Zfp148 Deficiency Causes Lung Maturation Defects and Lethality in Newborn Mice That Are Rescued by Deletion of p53 or Antioxidant Treatment

The transcription factor Zfp148 (Zbp-89, BFCOL, BERF1, htβ) interacts physically with the tumor suppressor p53 and is implicated in cell cycle control, but the physiological role of Zfp148 remains unknown. Here we show that Zfp148 deficiency leads to respiratory distress and lethality in newborn mice. Zfp148 deficiency prevented structural maturation of the prenatal lung without affecting type II cell differentiation or surfactant production. BrdU analyses revealed that Zfp148 deficiency caused proliferation arrest of pulmonary cells at E18.5–19.5. Similarly, Zfp148-deficient fibroblasts exhibited proliferative arrest that was dependent on p53, raising the possibility that cell stress is part of the underlying mechanism. Indeed, Zfp148 deficiency lowered the threshold for activation of p53 under oxidative conditions. Moreover, both in vivo and cellular phenotypes were rescued on Trp53 +/− or Trp53 −/− backgrounds and by antioxidant treatment. Thus, Zfp148 prevents respiratory distress and lethality in newborn mice by attenuating oxidative stress–dependent p53-activity during the saccular stage of lung development. Our results establish Zfp148 as a novel player in mammalian lung maturation and demonstrate that Zfp148 is critical for cell cycle progression in vivo.

Zfp148 has been targeted three times in mice with inconsistent results [11][12][13]. Takeuchi reported that haploinsufficiency for Zfp148 blocked germ cell differentiation in chimeric mice, thus precluding germline transmission of the targeted allele [13]. Another study suggested that Zfp148-deficient mice die at embryonic day (E) 9.5 with neural tube defects and anaemia [12]. Specific targeting of Zfp148 exon 4, finally, caused postnatal lethality in homozygous mice and aggravated intestinal inflammation after dextran sulphate treatment [11]. Those discrepancies could reflect different targeting strategies, but the fact that all three papers report distinct phenotypes without apparent similarities underscores that the physiological role of Zfp148 remains poorly understood.
Cell culture experiments on the other hand strongly suggest that Zfp148 is involved in cell cycle control [5,6,8,10,14,15]. For example, overexpressing Zfp148 in cancer cells induces growth arrest or apoptosis by regulation of Cdkn1a (p21) and Bak, respectively, and silencing Zfp148 induces senescence by regulation of p16 ink4a [5,6,[8][9][10]. Moreover, Zfp148 interacts with the tumour suppressor p53 and overexpression of Zfp148 increases p53 levels in the nucleus [14]. In consistency with those reports, transgenic expression of Zfp148 in intestinal epithelium induced apoptosis and reduced intestinal tumour formation in the APC Min/ + model of intestinal cancer [16]. The physiological relevance of this finding is, however, not clear since the protein was highly overexpressed.
In this study, we generate Zfp148-deficient mice and show that Zfp148 plays an important and unexpected role in lung development. Zfp148 maintains proliferation of pulmonary cells during the saccular stage of lung development by suppressing oxidative stress-dependent p53 activity, thus preventing respiratory distress and lethality in newborn mice. The results demon-strate for the first time that Zfp148 controls cell proliferation in vivo, in a p53-dependent manner.

Mice
All animal experiments were approved by the animal research ethics committee in Gothenburg, Sweden. Zfp148 +/gt mice were generated from ES cell clone XB878 (BayGenomics) and maintained on a 129P2/OlaHsd and C57Bl/6 mixed genetic background. Trp53 tm1Tyj (Trp53 +/ ) mice (The Jackson Laboratory) were maintained on a 129Sv and C57Bl/6 mixed genetic background. PCR primers used for genotyping are listed in Supporting Information Table S2.

Southern Blot Analysis
Genomic DNA was digested with EcoRI, HindIII, or KpnI and analyzed with a Zfp148 intron 4-specific probe using standard protocols.

PCR Analyses
Primers are listed in Supporting Information Table S2.

Real-time Quantitative PCR
TaqMan and Sybr Green assays were performed as described [17] using TaqMan/Sybr Green universal PCR mastermix (Applied Biosystems) and the pre-designed TaqMan assays (Applied Biosystems) or RT-PCR primers listed in Supporting Information Table S2. with the restriction enzymes E, H, and K, respectively (asterisk indicates mutant band). Fragment size in kb. (C) PCR amplification across the gene-trap insertion on tail DNA from pups of Zfp148 +/gt intercrosses confirmed gene-trap insertion at approximately 8500 base-pairs (bp) downstream of exon 4. The gene-trap specific band of 420 bp was amplified using a forward primer at 8256 bp downstream of exon 4 and a reverse primer in the genetrap. The wild-type band of 580 bp was amplified using the same forward primer and a reverse primer at 8813 bp downstream of exon 4. (D) Realtime RT-PCR of the exon 4-5 border on cDNA from tissues of adult Zfp148 gt/gt and wild-type mice, respectively. (E) Western blots showing expression of Zfp148 in Zfp148 gt/gt and wild-type mouse embryonic fibroblasts (MEFs). Actin was used as a loading control. (F) RT-PCR amplification across the indicated exon-exon borders of the Zfp148 cDNA isolated from liver and muscle in Zfp148 gt/gt and Zfp148 +/gt mice did not indicate cryptic splicing of the mRNA or alternative transcription start sites. Amplification of GAPDH was included as a loading control. (G) Whole mount image of X-Gal stained Zfp148 gt/gt mouse at E9.5. (H) Photomicrograph showing X-Gal stained transverse section of E9.5 Zfp148 gt/gt mouse. Scale bars, 100 mm. Neural tube (nt), notochord (nc) and paraxial mesoderm (pm). doi:10.1371/journal.pone.0055720.g001 Zfp148 Deficiency Causes Lung Maturation Defects PLOS ONE | www.plosone.org

X-Gal Staining
E9.5 embryos were dissected and immediately fixed at 4uC for 2 hours in 0.2% glutaraldehyde and 1.5% formaldehyde and wholemount stained at 37uC overnight as in [19].

Transmission Electron Microscopy
Lungs were fixed in 2% paraformaldehyde and 2.5% glutaraldehyde. Tissue slices were post-fixed in reduced osmium tetroxide and uranyl acetate, dehydrated, and embedded in epoxy resin. Ultrathin sections (Leica UC6 ultramicrotome) were examined in a LEO 912AB electron microscope.

Surfactant Lipid Analysis
Lipids from lung tissue were extracted as described [20]. Levels of cholesteryl esters, triglycerides and free cholesterol were determined by HPLC [21]. Analysis of phospholipids was performed by mass spectrometry [22].

Cell Culture
Primary MEFs were isolated from E13.5-14.5 embryos and cell proliferation assays were performed as described [18]. When indicated, the culture medium was supplemented with 100 mmol l 21 NAC (Sigma Aldrich) and changed daily.

Cell-cycle Analysis
Zfp148 gt/gt MEFs at passage 4 were trypsinized, fixed in 70% ethanol, incubated with propidium iodide and RNase A for 30 min at 37uC and analyzed in a FACScan flow cytometer with CellQuest Pro software (version 4.0.2, Becton Dickinson).

NAC Treatment
Pregnant females of Zfp148 +/gt intercrosses were administered NAC in the drinking water at a concentration of 1 g/l.

Generation of Zfp148-deficient Mice
To define the in vivo and cellular importance of Zfp148, we generated Zfp148-deficient mice from a gene-trap ES-cell clone. Southern blot analyses and genomic PCR confirmed that the gene-targeting vector had incorporated into the fourth intron of Zfp148, thus disrupting 87% of the coding sequence including the DNA binding zinc finger domains (Fig. 1A-C). ES-cells were injected into C57Bl/6 blastocysts to achieve germline transmission of the Zfp148 gt -allele. Heterozygous intercrosses of Zfp148 +/gt mice revealed that Zfp148 mRNA expression was reduced by .95% in Zfp148 gt/gt tissues and that the protein was undetectable in western blot analyses (Fig. 1D, E). There were no indications of exon skipping or cryptic splicing (Fig. 1F). The inserted gene-trap vector harbours a promoterless b-Geo reporter, under control of the endogenous Zfp148 promoter. X-Gal staining of embryonic day 9.5 (E9.5) Zfp148 gt/gt mice suggested that Zfp148 is ubiquitously expressed (Fig. 1G, H), in consistency with earlier results [2,4].
Zfp148 Deficiency Leads to Lethality in Newborn Mice, and Growth Retardation and Reduced Lifespan in Adult Mice Zfp148 gt/gt mice were identified at Mendelian ratios throughout embryonic development, but the frequency dropped to half the expected at postnatal day 1 (P1) suggesting that 50% of the mice died shortly after birth ( Fig. 2A). Zfp148 gt/gt mice at postnatal day P1 were pale and cyanotic indicating respiratory problems (Fig. 2B). Zfp148 gt/gt mice that survived the neonatal crisis were initially normal in size but displayed impaired growth and reduced lifespan ( Fig. 2C-F). To specifically assess the postnatal importance of Zfp148, mice that were moribund at five weeks of age with bronchioles excluded. (C) Distribution of Zfp148 genotypes of P1 pups of intercrosses between Zfp148 +/gt Trp53 +/+ and Zfp148 +/gt Trp53 +/2 mice. Scale bars, 100 mm. *P,0.05, **P,0.01, ***P,0.001. doi:10.1371/journal.pone.0055720.g007 Figure 8. Antioxidant rescue of defect lung maturation and neonatal lethality in Zfp148 gt/gt mice. (A) Photomicrographs showing lung morphology (hematoxylin and eosin; HE) and glycogen content (periodic acid-Shiff; PAS) and CC10 immunofluorescence in P1 lungs from Zfp148 +/+ , Zfp148 gt/gt , and NAC treated Zfp148 gt/gt mice, respectively (n = 6). Graphs show quantification of mean tissue area per total lung area, PAS-positive area with bronchioles excluded, and CC10-positive area with bronchioles excluded. (B) Distribution of Zfp148 genotypes of P1 pups of intercrosses between Zfp148 +/gt mice with and without NAC treatment. Scale bars, 100 mm. **P,0.01, ***P,0.001. doi:10.1371/journal.pone.0055720.g008 (study start) were excluded from the lifespan study. The study is therefore likely to underestimate mortality at young ages.

Respiratory Distress in Newborn Zfp148 gt/gt Mice is Caused by Defect Lung Maturation
The cyanotic appearance of Zfp148 gt/gt mice at P1 suggested problems with oxygenation. Histological analyses of lung tissue revealed poorly developed respiratory saccules with thickened primary septae, increased glycogen deposits and widespread ectopic expression of the Clara cell 10 kDa secretory protein CC10 (Fig. 3A). Analyses of lungs at E19.5 revealed that glycogen accumulation and ectopic CC10 expression occurred before birth (Fig. 3B). The results suggest a defect in sacculation, i.e. the transformation of distal lung buds into thin-walled terminal sacs that occurs at E18.5-P5 in mice.
Ectopic expression of CC10 protein indicated that cell differentiation could be disturbed in Zfp148 gt/gt lungs. However, differentiation of type II cells appeared normal, as judged by the expression of surfactant-related genes, the presence of lamellar bodies, and levels of the surfactant lipid dipalmitoylphosphatidylcholine ( Fig. 3C-E; Supporting Information Table S1). There was further no difference in the appearance of blood air barriers or expression of biomarker genes for type I cells (T1alpha, Aqp5) or endothelium (Pecam1, Tie2 Nos3) (Fig 3C, G). Notably, mRNA levels for the Clara cell markers CC10 and Pon1 were slightly reduced in Zfp148 gt/gt lungs, in spite of the ectopic expression of CC10 protein, indicating that Clara cell differentiation could be affected, although the cells appeared morphologically normal (3C, F). We conclude that Zfp148 deficiency prevents structural maturation of the prenatal lung without affecting type II cell differentiation or surfactant production.

Zfp148-deficiency Disrupts Cell Proliferation at the Saccular Stage of Lung Development
Zfp148 is implicated in cell cycle control suggesting that Zfp148 gt/gt lung defects could involve disturbed cell proliferation or apoptosis. To assess if cell proliferation was affected, E18.5 and E19.5 pregnant females were injected with 5-bromo-29-deoxyuridine (BrdU). Zfp148 gt/gt lungs at E19.5 contained fewer BrdUpositive cells compared to Zfp148 +/gt lungs and controls, indicating problems with cell proliferation (Fig. 4A). There was no difference in BrdU incorporation at E18.5 between Zfp148 gt/gt lungs and controls (Fig. 4B), suggesting that cell proliferation is interrupted during the saccular stage and that defective proliferation is coupled to the maturation defect. In consistency with those results, Zfp148 deficiency did not affect total lung DNA content or lung wet weight at P1 (Fig. 4C, D), as would have been expected if proliferation was disturbed at earlier stages. There was no difference in apoptosis as judged by staining for activated caspase 3 or the TUNEL assay (Supporting Information Fig. S1).

Zfp148 Deficiency Arrests Cell Proliferation in a Cell Autonomous Manner
To investigate mechanisms underlying the proliferation defect in Zfp148 gt/gt lungs, mouse embryonic fibroblasts (MEFs) were isolated from E13.5-14.5 embryos of heterozygous intercrosses. Zfp148 gt/gt MEFs proliferated poorly at early passages (Fig. 5A, B) but did not show increased apoptosis (Fig. 5C), compared to wildtype controls. Zfp148 gt/gt MEFs exhibited morphological signs of senescence and expressed the senescence marker p16 ink4a (Fig. 5D-F), in consistency with a previous study [8]. The proliferation arrest of Zfp148 gt/gt MEFs was abolished by restoring near physiological expression levels of Zfp148 by adenoviral transduction (AdZfp148) at passage 0 (Fig. 5G, H). AdZfp148 transduction at passage 2 did not reverse the proliferation arrest (Fig. 5I), suggesting that cell cycle arrest is irreversible once established. Zfp148 gt/gt MEFs were arrested regardless of their position in the cell cycle, as judged by cell cycle analysis at passage 4 (Fig. 5J).

Zfp148 Deficiency Lowers the Threshold for p53 Activation and Proliferative Arrest under Oxidative Conditions
Zfp148 deficiency is not likely to arrest cell proliferation under physiological conditions, since Zfp148 gt/gt embryos develop to term without gross malformations. The tumour suppressor p53 induces proliferative arrest in response to a wide range of stressors, including oxidative stress [23,24]. The interaction of Zfp148 with p53 opens up for the possibility that Zfp148 deficiency lowers the threshold for p53-activation. To assess this possibility, levels of phosphorylated p53 were measured under oxidative and nonoxidative conditions. We observed increased levels of phospho-p53 Ser18 (p53 Ser 15 in humans) protein in Zfp148 gt/gt MEFs cultured at 21% oxygen, but not at 3% oxygen (Fig. 6A). Similarly, levels of mRNA for the p53-target p21 were increased in Zfp148 gt/gt MEFs cultured at 21% oxygen, but not at 3% oxygen (Fig. 6B).
To assess if the proliferation arrest of Zfp148 gt/gt MEFs is mediated by p53, cells were obtained from Zfp148 gt/gt mice and controls, bred on a Trp53 null background. Loss of one or two copies of Trp53 abolished proliferative arrest in Zfp148 gt/gt MEFs (Fig. 6C, D and Supporting Information Fig. S2). In line with those results, induction of p21 was abolished in Zfp148 gt/gt MEFs on Trp53 +/2 or Trp53 2/2 background (Fig. 6B).
The selective activation of p53 in Zfp148-deficient MEFs cultured at 21% oxygen, suggest that oxidative stress is part of the underlying mechanism. To determine if oxidative stress contributes to the proliferative arrest, Zfp148 gt/gt MEFs were cultured in the presence of the antioxidant n-acetyl-L-cysteine (NAC) or at low oxygen concentrations (3%). Both treatments prevented the proliferative arrest (Fig. 6E, F).

Deletion of One or Two Copies of Trp53 Rescued Zfp148 gt/gt Mice from Proliferation Arrest, Respiratory Distress and Neonatal Lethality
We further investigated if the proliferation arrest and maturation defect of Zfp148 gt/gt lungs are mediated by p53. Indeed, loss of one copy of Trp53 restored cell proliferation in Zfp148 gt/gt lungs, as judged by BrdU incorporation (Fig. 7A). Moreover, loss of one or two copies of Trp53 prevented the thickening of the primary septae, glycogen accumulation, and ectopic CC10 expression in Zfp148 gt/gt lungs (Fig. 7B). The loss of one copy of Trp53 also rescued Zfp148 gt/gt mice from neonatal lethality (Fig. 7C).

Antioxidant Rescue of Respiratory Distress and Neonatal Lethality of Zfp148 gt/gt Mice
To determine if the respiratory distress and neonatal lethality of Zfp148-deficient mice are triggered by oxidative stress, we added NAC in the drinking water of pregnant females of heterozygous intercrosses. Strikingly, NAC treatment of pregnant females normalized lung morphology and prevented glycogen accumulation and ectopic CC10 expression of Zfp148 gt/gt offspring, and rescued them from neonatal death (Fig. 8A, B).

Discussion
In this study, we show that Zfp148 deficiency in mice causes respiratory distress and neonatal lethality, and that both of those effects were rescued on Trp53 +/2 and Trp53 2/2 genetic backgrounds and by antioxidant treatment. Mechanistic analyses showed that Zfp148 deficiency lowered the threshold for p53 activation under oxidative conditions thus disrupting cell proliferation and structural maturation of the prenatal lung.
Our data demonstrate that Zfp148 deficiency activates p53, and that structural maturation defects in Zfp148-deficient lungs are secondary to unchecked p53 activity. The mechanism by which p53 disrupts lung maturation is, however, not clear. Proliferation arrest of pulmonary cells probably contributes to the phenotype, but effects of p53-activation on cytoskeletal remodeling or cell migration, as described in [34][35][36][37], may play a role. Effects on cell differentiation are also possible, as judged by the ectopic expression of CC10 and reduced expression of Clara cell biomarker genes. Nevertheless, the rescue of Zfp148 gt/gt lungs upon deletion of Trp53 or antioxidant treatment shows that Zfp148 deficiency affects cell stress or stress responses, whereas Zfp148 is dispensable for lung development per se.
The results show for the first time that Zfp148 is required for cell proliferation in vivo, confirming previous in vitro data [8]. Cell proliferation was selectively arrested in prenatal lungs whereas the embryo as a whole developed to term without gross malformations, suggesting that Zfp148 is required for proliferation in certain contexts. The antioxidant rescue of Zfp148-deficient lungs and the rescue on Trp53 +/2 and Trp53 2/2 genetic backgrounds, suggest that the main function of Zfp148 is to protect cells from p53activation under oxidative conditions. In consistency with those results, Zfp148 deficiency lowered the threshold for p53-activation by oxygen in MEFs. Although Zfp148 deficiency may alter redox homeostasis, the previously demonstrated binding between Zfp148 and p53 opens up for the possibility that Zfp148 primarily regulates p53 [14]. Our demonstration of genetic interaction between Zfp148 and Trp53 in mice supports this possibility. Formal evidence of the proposed regulation will, however, require generation of Zfp148 mutants with deficient p53-binding. Future experiments should also establish whether concentrations of reactive oxygen species are elevated in Zfp148-deficient cells.
Prenatal lungs are particularly sensitive to the loss of Zfp148 and tissue hypoxia could be a contributing factor. Hypoxia increases production of reactive nitrogen or oxygen species in the mitochondria thus generating oxidative stress [38][39][40][41]. Mice targeted for Hypoxia-inducible factor 2a (Hif2a), or mice lacking Hif1a in airway epithelium, exhibit structural maturation defects reminding of those seen in Zfp148-deficient mice with additional effects on epithelial cell differentiation and surfactant production [25,42]. Hypoxia-inducible factors are only active under hypoxic conditions [43,44], suggesting that hypoxia is a driving force of prenatal lung maturation. Thus, levels of reactive nitrogen or oxygen species could be elevated in pulmonary cells at this stage, compared to other cells in the embryo, making them vulnerable to Zfp148 deficiency.
The defects of Zfp148-deficient lungs are reminiscent of clinical findings in newborns with bronchopulmonary dysplasia (BPD).
BPD is a developmental disorder with disrupted postnatal growth of the distal lung caused in part by oxidative damage inflicted by oxygen supplementation of preterm infants [45,46]. Our finding, that Zfp148 protects the prenatal lung from p53-induced proliferation arrest in response to oxidative stress in mice, opens up for the possibility that Zfp148 plays a role in the oxidative damage of preterm lungs in BPD. Studies in primates and mice show increased expression of p53 and p21 after postnatal oxygen supplementation [47][48][49], supporting that possibility.
The physiological role of Zfp148 has remained uncertain due to inconsistent results of gene targeting experiments in mice. A previous attempt to target Zfp148 resulted in sterile chimeric mice and it was suggested that haploinsufficiency for Zfp148 blocks germ cell differentiation [13]. Our study clearly demonstrates that Zfp148-heterozygous mice are fertile and that Zfp148 is dispensable for early mouse development. One possible explanation for this discrepancy is that the targeting strategy used by Takeuchi and co-workers (deletion of exon 9) generates a dominant negative phenotype. Alternatively, the 129P2/OlaHsd embryonic stem cell line that we used is more robust and maintains pluripotency better upon Zfp148 deficiency, compared to the AB1 strain used in Takeuchi's study. We conclude that in the current study, Zfp148 mRNA levels in Zfp148 gt/gt cells were reduced by .95% and the Zfp148 protein was undetectable.
Another study showed that mice generated from the same Zfp148 gene-trap mutation used in our study die at E8.5-E10.5 with unclosed neural tubes and anaemia [12]. We observed this phenotype in embryos of the first generation (F1) intercrosses. Importantly, this phenotype co-segregated with a gene-trap vector but not with the recombined Zfp148 locus (Supporting Information Fig. S3A-C). Moreover, the gene-trap vector was propagated to 79% of the brown offspring (F1 mice) of crosses between chimeric mice and C57Bl/6 mice. This deviates from the expected Mendelian distribution (62 of 78 mice, P = 1.5 610 27 , binomial distribution) but is consistent with a second integration of the genetrap vector in the embryonic stem cell line. Thus, the embryonic death is not caused by Zfp148 deficiency.
In summary, we show that Zfp148 is required for structural maturation of the prenatal lung by preventing oxidative stressdependent p53 activity during the saccular stage of lung development. The result demonstrates for the first time that Zfp148 plays a critical role for cell cycle progression in vivo, and establishes Zfp148 as a novel factor in mammalian lung development. Figure S1 Respiratory distress in Zfp148-deficient mice is not caused by apoptosis. TUNEL and cleaved caspase 3 (cl-CASP3) staining in lungs of P1 Zfp148 gt/gt and wt mice (n = 6). Sections of thymus from a 3-week-old wild-type mouse were used as a positive control. (TIF) Figure S2 Wt and Zfp148-deficient MEF proliferation on Trp53 +/2 or Trp53 2/2 backgrounds (n = 3). (TIF) Figure S3 Integration of a second gene-trap in the XB878 ES cell clone. (A) Dissection of 21 E9.5 embryos of F1 generation intercrosses identified a proportion of embryos with unclosed neural tubes and variable degrees of additional defects similar to those described in [12]. Importantly, these mice had at least one intact Zfp148 allele as judged by PCR-amplification of a DNA fragment spanning the gene-trap insertion site of the Zfp148 locus (Zfp148wt PCR). Moreover, the gene-trap vector was propagated to 79% of the brown offspring (F1 mice) of crosses between chimeric mice and C57Bl/6 mice, which deviates from the expected Mendelian distribution (62 of 78 mice, P = 1.5 x 10 27 , binomial distribution) but is consistent with the presence of two gene-trap alleles in the injected ES-cells. (B, C) Dorsal view of E9.5 embryos exhibiting unclosed neural tubes (arrows). (TIF) Table S1 HPLC and mass spec analysis of surfactant lipids from P1 lungs from wt and Zfp148 gt/gt mice, first two columns are means for respective genotypes (n = 4) followed by p-value. (PDF)