|Home | About | Journals | Submit | Contact Us | Français|
PRKAR1A inactivation leads to dysregulated cAMP signaling and Carney complex (CNC) in humans, a syndrome associated with skin, endocrine and other tumors. The CNC phenotype is not easily explained by the ubiquitous cAMP signaling defect; furthermore, Prkar1a+/− mice did not develop skin and other CNC tumors. To identify whether a Prkar1a defect is truly a generic but weak tumorigenic signal that depends on tissue-specific or other factors, we investigated Prkar1a+/− mice when bred within the Rb1+/− or Trp53+/− backgrounds, or treated with a two-step skin carcinogenesis protocol. Prkar1a+/− Trp53+/− mice developed more sarcomas than Trp53+/− mice (P < 0.05) and Prkar1a+/− Rb1+/− mice grew more (and larger) pituitary and thyroid tumors than Rb1+/− mice. All mice with double heterozygosity had significantly reduced life-spans compared with their single-heterozygous counterparts. Prkar1a+/− mice also developed more papillomas than wild-type animals. A whole-genome transcriptome profiling of tumors produced by all three models identified Wnt signaling as the main pathway activated by abnormal cAMP signaling, along with cell cycle abnormalities; all changes were confirmed by qRT–PCR array and immunohistochemistry. siRNA down-regulation of Ctnnb1, E2f1 or Cdk4 inhibited proliferation of human adrenal cells bearing a PRKAR1A-inactivating mutation and Prkar1a+/− mouse embryonic fibroblasts and arrested both cell lines at the G0/G1 phase of the cell cycle. In conclusion, Prkar1a haploinsufficiency is a relatively weak tumorigenic signal that can act synergistically with other tumor suppressor gene defects or chemicals to induce tumors, mostly through Wnt-signaling activation and cell cycle dysregulation, consistent with studies in human neoplasms carrying PRKAR1A defects.
Carney complex (CNC) is a multiple neoplasia syndrome that is inherited in an autosomal dominant manner and is characterized by several types of skin tumors and pigmented lesions, myxomas, schwannomas and endocrine neoplasms (1,2). Inactivating mutations of the PRKAR1A gene coding for the 1-α regulatory (RIα) subunit of protein kinase A (PKA) are responsible for the disease in most CNC patients (3,4). PRKAR1A mutations lead to a net increase in PKA activity and, hence, enhance cAMP signaling (3). Accordingly, Prkar1a haploinsufficiency in mice led to the development of tumors arising in cAMP-responsive tissues, such as the bone, Schwann and thyroid follicular cells (5).
However, Prkar1a+/− mice did not develop some of the most frequent CNC tumors, such as skin and heart myxomas, and pituitary adenomas; in addition, all mouse tumors were observed relatively late, typically after the first 6–9 months of life (5). Thus, tumor formation in Prkar1a+/− mice was not as efficient as in other mouse models of human syndromes leading to multiple neoplasias. One mitigating factor was the degree of increase in cAMP signaling that was observed in mouse tissues (5): generally, Prkar1a+/− mouse tissues did not demonstrate the same level of increased cAMP responsiveness of total PKA signaling seen in human CNC tissues. When a different mouse model was created that led to significantly higher Prkar1a down-regulation and consequently augmented cAMP signaling, more tissues were involved, life-span was decreased and the overall phenotype was more severe (6,7). Accordingly, tissue-specific complete Prkar1a deficiency that led to significant upregulation of PKA signaling in mouse pituitary and heart led to adenomas and myxomas, respectively (8,9).
The Prkar1a mouse model studies not only pointed to the significance of increased cAMP and/or PKA signaling for tumor formation but also suggested that additional factors may be necessary for Prkar1a haploinsufficiency to cause these tumors. Indeed, in vitro studies showed that dysregulation of cyclins and E2F1 were key changes in the process of immortalization of two cell lines, respectively: Prkar1a−/− mouse embryonic fibroblasts (MEFs) (10) and transformed human adrenal PRKAR1A-haploinsufficient cells (11).
Given the role of the p53 (the product of the Trp53 gene) and retinoblastoma (Rb, the product of the Rb1 gene) genes in controlling the cell cycle and the association of Rb with E2F1, we asked the following question: would the tumorigenic properties of Prkar1a haploinsufficiency be augmented or mimic more accurately the CNC phenotype in the background of Trp53 or Rb1 haploinsufficiency? Mouse models such as those for neurofibromatosis type 1 and Peutz-Jeghers syndrome, diseases that are similar to CNC, failed to reproduce exactly or with the same intensity the respective human phenotypes until the mouse gene was knocked out in the Trp53+/− background (12–14), even in the absence of any known involvement of p53 in the human tumors associated with these diseases. In addition, Trp53+/− mice developed bone tumors and bone is a tissue that is frequently affected by Prkar1a haploinsufficiency in mice (5). Rb1+/− mice developed intermediate lobe pituitary adenomas and thyroid C-cell hyperplasia or adenomas at high frequency (15) and mice deficient in both p53 and Rb develop primarily endocrine tumors (16).
And what about skin lesions in CNC and this disease's mouse models? Patients with CNC develop skin pigmented lesions, myxomas, collagenomas and fibromas (1,2); all these lesions, including the melanocytic ones, are always benign (17). As stated earlier, Prkar1a haploinsufficiency or down-regulation by an antisense transcript in mice did not lead to any skin tumors (5–7). There are several transgenic mouse models of skin tumors, but almost all are designed to address questions related to human cutaneous malignancies. A model that produces mostly benign, non-inflammatory, skin lesions (at least in the early phases) is the one that is the result of a two-stage protocol, in which tumor initiation on mouse skin is accomplished through a single topic application of 7,12-dimethylbenz(a)anthracene (DMBA) that causes an irreversible activation of the Hras oncogene in keratinocytes (18,19). Tumor promotion takes place when the initiated cells are expanded due to repeated applications of 12-O-tetradecanoylphorbol-13-acetate (TPA), a proliferation signal that promotes the formation of papillomas (20). Activation of the Hras signaling pathway has been postulated to be implicated in human PRKAR1A haploinsufficiency-induced proliferation (21,22).
Thus, we subjected the skin of Prkar1a+/− mice to DMBA and TPA treatment, an experiment that produced (like the other two described earlier) impressive results: in all three experiments, Prkar1a haploinsufficiency led to a greater number of tumors, and larger and more aggressive lesions. Interestingly, Prkar1a haploinsufficiency did not affect the pattern (the location or histology) of the tumors that these three very different mouse models developed; it only increased their number and/or size.
We then investigated what made these lesions different from those that developed without Prkar1a haploinsufficiency by studying their transcriptome. Beyond expected changes in cell cycle genes, Wnt signaling alterations were shared among all three experiments, despite their very different backgrounds. This finding, along with some recent data from human studies, e.g. the identification of secondary, somatic b-catenin (CTNNB1) mutations in PRKAR1A-haploinsufficient tumors (23,24) point to the central role of Wnt signaling abnormalities in mediating tumorigenesis initiated by Prkar1a haploinsufficiency.
Both the p53 and Rb1 proteins mediate arrest of the cell cycle at late G1. Rb1 functions through sequestration of members of the E2f transcription factor family and prevention of entry into the S phase (25), whereas p53 acts through activation of p21, which directly inhibits the G1 cyclin-dependent kinases that mediate cell cycle progression (26,27). Crosses were made between Trp53+/− mice and Prkar1a+/− mice in the mixed C57BL/6J genetic background and were monitored for subsequent tumor incidence and spectrum. Prkar1a+/− Trp53+/− mice exhibited significantly decreased survival relative to Trp53+/− and Prkar1a+/− mice (P < 0.0001) (Fig. 1A). By 1 year of age, Prkar1a+/− mice bred into the Trp53+/− background developed significantly more sarcomas than Trp53+/− mice (P < 0.001). Prkar1a+/− Trp53+/− mice (n = 30) developed 14 osteosarcomas and 11 splenic hemangiosarcomas, whereas Trp53+/− mice (n = 14) developed 2 splenic hemangiosarcomas, 1 cranial ostema and 1 fibrosarcoma (Fig. 2A). Thyroid neoplasms were also identified in Prkar1a+/− Trp53+/− mice (three follicular thyroid adenomas and one medullary thyroid carcinoma). Interestingly, a medullary thyroid carcinoma emerged from a Prkar1a+/− Trp53+/− mouse.
With the exception of few mild histologic abnormalities (data not shown), Prkar1a+/− mice did not develop any pituitary lesions, consistent with previously published data (5,9,28). Rb1+/− mice developed pituitary and thyroid tumors, as described previously (15). Rb1+/− Prkar1a+/− mice exhibited overall reduced survival when compared with Rb1+/− and Prkar1a+/− mice (P= 0.01) (Fig. 1B). Prkar1a+/− Rb1+/− mice also had a greater number of pituitary tumors and medullary thyroid carcinomas than Rb1+/− mice, and all lesions were larger in the former. Pars intermedia adenomas were identified in 50% of the Prkar1a+/− Rb1+/− mice, whereas only 17% in the Rb1+/− mice (Fig. 2B). Medullary thyroid carcinoma was diagnosed in all but one of the Prkar1a+/− Rb1+/− mice, significantly more than the Rb1+/− mice. In addition, medullary thyroid carcinomas were bilateral in most Prkar1a+/− Rb1+/− mice (data not shown) (Fig. 2C).
Finally, Prkar1a+/− and control mice [wild-type (WT)] from the same litter were treated with the DMBA/TPA carcinogenesis protocol (18–20). It should be noted that different mouse strains have variable susceptibility to this protocol and C57BL/6J mice (of the same background that Prkar1a+/− and WT mice were in this experiment) are relatively resistant to skin tumor induction (29). Prkar1a+/− mice developed more papillomas than WT mice during the 20-week-long treatment protocol (P = 0.004) (Fig. 1C). Prkar1a+/− mice (n = 7) treated with DMBA plus TPA developed a total of 25 papillomas after 20 weeks, whereas only 4 papillomas were observed in WT treated animals (n = 5) (Fig. 2D). Prkar1a+/− and WT mice treated with the vehicle (i.e. acetone) did not develop any skin tumors.
Tumors from the three mouse models were used for microarray analysis using the Illumina Beadarrays® system. All genes were displayed in the heat maps constructed by processing the data using unsupervised hierarchical clustering (Supplementary Material, Fig. S1) (30). The functional analysis of the whole-genome transcriptome profiling was performed using the DAVID Bioinformatic Resources 2008 (NIAID, NIH) (31,32). Multiple pathways are differentially regulated among tumors from single heterozygous and double heterozygous mice, as well as between papillomas from the Prkar1a+/− and WT mice (Table 1 and Supplementary Material, Table S1). The complete data are available at the National Center for Biotechnology Information's Gene Expression Omnibus (vog.hin.mln.ibcn@oeg) and are accessible through accession number GSE19576.
Cell cycle, apoptosis and mTOR-dependent pathways were enriched in Prkar1a+/− Trp53+/− sarcomas when compared with Trp53+/− sarcomas (P < 0.05). Pituitary and thyroid tumors from Prkar1a+/− Rb1+/− mice showed an increase in Wnt signaling, cell cycle and apoptosis pathways in relation to Rb1+/− tumors (P < 0.05). Genes related to Wnt signaling, cell cycle, apoptosis and oxidative phosphorylation pathways were over-expressed in Prkar1a+/− papillomas compared with WT papillomas (P < 0.05).
A qRT–PCR array carrying 84 genes of the cell cycle was carried out to confirm the microarray data (Fig. 3A). Key genes (E2f1, Cdk2, Cdk4, Cdknb1, Mdmd2, Chek1 and Mki67) were overexpressed in all Prkar1a+/− Trp53+/− sarcomas when compared with Trp53+/− sarcomas (P < 0.05). Five out of 84 genes (E2f1, Cdkn2a, Ccnb2, Ccna1 and Mki67) were significantly more expressed in pituitary tumors from the Prkar1a+/− Rb1+/− mice, compared with the Rb1+/− mice (P < 0.05). Expression levels of E2f1, Mdm2, Notch2 and Chek1 genes were also higher in Prkar1a+/− papillomas compared with papillomas from WT mice (P < 0.05). A strong staining for E2f1 was demonstrated in Prkar1a+/− papillomas and pituitary and thyroid tumors from Prkar1a+/− Rb1+/− mice (Fig. 3B). Cyclin D1 over-expression at the protein level was also detected in Prkar1a+/− papillomas, Prkar1a+/− Trp53+/− sarcomas, and pituitary and thyroid tumors from Prkar1a+/− Rb1+/− mice (Fig. 3C). Known downstream targets of PKA (Akt and c-fos) were over-expressed in tumors from both the double heterozygous animals and in Prkar1a+/− papillomas (Fig. 3D).
The Wnt signaling pathway was over-expressed in Prkar1a+/− sarcomas, pituitary and thyroid tumors, and papillomas, from the three experiments, respectively. The qRT–PCR array included 84 genes from this pathway (Fig. 4A). The Ctnnb1, Wnt3, Wnt3a and Csnk1d genes were over-expressed in Prkar1a+/−Trp53+/− sarcomas, compared with Trp53+/− sarcomas (P < 0.05). Seven out of 84 genes (Ctnnb1, Wnt3, Wnt3a, Lrp5, Jun, Frzb and Foxn1) were significantly more expressed in pituitary tumors from Prkar1a+/− Rb1+/− mice, compared with Rb1+/− mice (P < 0.05). Expression levels of Ctnnb1, Wnt3, Lrp5 genes were also higher in Prkar1a+/− Rb1+/− thyroid tumors (P < 0.05). In Prkar1a+/− papillomas, five genes (Ctnnb1, Lrp5, Myc, Foxn1 and Fzd6) of the Wnt signaling pathway were found to be over-expressed (P < 0.05). A strong Ctnnb1 protein expression was detected in Prkar1a+/− papillomas, Prkar1a+/− Trp53+/− sarcomas and Prkar1a+/− Rb1+/− pituitary tumors (Fig. 4B). Wnt3 immunoreactivity was clearly higher in Prkar1a+/− papillomas than in WT papillomas (Fig. 4C). The Lrp5 protein was over-expressed in several of the tumors from the two double heterozygote mouse models and in the Prkar1a+/− papillomas (Fig. 4D).
The Wnt3, Ctnnb1, E2f1 and Cdk4 genes that were significantly altered in all three experiments were selected as targets for confirmation by siRNA studies. A human adrenal cell line bearing a PRKAR1A-inactivating mutation (Carney cells) and Prkar1a+/− MEFs were used for these experiments; a fifth gene that was significantly altered in all three experiments, Lrp5, was not expressed in both cell lines and siRNA treatment for Lrp5 did not have any effect in cell proliferation (to be excluded) (Fig. 5A). On the other hand, siRNA inhibition of Wnt3, Ctnnb1, E2f1 or Cdk4 led to significant decrease in cell proliferation of Carney cells and MEFs (P < 0.05) (Fig. 5B). For example, E2f1 siRNA treatment reduced proliferation by 50 ± 2.3% and 59.9 ± 1.6% when compared with untreated Carney cells and MEFs, respectively.
Flow-cytometric analysis was performed to assess cell cycle distribution of the whole cell population (Fig. 5C). Cell cycle analysis of synchronized Carney cells and MEFs treated with siRNA for Wnt3, Ctnnb1, E2f1 and Cdk4 indicated that both cell lines significantly accumulated in the G0/G1 phase of the cell cycle compared with untreated or non-target siRNA treated cells (P < 0.05).
The presented data showed that Prkar1a+/− Trp53+/− and Prkar1a+/− Rb1+/− mice developed more sarcomas and endocrine tumors, respectively, and Prkar1a haploinsufficiency increased susceptibility to skin papillomas induced by DMBA/TPA; Prkar1a haploinsufficiency also reduced overall survival in the Trp53+/− and Rb1+/− backgrounds. This study is to our knowledge the first indication that Prkar1a haploinsufficiency in mice may act as a generic, synergistic, albeit weak, tumorigenic signal. Prkar1a haploinsufficiency appears to be a ‘generic’ tumorigenic signal in the sense that it increased lesions in all three mouse models that were tested, despite the fact that the Trp53+/−, Rb1+/− and DMBA/TPA-induced tumors are quite different in terms of pathways and tissues that are affected. Prkar1a haploinsufficiency appears to also have a synergistic and not an additive effect when combined with Trp53+/− and Rb1+/− defects: the signature lesions of Prkar1a+/− mice, tail and thyroid follicular tumors, and schwannomas (5) were conspicuously not increased in frequency or different in pathology in the Prkar1a+/− Trp53+/− and Prkar1a+/− Rb1+/− mice.
It is also noteworthy, and perhaps indicative of its weakness as a tumorigenic signal, that Prkar1a haploinsufficiency did not alter in any way the tissues or types of cells that are affected by Trp53+/− and Rb1+/− defects or by DMBA/TPA application, despite the almost ubiquitous Prkar1a expression. We selected these mouse models because of the cAMP-responsiveness of the main types of affected cells (sarcomas, thyroid C-cells and pituitary intermediate lobe) or due to the uniqueness of the phenotype and its similarity to CNC (skin papillomas). In all cases, Prkar1a haploinsufficiency simply augmented the previously described phenotype for the respective mouse model, without causing new tumors in other cAMP-responsive (or other) tissues.
R1α haploinsufficiency in human lymphocytes and mouse models causes an increase in total cAMP-stimulated kinase activity and enhances MAPK activity (5,22,33,34). Complete R1α deficiency causes constitutive PKA activation and immortalization of MEFs through up-regulation of cyclin D1; this increase occurred independently of other pathways known to increase cyclin D1 levels (10). The involvement of cyclin D1 in mediating Prkar1a haploinsufficiency signals was confirmed in our studies, along with that of other genes of the cell cycle, that were significantly altered not only in the Rb1+/− and Trp53+/− backgrounds (where comparison was made with the single heterozygotes, Rb1+/− and Trp53+/−, respectively) but also in the DMBA/TPA-induced tumors where no primary cell cycle gene was altered genetically.
Our previous studies of a human cell line harboring a PRKAR1A-inactivating mutation showed that the cell cycle factor E2F1 mediates proliferative effects of defective R1α (11). E2F1 is a downstream effector of Rb and has a pivotal role in controlling cell cycle progression (35,36). Loss of E2f1 in mice reduces the frequency of pituitary and thyroid tumors and lengthens the lifespan of Rb1+/− E2f1−/− mice (37). In the present study, siRNA disruption of E2f1 inhibited proliferation of Carney cells and Prkar1a+/− MEFs and arrested both cell lines at the G0/G1 phase of the cell cycle, preventing entry into and progression through S phase.
The implication of E2F1 in mediating R1α haploinsufficiency signals provides a link to the observed synergy of Prkar1a-haploinsufficiency in the Trp53+/− background and the observed activation of cyclin D1: recent studies suggested that transcriptional activation and repression of E2fs are linked to the p53 tumor suppressor gene (38–40). Specifically, targeted disruption of activator E2fs (E2F1, E2F2 and E2F3) led to p53 activation and sequential induction of p53 target genes, including CDKN1A (38).
The Wnt signaling pathway regulates a wide range of functions such as cell growth and differentiation (41). The functional analysis of array data revealed a significant enrichment of Wnt signaling pathway in Prkar1a+/− Trp53+/− sarcomas, Prkar1a+/− Rb1+/− pituitary and thyroid tumors, as well as in Prkar1a+/− papillomas. Ctnnb1, Wnt3 and Lrp5 over-expression was confirmed by qRT–PCR and immunostaining. In addition, the treatment of PRKAR1A-haploinsufficient cells (Carney cells) and Prkar1a+/− MEFs with siRNAs for Wnt3 and Ctnnb1 blocked the transition to the S phase in both cell lines. cAMP signaling via PKA and its target transcription factor CREB are required for Wnt-directed myogenic gene expression (42). It was previously demonstrated that glycogen synthase kinase (GSK)-3β is a downstream target of cAMP/PKA signaling in steroidogenic cells (43). Interestingly, expression studies in massive macronodular adrenocortical disease, a disorder associated with excess cAMP signaling (44) and primary pigmented nodular adrenocortical disease (PPNAD), a condition mostly associated with PRKAR1A-inactivating mutations (4), indicated the over-expression of genes that regulate or are part of the Wnt signaling pathway such as WISP2, glycogen synthase kinase-3β (GSK3B) and β-catenin (CTNNB1) (45,46). In addition, somatic mutations of the CTNNB1 gene have been found in adrenal tumors from patients with PPNAD, Carney complex and germline PRKAR1A-inactivating mutations (23,24). Recently, a microRNA profile analysis showed that cAMP and/or PKA via microRNA regulation affects the Wnt signaling pathway in both PPNAD (47) and MMAD (48). These studies along with the findings presented here suggest that the Wnt pathway is a primary mediator of tumorigenesis related to increased cAMP and/or PKA signaling and R1α haploinsufficiency, almost regardless of the cellular context in both human and mouse tissues.
We conclude that Prkar1a haploinsufficiency can act weakly but synergistically with other molecular defects in inducing tumor formation (Fig. 6). Dysregulation of cell cycle genes such as cyclin D1 and E2F1 and abnormal Wnt signaling are essential in these effects.
To exclude possible effects of the different genetic backgrounds on molecular signature of tumors, all mice were from a (CD-1 × C57BL/6) F1 hybrid background. Prkar1a+/− mice carrying a deletion of exon 2 have been previously described (5). Trp53+/− mice (C57BL/6; Trp53tm1Tyj) were purchased from Jackson Laboratories (JAX mice and Services, Bar Harbor, ME, USA) and were genotyped as described elsewhere (49). Rb1+/− mice (C57BL/6; Rb1tm1Tyj) were purchased from Jackson Laboratories and were genotyped as previously described (50). Prkar1a+/− mice were crossed with Trp53+/− or Rb+/− mice and the phenotype analyzed at 1 year of age. In the two-step carcinogenesis protocol, WT and Prkar1a+/− mice (both on C57BL/6 background) were treated with a single topical application of 7,12-dimethylbenz(a)anthracene (DMBA) and repeated applications of 12-O-tetradecanoylphorbol-13-acetate (TPA) once per week for 20 weeks (51). All animal work in this study was carried out in accordance with Institutional Laboratory Animal Care and Use Committee guidelines under animals protocol 06-033 (at the NIH, Bethesda, MD, USA).
Total RNA was extracted from tumor samples using the TRIZOL reagent method. Preparation of cRNA from total RNA, hybridization in Sentrix MouseRef-8 Expression BeadChips, scanning and image analysis was done as previously described (52). Preliminary analysis of Illumina data was performed using Illumina BeadStudio software (Illumina, San Diego, CA, USA), which returns the trimmed mean average intensity for each single gene probe type (non-normalized). Any gene consistently with a P detection value above 0.05 for all samples was eliminated from further analysis. This background filter resulted in the removal of ~40% of all the genes on the Illumina array. Z-transformation for normalization was performed for each Illumina sample/array (53). Microarray data are in compliance with the Minimal Information About a Microarray Experiment (MIAME) format. The raw and normalized array data have been deposited in National Center for Biotechnology Information's Gene Expression Omnibus http://www.ncbi.nlm.nih.gov/geo and are accessible through Gene Expression Omnibus Series accession number GSE19576. Heatmaps were made using Java Treeview (54).
The functional analysis of the whole-genome transcriptome profiling was performed using the Database for Annotation, Visualization and Integrated Discovery (DAVID) Bioinformatic Resources 2008 (NIAID, NIH, http://david.abcc.ncifcrf.gov/home.jsp) (31,32). The lists of genes (induced or repressed) were submitted to the DAVID database (http://david. abcc.ncifcrf.gov), which clusterizes genes according to a series of common keywords. The proportion of each keyword in the list is compared with the one in the whole genome, making it possible to compute P-values and enrichment scores (geometric mean of the inverse log of each P-value). The detailed information of gene alterations was systematically reported on KEGG pathways (Supplementary Material, Table S1).
Gene expression data were confirmed by quantitative real-time PCR using specific array plates for cell cycle and Wnt signaling pathways (SABiosciences, Frederick, MD, USA). Relative quantification was performed using the 2−ΔΔCT method (55).
Tumor tissues were removed from mice and fixed in formalin, processed and paraffin embedded for subsequent H&E staining and immunohistochemistry, as previously described (56). For the immunohistochemical analysis, the following antibodies were used: E2f1 (H-137, Santa Cruz Biotechnology, Santa Cruz, CA, USA), Cncd1 (ab16663, Abcam, Cambridge, MA, USA), p16 (ab54210, Abcam), AKT (H-136, Santa Cruz Biotechnology), c-fos (ab7963-1, Abcam), Wnt3 (ab32249, Abcam), Ctnnb1 (ab6302, Abcam) and Lrp5 (ab38311, Abcam).
The human adrenal cells bearing a PRKAR1A-inactivating mutation (Carney cells) and Prkar1a+/− MEFs were previously characterized (10,11). Carney cells and Prkar1a+/− MEFs were transfected with 100 nM ON-TARGETplus Smartpool siRNA (Dharmacon Thermo Scientific, Lafayette, CO, USA) specific to human (WNT3, LRP5, CTNNB1, E2F1 and CDK4) and mouse (WNT3, Wnt3, Ctnnb1, Lrp5, E2f1 and Cdk4), respectively, or ON-TARGETplus non-targeting pool using transfection reagent DharmaFECT1 (Dharmacon) as per manufacturer's instruction.
The efficiency of siRNA target gene disruption was confirmed by western blot analysis following standard procedures (57). The following antibodies were used: Wnt3 (ab32249, Abcam), Ctnnb1 (ab6302, Abcam), Lrp5 (ab38311, Abcam), E2f1 (H-137, Santa Cruz Biotechnology), Cdk4 (DCS-35, Santa Cruz Biotechnology) and beta-actin (ab8227, Abcam). Briefly, cells were lysed by homogenization in 20 mm Tris–HCl (pH 7.5), 100 mm NaCl, 5 mm MgCl2, 1% Nonidet P-40, 0.5% sodium deoxycholate and protease inhibitor cocktail I (EMD Biosciences, La Jolla, CA, USA) with subsequent centrifugation at 10 000 rpm for 10 min at 4°C. Equal amounts of protein lysate were subjected to SDS–PAGE, transferred to nitrocellulose membranes and probed with antibodies.
After 72 h of siRNA transfection, cell proliferation assays were performed on Carney cells and Prkar1a+/− MEFs using the CellTiter 96® AQueous One Solution Cell Proliferation Assay (Promega, Madison, WI, USA). Data represent mean ± SE of three separate experiments.
Flow-cytometry analysis was performed to assess cell cycle distribution of the whole cell population. Carney cells and Prkar1a+/− MEFs, synchronized in low-serum (0.5%) medium for 48 h, were released from cell cycle arrest by adding 10% FBS. We measured the cell cycle 48 h after addition of serum. Cells were harvested, fixed with ice-cold 70% ethanol, stained with propidium iodide (10 µg/ml) and ribonuclease A (100 µg/ml) (Sigma, St Louis, MO, USA), and subjected to cell cycle analysis using FACSCalibur (Becton Dickinson, Mountain View, CA, USA). The percentage of aneuploid cells and cell cycle distribution were calculated with ModFit LT cell-cycle analysis software (Verity Software House, Topsham, ME, USA). Data represent mean ± SE of three separate experiments.
All statistical analyses were performed with the SPSS 16.0 (SPSS Inc., Chicago, IL, USA). Survival analysis was performed using the log rank test. The X2 test was used to test the number of observed tumors between groups. Continuous data are expressed as mean ± SE. A two-sample t-test was used for statistical analysis of cell proliferation and cell cycle data. A P-value less than 0.05 was considered significant.
Conflict of Interest statement. None declared.
This work was supported by US National Institutes of Health, Eunice Kennedy Shriver National Institute of Child Health and Human Development intramural project Z01-HD-000642-04 (to C.A.S.).