049, Psen 1Mss208050, Psen 1Mss208051) were pur- 37 °C. After 48 h, the transfection medium was aspirated and replaced chased from Invitrogen (Carlsbad, CA). TACE inhibitor TAPI-1 was pur- with fresh transfection media prepared as described above. A second chased from Peptides International (Louisville, KY). siRNA cocktail of transfection step was done in the same manner after 48 h from the three siRNA sequences targeting TACE, Silencer Negative Control scram- first transfection and cultures continued for another 24 h resulting in a bled siRNA and Silencer GAPDH siRNA were purchased from Ambion (St total exposure time of 72 h. When transfecting the cells for the second Louis, MO). Glass 100 mm cell culture dishes were purchased from time the media were changed to serum free media, and some cells Pyrex/Corning (St Louis, MO). The murine lung epithelial cell line were treated with NRG (3.3 nM) for 24 h. Each experiment included MLE12 was purchased from the American Type Culture Collection the control conditions of scrambled siRNA and GAPDH siRNA. After 72 (Manassas, VA), Neuregulin 1β was produced using an expression vec- h of total exposure time, the MLE-12 cells in the 6-well plates were tor kindly provided by Kermit Carraway III (UC Davis, CA) and purified harvested. by Dr. Ann Kane, Phoenix Laboratory (Tufts Medical Center, Boston, MA). BCA Protein Assay Kit and RIPA buffer were obtained from Pierce/Thermo Scientific (Logan, UT), protease inhibitor cocktail from 2.4. Chemical blockade of TACE Sigma Aldrich. Invitrolon PVDF filter paper and NuPage 4–12% Bis-Tris pre-cast gels (1.0 mmX 12 wells) were obtained from Novex/Invitrogen. MLE12 cells were plated in 6 well plates. When the cells were Tris-Glycine-SDS 10X running buffer, transfer buffer (10X Tris-buffered 50–60% confluent they were first treated for 24 h with the TACE in- saline washing buffer) and TBS-Tween-20 (10X) were from Boston Bio hibitor TAPI. In preliminary experiments increasing concentrations Products (Ashland, MA). Methanol 100%, Kodak film (X-OMAT Blue (50 nM, 100 nM, 150 nM, 200 nM, 300 nM) were used to determine XB) and Restore Plus Western Blot Stripping Buffer were from Thermo the minimal effective dose. Thereafter cells were treated for 24 h with Fisher Scientific (Palm Beach, FL). Anti-Presenilin 1 N-terminal (1–65) 200 nM TAPI alone, TAPI plus NRG, NRG only or media only. Cells rabbit polyclonal antibody (cat#529591) was from Calbiochem/EMD were then harvested. N. Fiaturi et al. / Biochimica et Biophysica Acta 1843 (2014) 797–805 799

2.5. Protein extraction and quantification 2.9. Conditioned media

Cells were harvested by first washing with PBS 3 times. 500 μl RIPA MLE12 cells (3 × 105) were plated and treated for 72 h with siRNA or buffer: protease inhibitor mixture (1:50) was added to each well TAPI as described above. Conditioned media was collected probed for twice. Using a cell scraper, cells were scrapped off plates (on ice), the presence of the shed ectodomain of ErbB4 using western blot aspirated, incubated on ice for 60 min with vortexing every 10 min, cen- analysis. trifuged at maximum speed for 15 minutes and the supernatant trans- ferred to new tubes and stored at −20 °C. Protein concentration in 2.10. Statistics each sample was determined in duplicate (BCA protein assay, Thermo Scientific). Statistical analysis was done using non-parametric ANOVA with post-hoc analysis or two-tailed t-tests (Graph Pad Software, San Diego, CA) as appropriate with a level of significance of P = 0.05. All 2.6. Western blot analysis data are expressed as Means ± SEM.

The protein samples were heated at 70 °C for 10 min with sample re- 3. Results ducing buffer to denature the protein. Proteins were loaded into pre- cast NuPage 4–12% Bis-Tris gels and separated by gel electrophoresis. 3.1. PSEN-1 knockdown decreases SP-B and SP-C expression in MLE12 cells Subsequently the proteins were transferred to a PVDF membrane. The membrane was blocked with 5% dry milk in TBST for 2 h at room tem- To determine if PSEN-1 regulates SP-B and SP-C expression, we perature or at 4 °C overnight. The blots were probed with primary down regulated PSEN-1 using siRNA in MLE12 cells and measured the antibody against PSEN-1 or TACE, and then stripped and probed for effect on SP-C and SP-B mRNA using qRT-PCR and protein levels SP-B, SP-C and beta Actin. Primary antibodies were incubated overnight using western blot analysis. PSEN-1 mRNA was reduced to 38% ± 17 at 4 °C and secondary antibodies were incubated for 2 h at room tem- (mean ± SE, n = 8) compared to the scrambled control; PSEN-1 protein perature. Blots were developed using Western Lightning Plus ECL and was reduced to 50% ± 12 (mean ± SE, n = 5, p = 0.02) of the scram- detected with a Kodak film X-OMAT Blue XB. Densitometry was done bled control (Fig. 1A). At this level of PSEN-1 knockdown, Sftpb mRNA for all blots and beta actin was used as an internal control for protein was reduced to 67% ± 11 (mean ± SE, n = 8) compared to the loading. scrambled control. The level of SP-B protein was reduced to 60% ± 9.7 (mean ± SE, n = 5, p = 0.01) of the scrambled control (Fig. 1B). Sftpc mRNA was reduced to 57% ± 6 (mean ± SE, n = 8) compared to the 2.7. Choline incorporation into DSPC scrambled control. SP-C protein was decreased to 63% ± 4.2 (mean ± SE, n = 5, p =0.017) in response to PSEN-1 protein knock down MLE12 cells were cultured and transfected in 24-well plates as de- (Fig. 1C). The specificity of PSEN-1 knockdown was checked by assaying scribed above except that the cells were treated with 0.5 μCi/ml [3H] GAPDH protein, which showed no significant decrease at the highest choline for the final 24 h of transfection. At the end of the 72 h of trans- level of PSEN-1 knockdown (data not shown here). These results suggest fection, the cells were harvested, the lysates sonicated and protein con- that PSEN-1 is required for optimal SP-B and SP-C expression. centrations measured. DSPC was isolated by lipid extraction, osmium tetroxide treatment and then liquid chromatography as we have de- 3.2. PSEN-1 knockdown suppresses the stimulatory effect of neuregulin on scribed [25,7]. The isolated DSPC samples were placed in scintillation SP-B and SP-C protein expression fluid and DPMs counted using a beta scintillation counter. The results were calculated as DPM per μg protein and presented as percentage of Others and we have previously shown that NRG stimulates surfac- the experimental specificcontrolvalue. tant production through the ErbB4 receptor [4,5].Todetermineif PSEN-1 down-regulation affects the stimulatory effect of NRG, we knocked down PSEN-1 in MLE12 cells as described above and exposed 2.8. RT-PCR these cells to 3.3 nM NRG during the final 24 hours. Western blot anal- ysis of scrambled control cells treated with NRG showed significantly Real-time PCR was used to determine the mRNA levels for PSEN-1, increased expression of SP-B and SP-C to 150.7 ± 9; 148% ± 10 respec- SP-B (Sftpb) and SP-C (Sftpc). PSEN-1 primers: Forward sequence 5′ tively; means ± SE, n = 5, p = 0.01 and p = 0.002, respectively (Fig. 2A TCA/AGA/AAG/CGT/TGC/CAG/C 3′, Reverse sequence 5′ CGT/GGC/ and B). PSEN-1 knockdown completely abrogated the stimulatory effect GAA/GTA/GAA/CAC/GA 3′, Sftpb primers: Forward sequence 5′ AGG/ of NRG on SP-B and SP-C protein levels (Fig. 2C and D). These data sug- ATG/CCA/TGG/GCC/CT 3′, Reverse sequence 5′ TCA/GTG/TCC/TGT/ gest that the ability of NRG to up-regulate SP-B and SP-C is dependent AGT/GGC/CAT/T 3′, Sftpc primers: Forward sequence 5′ CCA/CTG/GCA/ upon adequate PSEN-1 expression. TCG/TTG/TGT/ATG 3′, Reverse sequence 5′GTA/GGT/TCC/TGG/AGC/ TGG/CTT/A 3′, all were purchased from Applied Biosystems (Woburn, 3.3. Release of the ErbB4 cytoplasmic fragment (4ICD) is decreased in MA). The 50 μl reaction master mix contained 25 μl Taq polymerase, MLE12 cells treated with PSEN-1 siRNA 1.25 μl Multiscribe and RNA inhibitor mix, 8 μMeachofforwardandre- verse primer, 5 μM probe and 1 μg of RNA sample. Amplification and de- The ErbB4 cytoplasmic domain (4ICD) is thought to be important to tection of specific products were done with the ABI PRISM 7900 control gene expression of surfactant proteins. To determine if PSEN-1 sequence detection system from Applied Biosystems. The amplification regulates the level of 4ICD, we knocked down PSEN-1 in MLE12 cells protocol consisted of an initial denaturation and enzyme activation at and measured 4ICD levels using Western blot analysis. Cells treated 95 °C for 10 min, followed by 45 cycles at 95 °C for 15 s and 60 °C for with PSEN-1 siRNA showed significantly decreased levels of the 4ICD 1 min. In order to normalize the PSEN-1 and Sftp levels, actin was used product (60% ± 2.5; mean ± SE, n = 5, p = 0.03) compared to SCR as an internal control. Samples were run in triplicate. The differences siRNA treated cells. Control samples (SCR siRNA) treated with NRG in the Ct values of the PSEN-1 siRNA transfected cells compared to the showed increased 4ICD levels (130% ± 8 of non-NRG treatment; cells transfected with a scrambled (scr) siRNA sequence were expressed mean ± SE, n = 5, p = 0.02). Thus, NRG up-regulated PSEN-1 process- as DDCT and presented as % of the controls (= scr siRNA). The message ing of ErbB4, increasing the level of the 4ICD fragment, but in PSEN-1 levels of Sftpb and Sftpc were calculated accordingly. knock-down cells NRG did not increase the 4ICD fragment. (Fig. 3). 800 N. Fiaturi et al. / Biochimica et Biophysica Acta 1843 (2014) 797–805

Fig. 1. Representative western blots and densitometry quantification showing PSEN-1 knockdown in MLE12 cells and the effect on SP-B and SP-C protein levels. Densitometry results are expressed as % of the scrambled (SCR) siRNA condition (controls). Bars are means ± SEM of N = 5 experiments. * = P b 0.05. (A) PSEN-1 knockdown in PSEN-1 siRNA treated cells, GAPDH siRNA treated cells and SCR siRNA treated cells. Values are expressed as % of the SCR condition (controls). (B) SP-B expression in PSEN-1 siRNA, GAPDH siRNA and SCR siRNA treated MLE12 cells. (C) SP-C expression in PSEN-1 siRNA, GAPDH siRNA and SCR siRNA treated MLE12 cells.

3.4. PSEN-1 knockdown decreases DSPC synthesis in MLE12 cells (Supplementary Fig. 1). These observations suggest that, in contrast to its effect on levels of the 4ICD cytoplasmic domain, PSEN-1 has no de- DSPC is the major phospholipid component and major surface active tectable effect on the release of the ErbB4 ectodomain. component of surfactant. We therefore assessed the effect of PSEN-1 knockdown on DSPC synthesis in MLE12 cells. Cells treated with PSEN-1 siRNA showed significantly decreased DSPC synthesis compared 3.6. TACE knockdown does not affect SP-B and SP-C expression In MLE12 to scrambled siRNA treated samples (Fig. 4). NRG increased the level of cells: DSPC synthesis in SCR-treated cells, and this stimulatory effect was lost following PSEN-1 knockdown (Fig. 4). These results indicate that PSEN- The necessary first step in proteolytic processing of ErbB4 leading to 1 acts to regulate baseline and NRG-induced DSPC synthesis in MLE12 ultimate release of the 4ICD fragment and non-canonical ErbB4 signal- cells, similar to its effect on SP-B and SP-C. ing is thought to be cleavage by TACE, a trans-membrane metallopro- teinase. To determine if TACE activity is required for NRG stimulation 3.5. Release of the ErbB4 ectodomain is not changed in MLE12 cells treated of surfactant production, we used siRNA to knock down TACE in with PSEN-1 siRNA MLE12 cells and measured SP-B and SP-C expression levels by west- ern blot analysis. TACE protein was reduced to 53% ± 3 (mean ± SE, To assess the effect of PSEN-1 on expression of the ErbB4 ectodomain n = 4, p = 0.003) of the scrambled control (Fig. 5A). MLE12 cells we collected conditioned media from cells treated with PSEN-1 siRNA. treated with TACE siRNA showed no change in SP-B and SP-C Conditioned media from NRG-treated cells showed increased free (Fig. 5B). Furthermore, TACE knockdown did not inhibit the stimula- ErbB4 ectodomain levels compared to untreated cells. Conditioned tion of SP-B and SP-C by NRG. Cells treated with NRG showed higher media from cells treated with PSEN-1 siRNA showed no change in the expression of SP-B and C in the SCR treatment condition (140.7% ± 9, levels of the ErbB4 ectodomain (96% ± 5; mean ± SE, n = 5) compared 120% ± 7 respectively; mean ± SE, n = 4; p = 0.02, p = 0.003 to SCR siRNA treated cells. The free ErbB4 ectodomain level was respectively) and this stimulatory effect appeared to be maintained increased in SCR cells treated with NRG to 120% ± 8.3 (mean ± SE, in TACE knockdown samples (Fig. 5C and D). These results indicate n = 5), and in PSEN-1 siRNA treated cells the level of the 120 kD that NRG stimulation of surfactant production is largely independent ErbB4 ectodomain was increased to 135% ± 5.2 (mean ± SE, n = 5) of TACE activity in MLE12 cells. N. Fiaturi et al. / Biochimica et Biophysica Acta 1843 (2014) 797–805 801

Fig. 2. The effect of NRG on SP-B and SP-C protein expression in MLE12 cells following PSEN-1 knock down. (A) SP-B expression in SCR siRNA treated MLE12 cells in the absence and pres- ence of NRG. (B) SP-C expression in SCR siRNA treated MLE12 cells in the absence and presence of NRG. (C) SP-B expression in PSEN-1 siRNA treated MLE12 cells in the absence and pres- ence of NRG. (D) SP-C expression in PSEN-1 siRNA treated MLE12 cells in the absence and presence of NRG. Bars are means ± SEM of N = 5 experiments. * = P b 0.05, ** = P b 0.005 compared to no NRG treatment.

Fig. 4. The effect of PSEN-1 knockdown on choline incorporation into DSPC. Bars show the Fig. 3. Representative western blots and densitometry quantification showing the effect of mean ± SEM of [3H]-choline incorporated into DSPC, expressed as disintigrations per min- PSEN-1 knockdown on the level of free ErbB4 cytoplasmic domain (4ICD) at baseline and ute (DPM) per microgram of protein × 103 in MLE12 cells treated with PSEN-1 siRNA com- following NRG treatment. Mean ± SEM, N = 5. * = P b 0.05 compared to SCR cells treated pared to SCR treated cells, in the absence or presence of NRG. * = P b 0.05 compared to SCR with NRG. with NRG; ** = P b 0.005 compared to SCR without NRG. 802 N. Fiaturi et al. / Biochimica et Biophysica Acta 1843 (2014) 797–805

Fig. 5. Representative western blot and densitometry quantification of TACE knockdown in MLE12 cells and the effect on surfactant protein expression. (A) TACE knockdown in TACE siRNA, GAPDH siRNA and SCR siRNA treated cells. Densitometry values were normalized to actin and compared to SCR siRNA treatment. Bars show mean ± SEM of N = 4 experiments. (B) Representative western blots of SP-B and SP-C expression in TACE siRNA, GAPDH siRNA and SCR siRNA treated MLE12 cells. (C) Western blots and densitometry of SP-B expression in TACE siRNA and SCR siRNA treated cells without and with NRG treatment. Densitometry values were normalized to actin and compared to the respective SCR value. Bars are means ± SEM of N = 4 experiments; * = P b 0.05. (D) Western blots and densitometry of SP-C expression in TACE siRNA and SCR siRNA treated cells without and with NRG treatment. Densitometry values are normalized to the respective SCR value. Bars are means ± SEM of N = 4 experiments; ** = P b 0.005.

3.7. Level of shed ErbB4 ectodomain, but not 4ICD is decreased in MLE12 activity. Dose–response experiments in MLE12 cells showed that 200 cells treated with TACE siRNA nM was sufficient to reduce active TACE levels (measured by western blot) to 30% ± 1.4 (mean ± SE, n = 4, p = 0.0004) compared to To assess the role of TACE in the release of both the ectodomain and media treated cells (Supplementary Fig. 2). the 4ICD components of ErbB4, we measured the ectodomain level in conditioned media and the 4ICD level in cell lysates following TACE knockdown via siRNA. Western blot analysis of cells treated with 3.9. TACE inhibition by TAPI does not affect SP-B, 4ICD expression TACE siRNA showed no change in 4ICD levels compared to SCR treated cells (98% ± 2.5 of SCR cells; mean ± SE, n = 4). NRG treatment in- We confirmed the absence of a role of TACE in ErbB4-mediated creased the level of 4ICD in both SCR treated and TACE siRNA treated regulation of SP-B by treating MLE12 cells with 200nM TAPI without cells (130.6% ± 2.1, 150% ± 2.3, means ± SE, n = 4 respectively) com- and with NRG. MLE12 cells treated with TAPI showed no change in pared to untreated scrambled control cells (Fig. 6A). To further demon- SP-B (99% ± 1.4, mean ± SE, n = 4) (Fig. 7A). These results indicate strate that TACE inhibition blocked proteolytic release of the ErbB4 that baseline SP-B production is largely independent of TACE activ- ectodomain, conditioned medium from TACE siRNA treated cells was ity, at least in MLE12 cells. MLE12 cells exposed to NRG alone collected after 24 h of treatment, and media assayed for presence of or NRG plus TAPI showed higher levels of SP-B protein. SP-B was sig- the released ectodomain fragment of ErbB4. Western blot analysis of nificantly increased in NRG treated cells to 214% ± 14 (means ± SE, the conditioned media from cells treated with TACE siRNA showed de- n = 4, p = 0.0004) and in NRG plus TAPI treated cells to 197% ± 11 creased ErbB4 ectodomain levels (40% ± 1.1; mean ± SE, n = 5, p = (means ± SE, n = 4, p = 0.004) (Fig. 7B). These results confirm that 0.003) compared to scrambled siRNA treated cells. NRG stimulated the unlike PSEN-1, TACE is not involved in NRG-stimulated mediation of level of released ectodomain in scrambled siRNA treated cells ErbB4 control of SP-B expression in MLE12 cells. Analysis of the (150% ± 8.7, mean ± SE, n = 4, p = 0.0173) compared to cells not treat- conditioned media showed that TAPI-treated cells shed significantly ed with NRG. The stimulatory effect of NRG on ectodomain shedding was less ErbB4 ectodomain (50% ± 3.4; mean ± SE, n = 4, p = 0.001) significantly lost in TACE siRNA treated samples (Fig. 6B). These observa- compared to untreated cells. NRG treatment stimulated ectodomain tions indicate that TACE specifically regulates ectodomain cleavage, but release in control cells (143.8% ± 5; mean ± SE, n = 4, p = 0.01). not cytoplasmic cleavage regulated by PSEN-1 leading to 4ICD release, As expected, this stimulatory effect was absent in conditioned contrary to current models of ErbB4 non-canonical signaling. media from TAPI treated cells (Fig. 8A). The presence of the 4ICD fragment in cell lysates from TAPI treated cells was not changed 3.8. Chemical inhibition of TACE activity (85.3% ± 5.8; mean ± SE, n = 3, p = 0.17) compared to media treated cells (Fig. 8B). These data indicate that TACE inhibition by To further validate the results obtained from TACE knockdown, we TAPI prevents TACE activity consistent with TACE knockdown using used TAPI, a specific and well-characterized chemical inhibitor of TACE siRNA. N. Fiaturi et al. / Biochimica et Biophysica Acta 1843 (2014) 797–805 803

Fig. 7. The effect of TACE inhibition by TAPI on SP-B levels. (A) Representative western blot and densitometry quantification of SP-B protein levels in MLE12 cells treated as control or Fig. 6. The effect of TACE knockdown on the levels of the 4ICD cleavage fragment (A) and with 200 nM TAPI. Inhibition of TACE by TAPI did not affect SP-B levels. (B) Representative the released ectodomain fragment (B) in MLE-12 cells. Figure shows representative west- western blot and densitometry quantification of the effect of TACE inhibition by TAPI on ern blots and densitometry values normalized to SCR without or with NRG treatments re- NRG stimulation of SPB protein levels. Bars represent mean ± SEM of N = 4 experiments; spectively. Bars are means ± SEM of N = 4 experiments; ** = P b 0.005 compared to SCR ** = P b 0.005, *** = P b 0.0005 compared to the respective controls. condition, T-test with Bonferroni correction for multiple comparisons. PSEN-1 or PSEN-2 in mice results in pulmonary phenotypes [23]. 4. Discussion While the pulmonary phenotypes have not been completely character- ized, it is known that newborn mice with PSEN-1 knockout have lungs In spite of advances in the treatment and prevention of RDS, includ- with poorly developed saccular structures and undergo neonatal ing the use of prenatal glucocorticoids and neonatal treatment with death with poorly expanded lungs. PSEN-2 knockout mice survive postnatal surfactant replacement, RDS remains one of the significant birth but develop progressive alveolar wall thickening, alveolar and air- causes of morbidity and mortality in premature infants [26,27]. There- way fibrosis, and hemorrhage. Double knockouts show an exacerbation fore, it is important to investigate the endogenous pathways regulating of the fetal/neonatal abnormalities of alveolar development [20,21]. the control of surfactant synthesis in the alveolar type II cell. We have Mice with an inactivating in TACE also have poorly developed shown the important role of ErbB4 receptor and its ligand NRG in sur- saccular structures and undergo neonatal death with poorly expanded factant synthesis [7,4,5,22], but the specific signaling mechanism(s) in lungs [28]. However, no study has addressed the interdependence of type II cells that promote surfactant synthesis is not well understood. TACE and PSEN-1 processing of ErbB4 in the ErbB4 signaling pathway Our previous work suggests that ErbB4 proteolytic processing and traf- controlling type II cell surfactant expression. ficking mechanisms are involved in its control of type II cell maturation We focused on PSEN-1 because of the more severe developmental and surfactant production [22]. In this study we defined the signifi- phenotype of alveolar maturation in the PSEN-1 knockout mouse. We cance of ErbB4 processing by PSEN-1 and TACE for stimulation of used siRNA to silence PSEN-1 expression in MLE12 cells. MLE12 cells type II cell surfactant DSPC synthesis and SP-B and SP-C expression treated with siRNA targeting PSEN-1 showed a reduction in the level to provide a basis for the development of novel therapeutic ap- of ErbB4 4ICD cytoplasmic cleavage product and no change in the proaches to prevent and treat RDS. In the process we found that pro- ectodomain shedding component of ErbB4. Moreover, these cells exhib- teolytic processing of ErbB4 by PSEN-1 does not require previous ited significant decreases in the surfactant proteins SP-B and SP-C and in processing by TACE. These are novel findings for both ErbB4 signal- the synthesis of the major phospholipid component of surfactant, DSPC. ing and type II cell biology. Further, cells treated with PSEN-1 siRNA showed a reduction in the re- Previous studies of the proteolytic processing of ErbB4 in NRG- leased ErbB4 4ICD cytoplasmic cleavage product and no change in the induced signaling indicate this is a two step process beginning with shedding of the ectodomain component of ErbB4. NRG treatment cleavage at the cell surface by TACE with shedding of the ectodomain, afforded the opportunity to evaluate the initiation of this ErbB4 process- followed by cleavage at the inner by γ-secretase to re- ing pathway. Our data indicate that cells treated with NRG showed in- lease the 4ICD fragment which traffics to the nucleus [12]. There are creased 4ICD levels in association with stimulated expression of SP-B, two separate enzymes that may make up the active component in the SP-C and DSPC synthesis. The fact that these stimulatory effects were γ-secretase complex, PSEN-1 and PSEN-2, which are transcribed from lost with PSEN-1 knockdown provides important evidence of the im- different . While there is some overlap in their function, studies portance of PSEN-1 processing in ErbB4-mediated type II cell surfactant show that γ-secretase complexes containing PSEN-1 or PSEN-2 have synthesis that strengthens our conclusions from previous studies [22]. functionally distinct phenotypes [20]. However, deletion of either Thus our data indicate that ErbB4 signaling in lung alveolar type II 804 N. Fiaturi et al. / Biochimica et Biophysica Acta 1843 (2014) 797–805

Fig. 8. The effect of TACE inhibition by TAPI on the levels of released ectodomain fragment (A) and 4ICD cleavage fragment (B) from ErbB4. Figure shows representative western blots and densitometry values normalized to SCR without or with NRG treatments respectively. Bars are means ± SEM of N = 4 experiments; * = P b 0.05, ** = P = 0.005 compared to respective controls. Bars are means ± SEM of N = 4 experiments. cells utilizes the pathway of the PSEN-1-dependent γ-secretase proteo- The results of these experiments showed that knockdown of TACE ei- lytic cleavage for promoting surfactant synthesis. ther by siRNA or TAPI does not inhibit SP-B and SP-C expression in In our studies of 4ICD levels in response to NRG treatment following MLE12 cells. More significantly we found that TACE inhibition did not PSEN-1 knockdown, the effect of NRG stimulation on the level of 4ICD suppress the stimulatory effect of NRG on surfactant protein levels. To was not completely removed. This could reflect some cleavage activity further probe the role of TACE we studied the effect of TACE inhibition by PSEN-2 or possibly the fact that with our knockdown the PSEN-1 pro- on γ-secretase mediated cleavage of ErbB4 in response to NRG by prob- tein level was reduced to 50%, possibly allowing a low residual PSEN-1 ing for the 4ICD fragment of ErbB4 in whole cell lysates of TACE siRNA activity. However at the same time the NRG effect on the levels of SPB, treated and TAPI treated cells. We found that even after TACE knock- SPC and DSPC were all significantly affected by the PSEN-1 knockdown. down or inhibition NRG stimulation continued to increase 4ICD levels. This fact, plus the fact that the change in 4ICD with NRG was small, we On the other hand the shedding of the ectodomain of ErbB4 was de- interpret as evidence that any residual 4ICD production was not suffi- creased in the conditioned media of these cells, as would be expected cient to allow NRG stimulation of surfactant proteins and DSPC. This with inactivation of TACE. These findings suggest that the two cleavage conclusion is also supported by our previous work in which a mutant steps in ErbB4 are not linked events that must both occur for ErbB4 pro- ErbB4 construct that lacked a gamma secretase binding site transfected cessing leading to signaling by the 4ICD component. However, we can- into fetal lung type II cells did not increase baseline or NRG-stimulated not rule out the possibility that the low remaining TACE activity was SP-B mRNA expression [22]. sufficient to prime PSEN-1 cleavage. Nevertheless, together with our γ-Secretase cleavage and nuclear transport of ErbB4 are important previous studies of the significance of nuclear localization of ErbB4 in components in regulation of fetal lung type II cell surfactant protein pro- NRG-mediated control of type II cell differentiation [22] these data indi- duction [22]. Because the accepted model of ErbB4 proteolytic process- cate that in this system proteolytic processing by TACE and by γ- ing includes cleavage by TACE followed by cleavage by γ-secretase, we secretase have separate functions for ErbB4 signaling. In conclusion then sought to demonstrate that shedding of the ectodomain through we find that ErbB4 signaling in type II epithelial cells utilizes the path- TACE cleavage is also a necessary first step for ErbB4 signaling in con- way of the PSEN-1-dependent γ-secretase . Our results em- trolling surfactant synthesis. We used both siRNA to knock down TACE phasize the importance of PSEN-1 for ErbB4 signaling in control of and chemical inhibition of TACE enzyme activity in MLE12 cells to alveolar type II cell differentiation. Our work also introduces novel in- study the importance of TACE processing of ErbB4 for SP-B and SP-C sight into the mechanism of ErbB4 cleavage processing to produce sig- production. For TACE inhibition we used TAPI, a hydroxamate inhibitor naling by the 4ICD component by showing that TACE activity is not a of TACE activity [29], which targets the zinc active motif on the enzyme. necessary antecedent for γ-secretase cleavage of ErbB4 in the control N. Fiaturi et al. / Biochimica et Biophysica Acta 1843 (2014) 797–805 805 of surfactant DSPC synthesis and SP-B and SP-C expression. This dissoci- [14] W. Zhou, G. Carpenter, Heregulin-dependent trafficking and cleavage of ErbB-4, – γ J. Biol. Chem. 275 (2000) 34737 34743. ation between TACE and -secretase processing of ErbB4 in lung type II [15] C. Goutte, M. Tsunozaki, V.A. Hale, J.R. Priess, APH-1 is a multipass membrane pro- cells may have important translational impact as investigators seek to tein essential for the in embryos, modulate ErbB4 signaling. In particular, this study may provide signifi- Proc. Natl. Acad. Sci. U. S. A. 99 (2002) 775–779. [16] D. Edbauer, E. Winkler, J.T. Regula, B. Pesold, H. Steiner, C. Haass, Reconstitution of cant clues for how to target ErbB4 biology in developing translational gamma-secretase activity, Nat. Cell Biol. 5 (2003) 486–488. approaches to benefit babies with RDS. [17] W.P. Esler, W.T. Kimberly, B.L. Ostaszewski, W. Ye, T.S. Diehl, D.J. Selkoe, M.S. Supplementary data to this article can be found online at http://dx. Wolfe, Activity-dependent isolation of the presenilin–gamma–secretase com- doi.org/10.1016/j.bbamcr.2014.01.015. plexrevealsnicastrinandagammasubstrate,Proc.Natl.Acad.Sci.U.S.A.99 (2002) 2720–2725. [18] R. Francis, G. McGrath, J. Zhang, D.A. Ruddy, M. Sym, J. Apfeld, M. Nicoll, M. Maxwell, Acknowledgment B. Hai, M.C. Ellis, A.L. Parks, W. Xu, J. Li, M. Gurney, R.L. Myers, C.S. Himes, R. Hiebsch, C. Ruble, J.S. Nye, D. Curtis, aph-1 and pen-2 are required for Notch pathway signal- ing, gamma–secretase cleavage of betaAPP, and presenilin protein accumulation, The authors wish to thank J. Russo, A. Chetty, M. Volpe, L. Pham, K. Dev. Cell 3 (2002) 85–97. Pringa and S. Mujahid, for helpful discussions and critical input on the [19] Y.M. Li, M. Xu, M.T. Lai, Q. Huang, J.L. Castro, J. DiMuzio-Mower, T. Harrison, C. Lellis, A. Nadin, J.G. Neduvelil, R.B. Register, M.K. Sardana, M.S. Shearman, A.L. Smith, X.P. study. The work was supported by NIH R01 HD046251, R01 HL085648, Shi, K.C. Yin, J.A. Shafer, S.J. Gardell, Photoactivated gamma-secretase inhibitors di- R21 HL097231 and a grant from the Peabody Foundation. rected to the covalently label presenilin 1, Nature 405 (2000) 689–694. [20] P. Mastrangelo, P.M. Mathews, M.A. Chishti, S.D. Schmidt, Y. Gu, J. Yang, M.J. Mazzella,J.Coomaraswamy,P.Horne,B.Strome,H.Pelly,G.Levesque,C. References Ebeling, Y. Jiang, R.A. Nixon, R. Rozmahel, P.E. Fraser, P. St George-Hyslop, G.A. Carlson, D. Westaway, Dissociated phenotypes in presenilin transgenic mice de- [1] A.H. Jobe, M. Ikegami, Surfactant metabolism, Clin. Perinatol. 20 (1993) 683–696. finefunctionallydistinctgamma-secretases,Proc.Natl.Acad.Sci.U.S.A.102 [2] M.E. Avery, J. Mead, Surface properties in relation to atelectasis and hyaline mem- (2005) 8972–8977. brane disease, Am. J. Dis. Child. 97 (1959) 517–523. [21] M.T. Lai, E. Chen, M.C. Crouthamel, J. DiMuzio-Mower, M. Xu, Q. Huang, E. Price, R.B. [3] A.V. Andreeva, M.A. Kutuzov, T.A. Voyno-Yasenetskaya, Regulation of surfactant se- Register, X.P. Shi, D.B. Donoviel, A. Bernstein, D. Hazuda, S.J. Gardell, Y.M. Li, cretion in alveolar type II cells, Am. J. Physiol. Lung Cell. Mol. Physiol. 293 (2007) Presenilin-1 and presenilin-2 exhibit distinct yet overlapping gamma-secretase ac- L259–L271. tivities, J. Biol. Chem. 278 (2003) 22475–22481. [4] K. Zscheppang, W. Liu, M.V. Volpe, H.C. Nielsen, C.E. Dammann, ErbB4 regulates fetal [22] K. Hoeing, K. Zscheppang, S. Mujahid, S. Murray, M.V. Volpe, C.E. Dammann, H.C. surfactant phospholipid synthesis in primary fetal rat type II cells, Am. J. Physiol. Nielsen, Presenilin-1 processing of ErbB4 in fetal type II cells is necessary for control Lung Cell. Mol. Physiol. 293 (2007) L429–L435. of fetal lung maturation, Biochim. Biophys. Acta 1813 (2011) 480–491. [5] W. Liu, E. Purevdorj, K. Zscheppang, D. von Mayersbach, J. Behrens, M.J. Brinkhaus, [23] A. Herreman, D. Hartmann, W. Annaert, P. Saftig, K. Craessaerts, L. Serneels, L. H.C. Nielsen, A. Schmiedl, C.E. Dammann, ErbB4 regulates the timely progression Umans, V. Schrijvers, F. Checler, H. Vanderstichele, V. Baekelandt, R. Dressel, P. of late fetal lung development, Biochim. Biophys. Acta 1803 (2010) 832–839. Cupers, D. Huylebroeck, A. Zwijsen, F. Van Leuven, B. De Strooper, Presenilin 2 defi- [6] N.V. Patel, M.J. Acarregui, J.M. Snyder, J.M. Klein, M.X. Sliwkowski, J.A. Kern, ciency causes a mild pulmonary phenotype and no changes in amyloid precursor Neuregulin-1 and human epidermal growth factor receptors 2 and 3 play a role in protein processing but enhances the embryonic lethal phenotype of presenilin 1 de- human lung development in vitro, Am. J. Respir. Cell Mol. Biol. 22 (2000) 432–440. ficiency, Proc. Natl. Acad. Sci. U. S. A. 96 (1999) 11872–11877. [7] C.E. Dammann, H.C. Nielsen, K.L. Carraway III, Role of neuregulin-1 beta in the devel- [24] J.I. MacDonald, F. Possmayer, Stimulation of phosphatidylcholine biosynthesis in oping lung, Am. J. Respir. Crit. Care Med. 167 (2003) 1711–1716. mouse MLE-12 type-II cells by conditioned medium from cortisol-treated rat fetal [8] G. Carpenter, ErbB-4: mechanism of action and biology, Exp. Cell Res. 284 (2003) lung fibroblasts, Biochem. J. 312 (Pt 2) (1995) 425–431. 66–77. [25] H.C. Nielsen, Epidermal growth factor influences the developmental clock regulating [9] E. Purevdorj, K. Zscheppang, H.G. Hoymann, A. Braun, D. von Mayersbach, M.J. maturation of the fetal lung fibroblast, Biochim. Biophys. Acta 1012 (1989) Brinkhaus, A. Schmiedl, C.E. Dammann, ErbB4 deletion leads to changes in lung 201–206. function and structure similar to bronchopulmonary dysplasia, Am. J. Physiol. [26] I. Gribetz, N.R. Frank, M.E. Avery, Static volume-pressure relations of excised lungs of Lung Cell. Mol. Physiol. 294 (2008) L516–L522. infants with hyaline membrane disease, newborn and stillborn infants, J. Clin. In- [10] B.K. Choi, X. Cai, B. Yuan, Z. Huang, X. Fan, H. Deng, N. Zhang, Z. An, HER3 intracel- vest. 38 (1959) 2168–2175. lular domains play a crucial role in HER3/HER2 dimerization and activation of [27] E. Kelly, H. Bryan, F. Possmayer, H. Frndova, C. Bryan, Compliance of the respiratory downstream signaling pathways, Protein Cell 3 (2012) 781–789. system in newborn infants pre- and postsurfactant replacement therapy, Pediatr. [11] W. Liu, M.A. Volpe, K. Zscheppang, H.C. Nielsen, C.E. Dammann, ErbB4 regulates sur- Pulmonol. 15 (1993) 225–230. factant synthesis and proliferation in adult rat pulmonary epithelial cells, Exp. Lung [28] J. Zhao, H. Chen, J.J. Peschon, W. Shi, Y. Zhang, S.J. Frank, D. Warburton, Pulmonary Res. 35 (2009) 29–47. hypoplasia in mice lacking tumor necrosis factor-alpha converting enzyme indicates [12] C.Y. Ni, M.P. Murphy, T.E. Golde, G. Carpenter, Gamma–secretase cleavage and nu- an indispensable role for cell surface protein shedding during embryonic lung clear localization of ErbB-4 receptor tyrosine kinase, Science 294 (2001) 2179–2181. branching morphogenesis, Dev. Biol. 232 (2001) 204–218. [13] M. Vecchi, G. Carpenter, Constitutive proteolysis of the ErbB-4 receptor tyrosine ki- [29] P.R. Murumkar, R. Giridhar, M.R. Yadav, Novel methods and strategies in the discov- nase by a unique, sequential mechanism, J. Cell Biol. 139 (1997) 995–1003. ery of TACE inhibitors, Expert Opin. Drug Discovery 8 (2013) 157–181.