The Fanconi anemia pathway controls oncogenic response in hematopoietic stem and progenitor cells by regulating PRMT5-mediated p53 arginine methylation

The Fanconi anemia (FA) pathway is involved in DNA damage and other cellular stress responses. We have investigated the role of the FA pathway in oncogenic stress response by employing an in vivo stress-response model expressing the Gadd45β-luciferase transgene. Using two inducible models of oncogenic activation (LSL-K-rasG12D and MycER), we show that hematopoietic stem and progenitor cells (HSPCs) from mice deficient for the FA core complex components Fanca or Fancc exhibit aberrant short-lived response to oncogenic insults. Mechanistic studies reveal that FA deficiency in HSPCs impairs oncogenic stress-induced G1 cell-cycle checkpoint, resulting from a compromised K-rasG12D-induced arginine methylation of p53 mediated by the protein arginine methyltransferase 5 (PRMT5). Furthermore, forced expression of PRMT5 in HSPCs from LSL-K-rasG12D/CreER-Fanca−/− mice prolongs oncogenic response and delays leukemia development in recipient mice. Our study defines an arginine methylation-dependent FA-p53 interplay that controls oncogenic stress response.

In response to oncogenic activation, normal cells induce genetically encoded programs, mainly growth arrest, apoptosis and senescence, which prevent deregulated proliferation and thus protect multicellular organisms from cancer progression [13][14][15]. It is known that oncogene-driven proliferation must be associated with inhibition of apoptosis and senescence to allow malignant outgrowth [16][17][18]. It has been postulated that functional loss of FA proteins renders cells a high predisposition to cancer transformation [1-3, 8, 19, 20]. However, how cells deficient in the FA pathway respond to oncogenic stress remains largely unknown.
The tumor suppressor p53 is a key component of the DNA damage response (DDR) network that activates vital damage containment procedures to restrict aberrant cell growth in response to DNA damage, oncogene activation, and loss of normal cell contacts by maintaining the balance between cell survival and apoptosis [21][22][23][24]. It is known that oncogene insults activate p53 [24,25]. Highly regenerative tissues, such as blood, possess effective DNA damage responses (DDR) that balance long-term Research Paper regeneration with protection from leukemogenesis [26]. Several recent studies using mouse models suggest a critical role for the p53 pathway in hematopoietic stem cell (HSC) self-renewal and quiescence [27][28][29][30][31] and precise regulation of p53 activity is likely to be important in determining the response of HSCs to DNA damage and oncogenic activation. Insufficient p53 activation would favor cell survival, but put cells at risk for loss of genomic integrity. In contrast, excessive p53 activation could compromise steady-state hematopoiesis and its recovery following exogenous marrow insult by causing too many cells to be eliminated [32]. In the context of FA, emerging evidence suggests that p53 deficiency may increase cancer development in patients with FA and FA mice [33][34][35][36]. Conversely, recent studies show that overactive p53 could cause hematopoietic stem and progenitor cell (HSPC) depletion in the BM of FA patients [37]. These studies corroborate a critical role of the FA proteins in cooperating with p53 in apoptosis and cell cycle checkpoint control after DNA damage induced by an exaggerated physiologic stress response.
In the present study, we employed an in vivo stressresponse model expressing the luciferase transgene under the control of the promoter of the stress-responsive gene Gadd45β and showed that HSPCs from mice deficient for the core complex components of the FA pathway, Fanca or Fancc, exhibited aberrant response to oncogenic stress. In a sharp contrast to wild-type controls, Fanca −/− or Fancc −/− HSPCs showed a short-lived response to oncogenic activation. Significantly, we demonstrated that disruption of the FA pathway compromised the oncogene K-ras G12D -induced arginine methylation of p53 mediated by the protein arginine methyltransferase 5 (PRMT5). Therefore, our study demonstrates for the first time that oncogenic stress orchestrates a p53-dependent response that is controlled by PRMT5-mediated arginine methylation and identifies the FA pathway as an integral part of this versatile cellular mechanism.

Disruption of the FA pathway induces a short-lived response to oncogenic stress in vitro
To investigate how cells with a defective FA pathway respond to oncogenic stress, we employed two inducible models of oncogenic activation: 1) the LSL-K-ras G12D /CreER mouse model [38], in which Cremediated recombination leads to deletion of a translational termination sequence (Lox-Stop-Lox) and expression of the oncogenic Kras G12D protein; 2) A retroviral vector delivered proto-oncogene Myc ER system [39], in which oncogene activation is induced and controlled by 4-OHT or tamoxifen exposure. The reason that we chose the K-ras G12D model for these detailed studies was two-fold: 1) it is an authentic in-vivo knock-in model, which enabled us to analyze oncogenic response under near physiological conditions; and 2) it is an established myeloid leukemia model, which has relevance to FA disease progression. We first examined the sensitivity of hematopoietic stem and progenitor (HSPC; LSK) cells ( Figure S1A), isolated from LSL-K-ras G12D /CreER mice or infected with the Myc ER retrovirus, to oncogene activation by in vitro culturing the cells in the presence of 4-Hydroxytamoxifen (4-OHT) for 48 hours [40,41] followed by plating in cytokinesupplemented methycellulose medium. We found that K-ras or Myc activation reduced colony formation in both WT and Fanca −/− or Fancc −/− progenitors (Figures 1A, S1B), which accompanied by increased apoptosis 24-96 h after 4-OHT induction ( Figures 1B, 1C, S1C, S1D).

Disruption of the FA pathway induces a short-lived response to oncogenic stress in vivo
We next examined oncogenic response in vivo by crossing the FA mice to the Luc-Gadd45β mice, which express the luciferase transgene under the control of the promoter of the stress-responsive gene Gadd45β [44] and allow for non-invasive in vivo imaging stressinduced expression of the luciferase marker. Gadd45β is well established for its diverse roles in cell cycle control, cell survival, apoptosis, DNA damage repair and the maintenance of genomic stability [45]. Gadd45β can also act as a stress sensor in the development of hematopoietic malignancies like leukemia [46]. We assessed oncogenic response in the recipients with similar donor chimerism ( Figure S2), transplanted with BM LSK cells expressing Luc-Gadd45β and LSL-K-ras G12D /CreER or Myc ER , by monitoring the kinetics of luciferase expression at different time points after tamoxifen injection. Consistent with the in vitro observations described above, the recipients transplanted with WT (Fanca +/+ or Fancc +/+ ) LSK cells expressing LSL-K-ras G12D /CreER (Figure 2A, 2B, 2E) or Myc ER ( Figure 2C, 2D, 2F) exhibited a persistent induction of luciferase expression even 48 hours after tamoxifen injection. In contrast, the recipients transplanted with FA-deficient (Fanca −/− or Fancc −/− ) LSK cells expressing LSL-K-ras G12D /CreER (Figure 2A, 2B, 2E) or Myc ER ( Figure 2C, 2D, 2F) showed a typical short-lived response www.impactjournals.com/oncotarget with a peak at 4 hours and declined from 6-24 hours after induction. Taken together, HSPCs deficient for the Fanca or Fancc gene displayed abnormal short-lived response to oncogenic stress. In addition, we observed a similar pattern of less long-lived oncogenic response in primary mouse embryonic fibroblasts (MEFs) isolated from Luc-LSL-K-ras/CreER-Fanca −/− mice ( Figure S3), suggesting that this abnormal oncogenic response is not restricted to HSPCs.

Deregulated G 1 checkpoint regulators correlated with short-lived oncogenic response
The observation that Fanca −/− and Fancc −/− HSPCs exhibited an abnormal short-lived response to K-ras and c-Myc activation prompted us to investigate the underlying molecular mechanism. Since activation of oncogenes including K-ras and c-Myc induces senescencelike cell growth arrest in a variety of cell types [47,48] and since we observed a short-lived G 1 arrest induced by K-ras or c-Myc activation in Fanca −/− and Fancc −/− LSK cells ( Figures 1D, 1E, S1E, S1F), we analyzed oncogeneinduced expression of G 1 arrest-associated cell-cycle regulators in K-Ras G12D -expressing HSPCs with a focus on p15 INKb , p16 INK4a , ARF and p21 Cip1/WAF1 [49,50]. Indeed, it is known that the K-ras G12D mutation is frequently associated with juvenile myelomonocytic leukemia (JMML) and acute myeloid leukemia (AML) [51]. LSK cells freshly isolated from WT and Fanca −/− littermates expressing the LSL-K-ras G12D /CreER transgenes were treated with 4-OHT at different time points to determine the kinetics of oncogenic induction of the G 1 arrest-associated cellcycle regulators. Interestingly, three of the four G 1 arrest-associated cell-cycle regulators, namely p16 INK4a , www.impactjournals.com/oncotarget p19 Arf and p21 Cip1/WAF1 , exhibited the pattern of short-lived response to K-ras activation in Fanca −/− HSPCs. That is, K-ras activation orchestrated the expression of these cell-cycle regulators that burst at 4 hours then receded to the basal level at 24 hours ( Figure 3A). In contrast, the WT LSK cells displayed a persistent response during the 4-24 hours period ( Figure 3A). These results suggest that a pathway regulating the expression of p16 INK4a , p19 Arf and p21 Cip1/WAF1 may be responsible for the short-lived oncogenic stress response in Fanca −/− HSPCs.

Short-lived oncogenic response involves p53 signaling
Since p21 Cip1/WAF1 is a p53 target and p19 Arf induction by K-ras requires p53 [49,50,52], we determined if p53 was essential for the prolonged oncogenic response in HSPCs. We deleted the Trp53 gene in our Fanca −/− LSL-K-ras G12D /CreER mice, and evaluated the requirement of p53 for K-ras-induced OIS (oncogene-induced senescence) [53] using p16 INK4a expression as a surrogate. Results are means ± standard deviation (SD) of 3 independent experiments (n = 9 per group). (D) Activation of K-ras leads to short-lived G 1 arrest in FA cells. Cells described in (B) were cultured in the presence of 4-OHT for 2 hours then released in fresh medium for the indicated time intervals, followed by cell cycle profiling by Hochest33324/Ki67 staining. Representative images (left) and quantification (right) were shown. Results are means ± standard deviation (SD) of 3 independent experiments (n = 6 per group). (E) Activation of Myc leads to short-lived G 1 arrest in FA cells. Cells described in (C) were cultured in the presence of 4-OHT for 2 hours then released in fresh medium for the indicated time intervals, followed by cell cycle profiling by Hochest33324/Ki67 staining. Representative images (left) and quantification (right) were shown. Results are means ± standard deviation (SD) of 3 independent experiments (n = 9 per group).

Aberrant oncogene-induced arginine methylation of p53 contributes to short-lived oncogenic response
We next examined the status of p53 in WT and Fanca −/− LSL-K-ras G12D /CreER HSPCs. While Fanca −/− BM lineage-negative (Lin -) cells (enriched for HSPCs) showed higher levels of p53 protein than WT cells, we did not observe significant difference in the kinetics of p53 induction between WT and Fanca −/− cells during the 2-24 hours period of K-ras activation ( Figure 3C). This suggests the aberrant oncogenic response observed in Fanca −/− HSPCs was most likely resulted from an altered p53 activity. Since most if not all of the p53 activities are controlled by post-translational modifications [54], we determined phosphorylation, acetylation and methylation of p53 in WT and Fanca −/− LSL-K-ras G12D /CreER HSPCs. The p53 protein was enriched by immunoprecipitation using lysates of BM Lincells treated with 4-OHT for different time intervals, and probed for forms of p53 modified by phosphorylation, acetylation or methylation under oncogenic stress. We found similar levels of phosphorylation, acetylation, or lysine methylation of p53 between WT and Fanca −/− cells at different time points after oncogene induction ( Figure 3D). However, K-rasinduced arginine methylation of p53 was peaked at 2 hours and almost completely abated after 8-hour induction with 4-OHT in Fanca −/− Lincells ( Figure 3D), reminiscent of the short-lived oncogenic response described above. Thus, these results identified an aberrant oncogene-induced arginine methylation of p53 as a potential mechanism for the short-lived oncogenic stress response in FA HSPCs.

Ectopic expression of PRMT5 prevents short-lived oncogenic response
Protein arginine methylation is mediated by a group of protein arginine methyltransferases (PRMTs) [55]. Three of the PRMTs (PRMT1, 4 and 5) have been implicated in regulating p53-mediated transcription [50][51][52]. We thus examined the levels of these PRMTs in Fanca −/− LSL-K-ras G12D /CreER Lincells transduced with Venus or Venus-FANCA lentivirus and then subjected to oncogenic stress. We found that the level of PRMT5 underwent progressive decline after 4 hours of 4-OHT induction in Fanca −/− LSL-K-ras G12D /CreER Lincells, and re-expression of the FANCA gene in these cells stabilized (A) Short-lived response of G 1 -arrest-associated cell-cycle regulators in response to oncogenic stress in FA HSPCs. LSK cells isolated from Luc-LSL-Fanca +/+ /K-ras/CreER or Luc-LSL-Fanca −/− /K-ras/CreER mice were cultured in the presence of 4-OHT for 2 hours then released in fresh medium for the indicated time intervals. RNA was extracted at different time points for qPCR analysis using primers listed in Table S1. Levels of the expression in each sample were normalized to the level of GAPDH mRNA, and the expression levels of the Fanca +/+ samples at 2 h were normalized as 100. (B) p53 is essential for the prolonged oncogenic response in HSPCs. LSK cells isolated from LSL-Fanca +/+ /K-ras/CreER-p53 +/+ , LSL-Fanca −/− /K-ras/CreER-p53 +/+ , LSL-Fanca +/+ /K-ras/CreER-p53 −/− , LSL-Fanca −/− /K-ras/ CreER-p53 −/− mice were cultured in the presence of 4-OHT for 2 hours then released in fresh medium for the indicated time intervals. RNA was extracted at different time points for qPCR analysis using primers for p16 (Upper) and p21 (Lower) Samples were normalized to the level of GAPDH mRNA. (C) K-ras activation does not alter p53 level. BM Lincells from Luc-LSL-Fanca +/+ /K-ras/CreER or Luc-LSL-Fanca −/− /K-ras/CreER mice were cultured in the presence of 4-OHT for 2 hours then released in fresh medium for the indicated time intervals. Whole cell lysates were prepared then subjected to immunoblot using antibodies against p53 or β-actin. p53:β-actin ratio in each sample was calculated. The WT control sample at 0 h time point has been normalized as 1. (D) FA deficiency leads to short-lived arginine methylation of p53 in response to oncogene activation. BM Lincells from Luc-LSL-Fanca +/+ /K-ras/CreER or Luc-LSL-Fanca −/− /K-ras/ CreER mice were cultured in the presence of 4-OHT for 2 hours then released in fresh medium for the indicated time intervals. Whole cell lysates were prepared, and subjected to immunoprecipitation using anti-p53 antibody followed by immunoblotting with antibodies against phosphor-serine, acetylated lysine, methylated lysine, mono-methyl arginine or p53, respectively. the PRMT5 protein during the 24-hour period of 4-OHT induction ( Figure 4A).
The correlation between aberrant oncogene-induced arginine methylation of p53 accompanied by a progressive decline in PRMT5 and the short-lived oncogenic stress response in FA HSPCs prompted us to determine whether restoring the level of PRMT5 could lengthen the oncogenic response in FA HSPCs. To this end, we ectopically expressed GFP or GFP-PRMT5 in WT and Fanca −/− Lincells expressing the Luc-LSL-K-ras G12D /CreER cassette by lentiviral transduction (Figure S4A). The transduced (GFP + ) cells were then sorted for LSK cells and used for analyzing oncogenic response in vitro and in vivo. First, we analyzed the sorted GFP + LSK cells for the kinetics of p16 INK4a expression at different time-points of K-ras G12D induction following 4-OHT treatment. We found that ectopic expression of PRMT5 completely abrogated the short-lived K-ras-induced p16 INK4a expression in Fanca −/− LSK cells ( Figure 4B). Lentiviral expression of PRMT5 did not have effect on K-ras-induced p16 INK4a expression in WT LSK cells ( Figure S4B).
To specifically demonstrate the effect of ectopic expression of PRMT5 on p53 response, we determined the kinetics of oncogene-induced expression of the p53 target p21 Cip1/WAF1 using vector-or PRMT5-transduced LSK cells from Fanca −/− LSL-K-ras G12D /CreER mice with WT or p53-null background. As illustrated in Figure 4C, forced expression of PRMT5 showed a prolonged pattern, similar to that seen in WT LSK cells ( Figure 3B), of p21 Cip1/WAF1 expression in response to K-ras activation in p53-sufficient Fanca −/− LSL-K-ras G12D /CreER LSK cells but not in p53defficient Fanca −/− LSL-K-ras G12D /CreER LSK cells ( Figure 4C). This result suggests that p53 is at least one of the targets of ectopically expressed PRMT5 in response to K-ras G12D in these LSK cells.

Forced expression of PRMT5 delays leukemia
To determine the effect of ectopic expression of PRMT5 on oncogenic response in vivo, we transplanted the transduced (GFP + ) WT and Fanca −/− Luc-LSL-Kras G12D /CreER LSK cells into lethally irradiated BoyJ recipients. Four months later, the recipients were treated with tamoxifen and assessed oncogenic response by monitoring the kinetics of luciferase expression at different time points. It appeared that exogenous expression of PRMT5 not only lengthened but also enhanced the activated K-ras response in the progeny of donor Fanca −/− Luc-LSL-K-ras G12D /CreER HSPCs (Figures 4D, S4C). Lentiviral expression of PRMT5 also augmented K-rasinduced luciferase expression in WT donor-derived cells ( Figure S4D). Furthermore, K-ras G12D -induced leukemia was significantly delayed in recipient mice transplanted with Fanca −/− Luc-LSL-K-ras G12D /CreER LSK cells expressing PRMT5 compared with the recipient mice of donor Fanca −/− Luc-LSL-K-ras G12D /CreER LSK cells carrying the vector control virus ( Figure 4E). Further characterization of the leukemic mice showed that the recipient mice transplanted with the Fanca −/− Luc-LSL-K-ras G12D /CreER LSK cells developed myeloid leukemia, as characterized by increased white blood cells (WBCs) in the peripheral blood, anemia, and infiltration of myeloid blasts in the spleen and bone marrow ( Figure 4F-4H). Collectively, these results link attenuated PRMT5 activity to the short-lived oncogenic stress response and leukemia development in FA HSPCs.
We further examined the molecular and cellular effects of forced expression of PRMT5 by analyzing the levels of p53 arginine methylation, the transcriptional activation of p53, the activity of DNA damage response kinases ATM and ATR, and cell cycle status in the leukemic cells. As expected, ectopic PRMT5 expression increased arginine methylation of p53 in K-ras G12Dexpressing WT and Fanca −/− leukemic cells ( Figure 5A). In contrast to an increased expression of the p53 target gene in cell-cycle control, p21 Cip1/WAF1 ( Figure 4C), forced expression of PRMT5 did not enhance expression of several major p53 apoptotic target genes like Bax, Puma, or Noxa in these leukemic cells ( Figure 5B). PRMT5 expression did not increase the activity of the key DNA damage response kinases ATM and ATR, using phosphorylation of CHK1 and CHK2 as the respective surrogate ( Figure 5C). However, forced expression of PRMT5 prolonged the short-lived G 1 arrest induced by K-ras activation in Fanca −/− leukemic cells ( Figure 5D). These results suggest that PRMT5 may exert its function via the p53/G 1 cell-cycle checkpoint rather than the ATM/ ATR/G 2 DNA damage checkpoint in our oncogenic response model.

DISCUSSION
The fact that patients with Fanconi anemia (FA) gene mutations have high predisposition to leukemia and other cancers indicates that FA disease progression involves oncogenic transformation. Little is known about the mechanism by which these mutant cells respond to oncogenic activation. In the present study, we investigated the response of FA HSPCs to oncogene insults and found that the FA pathway plays an important role in response to oncogenic stress. There are several findings that highlight the significance of our study: 1) FA HSPCs displayed an aberrant short-lived response to oncogenic stress induced by activated K-ras or c-Myc; 2) Fanca deficiency compromises K-ras G12D -induced arginine methylation of p53 accompanied by downregulated PRMT5; 3) forced expression of PRMT5 in Fanca −/− HSPCs prolonged oncogenic response and delayed leukemia development in irradiated recipient mice. Based on these observations, we propose a model in which the FA proteins in WT cells may play a role in regulating PRMT5-mediated p53 arginine methylation in response to oncogenic stress leading to prolonged oncogenic response. On the other hand, the regulatory function of PRMT5 on p53 is diminished in Fanca-or Fancc-deficient cells, which leads to decreased p53 arginine methylation and short-live oncogenic response ( Figure 5E).
An important and novel finding of the present study is the short-lived response of FA HSPCs to oncogenic insults. It is known that FA patients have an increased susceptibility to cancer including leukemia [1-3, 8, 19, 20]. This may be resulted from increased genomic instability, as cells from FA patients are defective in repair of DNA damage induced by certain genotoxic agents [4,5,8,20]. Genes encoding proteins with cancerpromoting (oncogenes like K-ras and Myc) or suppressive (tumor suppressors like p53 and ATM) activities in FA cells may be vulnerable to alterations due to a deficit in DNA repair. These alterations may create mutations that activate oncogenes or inactivate tumor suppressors leading to cancer transformation. How FA HSPCs respond to oncogene activation has not been studied. Here we employed two oncogenic systems, the LSL-K-ras G12D / CreER transgenic mouse model and the inducible Myc ER model, and observed a unique phenomenon featuring a short-lived response of FA HSPCs to oncogene activation both in vitro and in vivo. Whether this aberrant short-lived response is the consequence of mutations that disrupt oncogenic signaling pathways remains to be investigated.
We have made effort to identify the signaling pathway that is responsible for the observed short-lived oncogenic response in FA HSPCs. Our results implicate a role of the tumor suppressor p53. Several lines of evidence support this notion: first, the kinetics of oncogene-induced expression of p53 downstream signaling targets p19 Arf and p21 Cip1/WAF1 displayed the exact pattern of the expression of the Gadd45β-luciferase transgene and p16 INK4a in Fanca −/− HSPCs (Figures 2A, 2B, 2C, 2D and Figure 3A). Second, deletion of the Trp53 gene abrogated both the long-lasting and the short-lived oncogenic responses in WT and Fanca −/− LSL-K-ras G12D /CreER LSK cells, respectively ( Figure 3B, Upper). Third, p53 loss completely abolished p21 Cip1/WAF1 expression, which is absolutely required for the survival of HSPCs, in response to K-ras activation in both WT and Fanca −/− LSL-K-ras G12D /CreER LSK cells ( Figure 3B, Lower). Reduced level of p21 may lead to increased apoptosis upon oncogene induction and p53/p21 deficiency could simultaneously cause DNA damage and de-regulate cell cycle leading to mitotic catastrophe. These results thus identify the p53 tumor suppressor pathway that interplays with the FA pathway in oncogenic response. In addition, our finding further corroborates the previous reports that loss of p53 increases cancer development in patients with FA and FA knockout mice [34][35][36].
Our results also suggest that an alteration of p53 protein level or activity during time window of 4-24 hours following K-ras activation might account for the short-lived oncogenic response in Fanca −/− LSL-K-ras G12D / CreER LSK cells. The observation that there was no difference in p53 protein level between WT and Fanca −/− LSK cells during the 2-24 hours period of K-ras activation ( Figure 3C), shifted our attention to an altered p53 activity. In normal cells, cellular stresses including oncogenic activation initiates a signaling network involving enzymes that target p53 in the form of post-translational modifications, such as phosphorylation, acetylation and methylation [54]. Protein arginine methylation is mediated by a group of protein arginine methyltransferases (PRMTs) an important process and involved in the regulation of gene expression, RNA metabolism and protein function [55]. Remarkably, we showed that Fanca deficiency compromised K-ras G12D -induced arginine methylation of p53 accompanied by downregulated PRMT5 and that forced expression of PRMT5 in Fanca −/− HSPCs prolonged oncogenic response and delayed leukemia development in irradiated recipient mice (Figure 4). Thus, these studies defined a compromised oncogene-induced arginine methylation of p53 by PRMT5 as one mechanism for the short-lived oncogenic stress response in FA HSPCs.
The apparent tumor-suppressor role of PRMT5 identified in this work is in contrast to a recent study of Li et al. who demonstrated that PRMT5 for the inhibition of p53-dependent tumor suppression in response to oncogenic insults [57]. We noted three major differences in the experimental systems between our studies and those of Li et al. Firstly, while the study by Li et al. shows that PRMT5 is required for lymphomagenesis driven by multiple oncogenes including cyclin D1, c-MYC, NOTCH1, and MLL-AF9, the authors employed the Cyclin D1 D1T286A oncogenic model for PRMT5related mechanistic studies. On the other hand, we used the K-ras G12D model. Secondary, all of the studies by Li et al. were performed in T-cell lymphoma/Leukemia, which is different from our myeloid leukemia model. Lastly, our mouse model is deficient in the FA DNA repair pathway. In addition to these differences, our results appear to be consistent to a recent report, which shows that over-expression of PRMT5 caused cell cycle arrest while PRMT5 depletion resulted in apoptosis [58]. In this context, our study reveals the multiple facets of PRMT5 response to oncogenic activation. The mechanism by which PRMT5 is deregulated in the FA-deficient cells is not known at the moment. However, our recent proteomic study with a Fancd2 knock-in mouse model shows that Fancd2 interacts with the murine Prmt5 in vivo (data not shown). We speculate that the FA pathway may play a role in maintaining PRMT5 activity in response to oncogenic stress leading to increased p53 arginine methylation and prolonged oncogenic response.
In conclusion, the present study, using an innovative in vivo model integrating the stress-responsive Gadd45βluciferase transgene and inducible oncogenes (LSL-K- ras G12D and Myc ER ), uncovers two novel and significant findings: (i) normal HSPCs utilize arginine methylation of p53 by PRMT5 to orchestrate long-lasting oncogenic response. (ii) FA deficiency leads to a complete reversal of the response pattern characterized by short-lived oncogenic stress response accompanied with aberrant oncogene-induced arginine methylation of p53 resulted from a progressive decline in PRMT5 in FA HSPCs. Our studies indicate that a dysregulated PRMT5 in FA HSPCs plays an important role in suppressing p53-dependent tumor surveillance and suggest that PRMT5 may be a putative therapeutic target in FA tumorigenesis.
For Cre-mediated gene deletion, animals were injected i.p. with 100 μl of tamoxifen (20 mg/ml; Sigma-Aldrich, St. Louis, MO). Animals were maintained in the animal barrier facility at Cincinnati Children's Hospital Medical Center. All experimental procedures conducted in this study were approved by the Institutional Animal Care and Use Committee of Cincinnati Children's Hospital Medical Center.
For apoptosis staining, LSK cells were cultured in the presence of 4-OHT (125 mM; Sigma-Aldrich, St. Louis, MO) for 24 hours followed by Annexin V and 7AAD staining using BD ApoAlert Annexin V kit (BD Pharmingen, San Jose, CA) in accordance with the manufacturer's instruction. Apoptosis was analyzed by quantification of Annexin V-positive cell population by flow cytometry.
For cell cycle analysis, LSK cells were cultured in the presence of 4-OHT for 2 hours and released for indicated time intervals followed by staining with propidium iodide (PI) solution containing 1 mg/mL RNase A. In another set of cell cycle analysis, treated stained cells were fixed and permeabilized with Cytofix/Cytoperm buffer (BD Pharmingen, San Jose, CA) followed by intensive wash using Perm/Wash Buffer (BD Pharmingen, San Jose, CA). Cells were incubated with anti-mouse Ki67 antibody (BD Pharmingen, San Jose, CA) and Hochest 33342 (Sigma-Aldrich, St. Louis, MO) followed by Flow cytometric analysis.

IVIS imaging
Subjected mice pre-treated with indicated chemicals were injected by an intraperitoneal route with a luciferin solution (15 mg/ml in PBS (150 mg/kg)) that is allowed to distribute in awake animals for about 5-15 minutes. The mice are placed into a clear plexiglass anesthesia box (2.5-3% isofluorane) that allows unimpeded visual monitoring of animals. After the mice are fully anesthetized, they are transferred from the box to the nose cones attached to the manifold in the imaging chamber. The first image is taken approximately 5 minutes after luciferin injection. Continue to take images at indicated time points after the injection using IVIS series preclinical in vivo Imaging systems (PerkinElmer, Santa Clara, CA). The bioluminescence image signals were quantified by measuring photon flux using LiveImage Pro. 2.0 software (Caliper Life Science, Hopkinton, MA) as previously reported [62]. Briefly, specific areas of the image were analyzed by creating regions of interest (ROI) and measured the average # of photo/second. After subtracting the signal background of luciferin injected animals at time 0, the net photon flux (p/s/cm 2 /sr) and the photon flux were normalized to the respective values obtained at different time points after Tamoxifen injection and were plotted versus days after injection. GraphPad Prism was subsequently used to generate graphs for this manuscript.

Isolation of BM lineage-depleted cells
The femora and tibiae were harvested from the mice immediately after their sacrifice with CO 2 . Bone marrow (BM) cells were flushed from bones into Iscove's modified Dulbecco's medium (IMDM; Invitrogen, Grand Island, NY) containing 10% FCS, using a 21-guage needle and syringe. Low-density BM mononuclear cells (LDBMMNCs) were separated by Ficoll Hypaque density gradient (Sigma-Aldrich, St. Louis, MO) and washed with IMDM medium. LDBMMNCs were depleted of lineage-committed cells using a lineage cell depletion kit (Miltenyi Biotec, San Diego, CA) in accordance with the manufacturer's instruction.

Colony-forming unit assay
LSK cells (Lin -Sca1 + c-kit + cells) were cultured with 4-OHT (125 mM; Sigma-Aldrich, St. Louis, MO) for 48 hours followed by plating in a 35-mm tissue culture dish in 4 mL of semisolid medium containing 3 mL of MethoCult M3134 (Stem Cell Technologies, Vancouver, BC, Canada) and the following growth factors: 100 ng/mL SCF, 10 ng/mL IL-3, 100 ng/mL GM-CSF, and 4 units/mL erythropoietin (Peprotech, Burlington, NC). On day 7 after plating, erythroid and myeloid colonies were enumerated. Hematopoietic clonal growth results were expressed as means (of triplicate plates) ± SD of three experiments.

Quantitative PCR analysis
LSK cells were cultured in the presence of 4-OHT for 2 hours then released in fresh medium for the indicated time intervals. RNA was then extracted for qPCR analysis using primers listed in Table S1. Samples were normalized to the level of GAPDH mRNA.

Viral expression vectors
The lentiviral vector expressing FLAG-tagged fulllength FANCA was described previously [63]. The Myctagged full-length human PRMT5 cDNA was subcloned into lentiviral vector pLVX-IRES-GFP vector (Clontech, Mountain View, CA), using the gateway system (Life Technologies, Grand Island, NY). The MSCV-IRES-GFP-Myc ER retroviral vector was a gift from Dr. Cleveland JL at St. Jude Children's Research Hospital [39]. The plasmids (10 μg each) were used to produce retroviral supernatant.
For immunoprecipitation, protein A/G agarose beads (Santa Cruz Biotechnologies, Dallas, Texas) precleaned cell lysates were incubated with p53 antibody (R&D, Minneapolis, MN) by gentle rocking overnight at 4°C followed by incubation with protein A/G agarose beads for additional 1 hour at 4ºC. Pellets were then washed with 500 ml of lysis buffer and resuspended with 20 ml of 4X SDS sample buffer. Immunoblots were then incubated with primary antibodies specific for phosphor-serine (Millipore, Billerica, MA), acetylated lysine (Abcam, Cambridge, MA), methylated lysine (Abcam, Cambridge, MA), or mono-methyl arginine (Cell Signaling, Danvers, MA) antibodies for 12 to 16 h at 4 degree. Signals were revealed after incubation with anti-mouse or anti-rabbit secondary antibodies.

Statistics
Data were analyzed statistically using a Student's t test. The level of the statistical significance stated in the text was based on the P values. P < 0.05 was considered statistically significant.