CCAAT/Enhancer-Binding Protein-α Suppresses Lung Tumor Development in Mice through the p38α MAP Kinase Pathway

The transcription factor CCAAT/enhancer-binding protein α (C/EBPα) is a basic leucine zipper transcription factor and is expressed in alveolar type II cells, alveolar macrophages and Clara cells in the lung. Although decrease or absence of C/EBPα expression in human non-small cell lung cancer suggests a possible role of C/EBPα as a lung tumor suppressor, there is no direct proof for this hypothesis. In this study, we investigated, for the first time, the role of C/EBPα in lung tumors in vivo using transgenic mice with lung epithelial specific conditional deletion of Cebpa (CebpαΔ/Δ mice) and a urethane-induced lung tumor model. C/EBPα expression in the lung was dispensable, and its deletion was not oncogenic under unstressed conditions. However, at 28 wk after urethane injection, the number and size of tumors and the tumor burden were significantly higher in CebpαΔ/Δ mice than in littermate control mice. Urethane-injected CebpαΔ/Δ mice showed highly proliferative adenomas and adenocarcinomas in the lung, and survival time after urethane-injection was significantly shorter than that in control mice. In control mice, C/EBPα was strongly induced in the tumor tissues at 28 weeks after urethane-injection, but became weakened or absent as tumors progressed after long-term observation for over 1 year. Using intraperitoneal injection of p38 inhibitor (SB203580), we demonstrated that the induction of C/EBPα is strongly regulated by the p38 MAP kinase in murine alveolar epithelial cells. A high correlation was demonstrated between the expression of C/EBPα and p38α MAP kinase in tumor cells, suggesting that C/EBPα silencing in tumor cells is caused by down-regulation of p38α MAP kinase. In conclusion, the role of C/EBPα as a lung tumor suppressor was demonstrated for the first time in the present study, and the extinguished C/EBPα expression through p38α inactivation leads tumor promotion and progression.


Introduction
Exposure to carcinogens, including ionizing radiation, viral infection, and tobacco smoking causes cumulative changes in the DNA of lung tissue [1] and initiates lung cancer by activating oncogenes and/or inactivating tumor suppressor genes [2]. CCAAT/enhancer-binding protein-a (C/EBPa) is a basic leucine zipper transcription factor that is expressed in many tissues [3]. C/ EBPa plays an important role in normal tissue development, namely, in the regulation of cell proliferation and cell differentiation [4,5].
In the lung, C/EBPa is expressed in alveolar type II cells (ATII cells), Clara cells, and alveolar macrophages from late gestation through adulthood [6][7][8][9]. The mechanisms that regulate lung epithelial development are often linked to lung injury-repair processes and lung disease. Experiments in transgenic mice with lung epithelial specific conditional deletion of C/EBPa (Cebpa D/D mice) have proved that C/EBPa is required for pulmonary maturation in late gestation [7] and potentially plays an emergent role to maintain lung homeostasis following lung injury and repair in adult mice [8,9].
Previous studies have demonstrated decreased or absent of C/ EBPa expression in 50% of stage II and IIIA lung adenocarci-nomas, suggesting that C/EBPa may function as a tumor suppressor in the lung [10][11][12]. However, direct evidence for this hypothesis is lacking, and the mechanism underlying the decrease in C/EBPa expression in lung adenocarcinoma has not been well studied. Unlike hematopoietic malignancies, in which the CEBPA gene has been demonstrated to be mutated [13], CEBPA mutations are rare in non-small cell lung cancer (NSCLC) [14]. Absence of C/EBPa by the aberrant DNA methylation of C/ EBPa promoter has been observed in vitro in human lung cancer cell lines and in human lung cancer tissues [12]. In addition, down-regulation of C/EBPa by highly expressed tribbles homo-log2 (TRIB2) has been shown in lung cancer cell lines [15]. In animal model, MAP kinase14 D/D Kras G12V mice with extensive lung stem cell numbers and tumor progression have demonstrated decreased C/EBPa expression in vivo, suggesting that C/EBPa is a possible downstream target of MAPK14 (p38a) in the lung [16].
In the present study, we tested the hypothesis that C/EBPa plays a role in the suppression of lung tumorigenesis in vivo. The deletion of C/EBPa in Cebpa D/D mice strikingly increased the urethane-induced lung tumor incidence and the malignant tumors. Furthermore, we demonstrated that the expression of p38a mitogen activated protein kinase (p38a MAP kinase) is required for C/EBPa expression in the lung.

Materials and Methods
This study was carried out in strict accordance with the recommendations in the Guide for the Care and Use of Laboratory Animals of the National Institutes of Health. All study protocols were approved by the Institutional Animal Care and Use Committee of the Cincinnati Children's Hospital Research Foundation (Permit Number: 9D05044). All surgery was performed under sodium pentobarbital anesthesia, and all efforts were made to minimize suffering.
Doxycycline (625 mg/kg; Harlan Teklad, Madison, WI) administered in the chow to the dams from embryonic day (E)0 to postnatal day (P)14 (the early-deletion study), provided sufficient doxycycline to the newborns through the milk ( Figure S1A).
For the late-deletion study, doxycycline-containing chow was given to Cebpa D/D mice for 3 wk beginning at 14 wk of age, which was 8 to 11 wk after urethane injection. The comparison group of Cebpa D/D mice was given no-doxycycline containing normal food (No-Dox group).

Tumor Induction by Urethane
For both the early-deletion and late-deletion studies, 1 mg/g body weight of urethane (Sigma, St. Louis, MO) dissolved in 0.9% NaCl (0.1 ml/g) was intraperitoneally injected to 6 week old mice twice. The second dose was injected 3 days after the first injection. The general condition of each mouse was monitored daily and the body weight was recorded weekly. Mice were euthanized at 10, 20, and 28 wk after the first urethane injection. In the long term study, control mice were euthanized when they lost more than 10% of their body weight at 6 wk of age.
The tumors on the lung surface were counted and diameter (r) was measured using microcallipers under a dissecting microscope [17]. The volume of tumor was calculated as T(n) = 3/4pr (n) 3 , where (n) is the number of tumors in the mouse and the tumor burden was calculated as P n k~1 T(k). Fixed lung tumor sections were stained by haematoxylin and eosin (H&E) and were histologically classified according to the previously recommended criteria [18].
Mitotic index was calculated to evaluate cell proliferation in tumors. Mitotic index (%) = Ki-67 positive cells/total cells 6100.
TUNEL staining was performed using a TumorTACS TM In Situ Apoptosis Detection Kit (TREVIGEN, Gaithersburg, MD) according to the manufacturer's protocol.

Real-time PCR
Homogenized left lung lobes were lysed in buffer and total RNA was isolated using an RNeasy Plus Mini-Kit (QIAGEN, Valencia, CA). Selected tumor samples on slide from paraffin embedded sections were isolated and RNA extraction was performed using the FFPE RNeasy-kit (Invitrogen). One microgram of total RNA was reverse transcribed to cDNA using Superscript VILO (Invitrogen). Quantitative

Administration of p38 MAP Kinase Inhibitor
Control mice (6 wk of age) were injected intraperitoneally with p38 MAP kinase inhibitor, SB203580 (1 mM/kg, LC Laboratories, Woburn, MO), or 0.9% NaCl 3 times, at 12 hours intervals. The mice were euthanized for analysis 3 hours after last injection.

DNA Methylation Analysis
All tumors were identified according to C/EBPa expression by immunohistochemistry. DNA and RNA analysis was performed on material isolated by hand under a dissection microscope from 4 serial slides for each tumor. DNA was isolated from each single tumor on paraffin sections using the FitAmp TM Paraffin Tissue Section DNA Isolation Kit (EPIGENTEK). For promoter methylation analysis, bisulfite modification was performed using Bisulflash TM DNA Modification Kit (EPIGENTEK). A DNA fragment in the C/EBPa promoter region was amplified by PCR using TaqGold PCR kit (Applied Biosystems) with primers designed by the web-based program MethPrimer (http://www. urogene.org/methprimer/index.html) (primer sequences: 2337to2140: 59GTTTTGGAAAGTTATAGGAGAAGG, 39CACCCAATACCCCAACTAACTC, 2140to+49: 59TTTTTAGTGTTGGTTGGAAGTG, 39CCTTCTCCTA-TAACTTTCC). The obtained DNA fragments were ligated with pGEM-T vector (Promega, Madison, WI) and transfected into Escherichia coli. Plasmid DNA purified from each colony was submitted to the Cincinnati Children's Hospital Medical Center DNA Sequence Core Unit for sequencing.

Alveolar type II cell isolation
Alveolar type II cells were isolated from lungs using collagenase as previously reported [8].

Statistical Analysis
Data were expressed as the means 6 SEM. Statistical analyses were performed using GraphPad Prism 5.0c (GraphPad Software, La Jolla, CA). Comparisons between two groups were evaluated by Mann-Whitney U test. Non-parametric correlation analysis was performed using Spearman's correlation test. For multiple comparisons, analysis of variance (ANOVA) was used, followed by Bonferroni's post hoc test for significance. Kaplan-Meier survival analysis was used to compare the lifespans between groups.

Deletion of C/EBPa in the Lung is not Oncogenic in Mice
In the present study, the Scgb1a1-rtTA promoter was used to induce lung epithelial specific deletion of C/EBPa. Scgb1a1 is expressed in Clara cells from E16-E17 of gestation, and in ATII cells after birth. Thus, using this Scgb1a1-rtTA system for Cebpa D/D mice, the deletion of C/EBPa occurred in these Scgb1a1expressing cells ( Figure S1B). Lung structure was normal and no cell hyper-proliferation was observed in the lungs of Cebpa D/D mice as reported previously [9]. C/EBPa was clearly deleted from ATII cells at 6 wk of age.
Cancer development occurs in three major stages: initiation, promotion, and progression. To evaluate whether the absence of C/EBPa initiates tumor or not, spontaneous lung tumor in lung was observed up to 18 mo of age. We demonstrated that spontaneous lung tumors were not detected at 4 and 8.5 mo of age in both control and Cebpa D/D mice (n = 6/group). All mice enrolled in 18 months study (n = 25/group) survived for 18 months and spontaneous lung tumors were detected in 4 out of 25 control mice (16%) and 5 out of 25 Cebpa D/D mice (20%). The incidence of spontaneous lung tumors in this study was similar to that in previous reported study using FVB/N mice (13% at 14 mo of age and 41% at 24 mo) [19]. Thus, the deletion of C/EBPa in the lung epithelial cells did not alter the incidence of spontaneous lung tumors, demonstrating that the deletion of C/EBPa in the lung is not oncogenic in mice.

Urethane-induced Lung Tumor Development is Significantly Increased in Cebpa D/D Mice
To investigate the role of C/EBPa in primary lung tumor promotion and progression, the lung tumor inducer urethane was injected into both control and Cebpa D/D mice at 6 wk of age. Three parameters for tumor enumeration including the number of tumors on the lung surface, the average diameter and the tumor burden were evaluated at 10, 20, and 28 wk after urethane injection (n = 10-17/group). All three parameters were similar at 10 wk, but significantly higher in Cebpa D/D mice at 20 and 28 wk after urethane injection compared to control mice (p,0.01) ( Figure 1A). The tumors grew from 20 to 28 wk by merging with neighboring tumors, which resulted in larger tumor sizes and greater tumor burden, but fewer individual tumors at 28 wk compared to 20 wk after urethane injection. As shown in the representative photographs, lung tumors were larger and more numerous in the lung of Cebpa D/D mice compared to control mice 28 wk after urethane injection ( Figure 1B, left column). All of the tumors in control mice were histologically classified as adenomas at 28 wk after urethane injection by H&E staining. The tumors over 5 mm in diameter that invades at least one large airways or the pleura were diagnosed as adenocarcinomas [18]. In Cebpa D/D mice, at least one adenocarcinoma was detected on each of the H&E stained slides ( Figure 1B, middle column). Urethane-induced tumors in control mice at 28 wk after injection showed strong expression of C/EBPa protein ( Figure 1B, right column) and Cebpa mRNA was significantly increased in urethane-injected control mice ( Figure 1C). Furthermore, Cebpa mRNA expression was enriched in tumor tissues compared to non-tumor lung tissues in control mice ( Figure 1C). The expression of Cebpa mRNA remained low in Cebpa D/D mice after urethane injection ( Figure 1C). In survival study, Cebpa D/D mice showed significantly shorter survival than control mice after urethane injection (control mice: 63.562.7 wk, Cebpa D/D mice: 28.560.9 wk, p,0.00001, n = 20/group) ( Figure 1D).
The proliferation of lung tumor cells was evaluated by immunohistochemistry for Ki-67 and the mitosis-specific marker phospho-histone 3 (pH 3). Both markers revealed more mitotic cells in the tumors in Cebpa D/D mice than in control mice (Figure 2A). Mitotic index in the tumors was significantly higher in Cebpa D/D mice than in control mice at 20 wk after urethane injection, suggesting that the deletion of C/EBPa enhanced tumor proliferation ( Figure 2B). As in the previous study of urethaneinduced tumor model [20], apoptotic cells as evaluated by TUNEL and cleaved caspase-3 staining was rare and was not strongly detected in either control or Cebpa D/D mice. Neither TUNEL nor cleaved caspase-3 staining was affected by the deletion of C/EBPa ( Figure 2C). Therefore, the deletion of C/ EBPa affected tumor cell development primarily by enhancing cell proliferation.

Promotion and Progression of Lung Tumor Stimulated by the Late-Deletion of Cebpa
To study how absence of C/EBPa affects tumor development, C/EBPa was deleted from the lung at 8 wk after urethane injection by administrating doxycycline at 14 wk of age for a total of 3 weeks ( Figure 3A). Three weeks of doxycycline treatment was required to significantly delete C/EBPa from the lung of Cebpa flox/flox mice ( Figure 3B). Lung tumors were evaluated at 20 wk after urethane injection (9 wk after C/EBPa deletion). Late-deletion of C/EBPa significantly increased the number and size of lung tumors and the tumor burden in on-Dox Cebpa D/D mice compared to no-Dox Cebpa flox/flox mice ( Figure 3C). Tumor proliferation was also significantly enhanced by C/EBPa deletion ( Figure 3B, D). Due to the shorter period of C/EBPa deletion, the tumor size and tumor burden in the late-deletion model were both smaller than those in the early-deletion model at 20 wk after urethane injection ( Figure 1B). Interestingly, doxycycline administration rapidly increased the number of tumors in the late-deletion model and the number of tumors at 20 wk was similar to that of the earlydeletion model (the number of tumors; late vs. early: 14.461.7 vs. 11.561.7, p = 0.25). These results indicated that loss of C/EBPa promotes tumor promotion and progression rather than tumor initiation.

C/EBPa Expression in Tumor was Lost during Tumor Progression
Although C/EBPa was strongly induced in all tumors in control mice at 28 wk after urethane injection, we found that this expression became weakened or absent as tumors progressed, similar results observed in human NSCLC [10,11]. To clarify the mechanisms of decreased C/EBPa in progressed lung tumors, urethane injected control mice (at 6 wk of age) were followed until more than 10% of their body weight at 6 wk of age was lost. Average date of euthanasia was 63.761.2 wk after urethane injection. All of the 26 control mice had lung adenomas and/or adenocarcinomas evaluated by H&E staining ( Figure 4A, upper panels). Immunohistochemistry of C/EBPa demonstrated 3 staining patterns: 1) entirely-positive, 2) partially-negative and 3) entirely-negative in lung tumors ( Figure 4A, lower panels). Partially-negative C/EBPa tumors were found in all of the 26 mice, while 6 out of 26 mice had C/EBPa entirely-negative tumors.

Loss of C/EBPa in Tumors is not by Aberrant Promoter DNA Methylation in Mice
To investigate the relation between C/EBPa expression and DNA methylation, tumors were stained with 5-mC. C/EBPanegative cells in tumors were highly methylated compared to C/ EBPa positive tumor cells detected by immunofluorescence of 5-mC ( Figure 4B). Our findings suggest that C/EBPa expression might be lost from lung tumor cells by aberrant promoter DNA methylation as previously shown in human NSCLC [12]. Because methylation is more critical in the core region than in non-core region for DNA silencing in cancer cells [21,22], the methylation of the core region in the Cebpa gene sequence was evaluated using isolated C/EBPa entirely-negative tumors. C/EBPa entirelypositive tumor tissues isolated from control mice at 28 wk after urethane injection were used for comparison. Contrary to our assumption, our mouse model represented no specific aberrant DNA methylation in isolated C/EBPa entirely-negative tumors ( Figure 4C). Thus, there was no relationship between aberrant DNA methylation of core region and C/EBPa absence in tumor.

C/EBPa is Induced by p38 MAP Kinase Activation in Lung Tumors
The strong C/EBPa expression in lung tumors was lost through an unknown mechanism as the tumors progressed. To determine if p38 MAP kinase regulates C/EBPa expression in normal lung, the p38 MAP kinase inhibitor (SB203580) was intraperitoneally injected into control mice and C/EBPa expression was determined by immunohistochemistry. As shown in Figure 5A and 5B, SB203580 successfully inhibited C/EBPa in alveolar type II cells and significantly inhibited Cebpa mRNA expression in the lungs. These data demonstrate that alveolar epithelial C/EBPa is regulated by p38 MAP kinase in vivo.
To investigate the mechanism of extinguished C/EBPa expression in tumors of long-observed model, total RNA was isolated from paraffin-embedded adenocarcinoma tissue from 13 urethane-injected mice followed by analysis of Cebpa and MAPK14 by qRT-PCR. Although MAPK14 (p38a) is a known regulator of C/EBPa expression in lung stem cells during cell differentiation [16,23], its role in lung tumor cells are unknown. We found that Cebpa mRNA expression strongly correlated with MAPK14 mRNA in tumors (r 2 = 0.832, p,0.0001, Spearman's correlation) ( Figure 5C). As demonstrated by double immunofluorescence of C/EBPa and p38 MAP kinase, the majority of the tumor cells were either double positive or double negative ( Figure 5DC). Collectively, these data suggest that the p38a MAP kinase pathway is a key regulator of C/EBPa expression in urethane-induced tumors and that decreased or absence of p38a MAP kinase in adenocarcinoma regulates C/EBPa expression in tumors.

Discussion
In the present study, we demonstrated that decreased C/EBPa expression in the lung epithelial cells does not initiate lung tumor, but enhances tumor promotion and progression which results in short survival in mice. No differences for spontaneous tumor occurrence were detected after the long term C/EBPa deletion in the lung. Therefore, the deletion of C/EBPa is not oncogenic. Although C/EBPa expression was dispensable for lung under the normal condition, C/EBPa was strongly induced in urethaneinduced tumors and suppressed tumor promotion/progression by regulating cell proliferation. This is the first study to directly  demonstrate that C/EBPa exhibits lung tumor suppressor activity in vivo.
Frequency of spontaneous tumors and susceptibility to urethane differ between mouse strains. Urethane is metabolized by CYP2E1 to epoxide, which interacts with DNA to form carcinogenic DNA adducts [24]. CYP2E1 and NFkB activities determine susceptibility to urethane in mice [17,24]. Although FVB/N mouse has higher susceptibility and C57BL6/J has less susceptible to urethane, our mice showed consistent susceptibility for tumor formation, suggesting a homogeneous mixed-strain background. C/EBPa has been shown to have a critical role in the lung under abnormal conditions such as lung injury and repair. Our previous studies highlighted the critical cytoprotective functions of C/EBPa in the lung and indicated that it may suppress tumor development. C/EBPa protects the lung from hyperoxia by regulating surfactant synthesis and oxidative sensor [8], suggesting that changes in the redox status might affect tumor development in mice. C/EBPa also regulates the protease/anti-protease balance in Clara cells during bronchiolar injury and repair [25]. Cebpa D/D mice lack lympho-epithelial Kazal-type-related inhibitor (LEKTI, coded by Spink5), which is a serine protease inhibitor that strongly inhibits kallikrein-related peptidase (KLK) activity [9,26]. Although we have not evaluated the activity of LEKTI in tumors, Spink5 mRNA was significantly up-regulated in control mice at 28 wk after urethane injection, but unchanged in Cebpa D/D mice ( Figure S2). The tumor derived KLKs promote tumor cell migration by activating PAR1 [27,28], therefore, LEKTI might be a potential functional downstream target of C/EBPa.
The tumor suppressor function of C/EBPa has been well documented in non-lung tissues, particularly in the liver, skin and myeloid malignancies [29][30][31]. In our study, tumors in Cebpa D/D mice showed increased proliferation determined by increased Ki-67 and mitosis-specific pH 3 expression. The cell cycle regulatory function of C/EBPa has been demonstrated in non-lung cells by interacting with p21 [13], E2Fs [32,33], Cdk2 [34], Cdk4 [34], SWI/SNF complex [35], Rb family members [36], and 1, 25 2 dihydroxyvitamin D3 [13,[32][33][34][35][36][37], but its role in lung cells remains poorly understood. Unlike hematopoietic cells, the human lung cancer cell line H358 did not exhibit any changes in the cell cycle upon transfection with CEBPA, and the CEBPA transgene stimulated apoptosis only in the context of toxic stimulation by zinc [3,11]. We evaluated the inhibition of CEBPA by small interfering RNA in A549 cells, but cell proliferation evaluated by XTT assay was same as control cells (Data not shown). Since C/ EBPa has an emergent role, urethane-induced inflammation might modify C/EBPa by the mechanisms such as phosphorylation, sumoylation or binding to co-factors as previously reported [38][39][40][41][42].
The present study demonstrated that p38 MAP kinase activation is required for C/EBPa expression in alveolar epithelial cells and that Mapk14 expression was strongly correlated with Cebpa expression in tumors. The interaction between the two factors has been studied for cell differentiation in non-lung cells including keratinocyte, hepatocyte and neutrophils [43][44][45]. Hui et al. reported that MAPK14 D/D mice died before postnatal day 4 due to respiratory distress with impaired ATII cell maturation [46], which is the same phenotype observed when Cebpa was deleted embryonically [7]. In another study, the K-Ras G12Vinduced tumor model intercrossed with MAPK14 D/D mice (MAPK14 deletion in adult) demonstrated high tumor frequency with low expression of C/EBPa and FoxA2 in the lung [16]. This study also demonstrated the down-regulation of p38a in human NSCLC [16], consistent with the results in our long-observed study. These two studies focused on lung stem cell or cancer stem cell differentiation in response to p38a MAP kinase pathway and claimed that p38a MAP kinase inactivity by using p38 MAP kinase inhibitors impaired lung stem cell differentiation and increased the tumor susceptibility [16,23]. In the present study, we demonstrated that p38 MAP kinase regulates C/EBPa in normal lung epithelial cells and that C/EBPa expression in tumors was silenced as tumors progressed in long-observed model. Furthermore, CC10 + and SP-C + double positive putative stem cells were rarely observed in urethane-induced tumors, either in Cebpa D/D mice or long-observed control mice group (data not shown), suggesting that the tumor progression in Cebpa D/D mice might not be associated with lung stem cell expansion. In long-observed model, we could not find the difference in survival time after urethane-injection and mitotic index in tumor cells between having and not having C/EBPa negative tumor in mice (data not shown), suggesting that C/EBPa silencing is not an only factor determining the lifetime risk due to lung malignancy, which is consistent with the previous study for NSCLC [10].
p38 MAP kinase inhibitors have been under the clinical trials for cardiovascular disease and chronic obstructive lung disease, and one of p38 MAP kinase inhibitors demonstrated significant anti-inflammatory effects by ameliorating disease biomarkers [47,48]. Although we showed C/EBPa silencing is not oncogenic in the lung, a potential risk of C/EBPa silencing in the lung by treatment with p38 MAP kinase inhibitors should be given a caution and these patients require screening for lung cancer. A dose threshold of p38 MAP kinase inhibitor for anti-inflammation and the silencing of C/EBPa and p38a expression should be evaluated to avoid lung cancer progression.
FoxA2, also regulated by p38a MAP kinase, is downregulated in NSCLC and is a candidate lung tumor suppressor [49]. The induction of FoxA2 in H358 cells has been shown to result in proliferation arrest and apoptosis in vitro [50]. Although FoxA2 is described as a downstream target of C/EBPa in the fetal lung and in vitro [7,50], FoxA2 was strongly expressed in the urethaneinduced tumors of Cebpa D/D mice ( Figure S3A) and Foxa2 mRNA was also induced in the tumors from Cebpa D/D mice ( Figure S3B), suggesting that FoxA2 does not suppress urethane-injected tumor in Cebpa D/D mice and is not a downstream target of C/EBPa in adult mice.
The mechanisms of C/EBPa absence in NSCLC have been demonstrated by using lung cancer cell lines. TRIB2 was demonstrated as a suppressor of C/EBPa in human NSCLC [15], but Trib2 mRNA expression in urethane-induced tumors of our long-observed model was below the detection limit compared to normal lung tissue (Data not shown). In spite of high 5-mC expression in C/EBPa negative tumor cells, our data showed that core region of Cebpa was not strongly methylated. Since Mapk14 expression was downregulated similarly as Cebpa expression in tumors, we assumed that signal upstream of MAP kinase14 might be silenced by DNA methylation. One limitation of this study is the use of a mouse model. In humans, the CEBPA gene is located on chromosome 19, whereas in mice, the Cebpa gene is located on chromosome 7. Furthermore, the promoter sequence is not identical between humans and mice. Thus, susceptibility for the promoter DNA methylation might be different between species. Our study demonstrated that C/EBPa absence was not due to the aberrant DNA methylation of the core promoter region in a murine model, but promoter hypermethylation cannot be ruled out as a mechanism of C/EBPa silencing in human tumors.
In summary, the present study demonstrated a cell specific role of C/EBPa as a lung tumor suppressor in vivo. C/EBPa suppressed tumor promotion/progression in a urethane-induced tumor model by regulating cell proliferation. C/EBPa expression in lung epithelial cells was regulated by p38a MAP kinase and p38 MAP kinase inactivation resulted in absence of C/EBPa expression. The p38a MAP kinase pathway is critical for tumor suppression and C/EBPa is one of the effective downstream targets of this pathway. Re-activation of the p38a MAP kinase pathway might be a useful therapy for lung cancer. Figure S1 Lung epithelial C/EBPa deletion in Cebpa D/D mice. (A): Triple transgenic system for the lung epithelial specific deletion of C/EBPa by doxycycline administration. By using Scgb1a1 promoter, Cre was expressed in lung epithelial cells. The targeting construct deletion was mediated by Cre/LoxP system. (B): Cebpa expression by qRT-PCR in whole lungs and isolated type II cells. Cebpa expression is significantly lower in Cebpa D/D mice in both lungs and isolated ATII cells (*p,0.01, n = 4/group). (TIF) Figure S2 mRNA expression of Spink5 in whole lungs. In control mice, Spink5 expression at 28 wk was significantly higher in urethane-injected mice than saline-injected mice (*p,0.05, n = 4/ group). (TIF) Figure S3 Significant FoxA2 expression in Cebpa D/D mice. (A): Double immunofluorescence of C/EBPa and FoxA2 in lung tumors at 28 wk after urethane injection. In control mice, both C/EBPa and FoxA2 were expressed in tumor. Although C/ EBPa was absent in the tumors of Cebpa D/D mice, FoxA2 was strongly expressed in tumors, suggesting that FoxA2 expression is independent of C/EBPa. Tu: tumor (B): Foxa2 mRNA expression in whole lungs was significantly up-regulated in both control and Cebpa D/D mice 28 wk after urethane injection (n = 4/group, *p,0.05). This expression was significantly stronger in tumor tissues than in non-tumor tissues in both control and Cebpa D/D mice (n = 4/group, *p,0.05). (TIF)