ANP32B deficiency impairs proliferation and suppresses tumor progression by regulating AKT phosphorylation

The acidic leucine-rich nuclear phosphoprotein 32B (ANP32B) is reported to impact normal development, with Anp32b-knockout mice exhibiting smaller size and premature aging. However, its cellular and molecular mechanisms, especially its potential roles in tumorigenesis, remain largely unclear. Here, we utilize 'knockout' models, RNAi silencing and clinical cohorts to more closely investigate the role of this enigmatic factor in cell proliferation and cancer phenotypes. We report that, compared with Anp32b wild-type (Anp32b+/+) littermates, a broad panel of tissues in Anp32b-deficient (Anp32b−/−) mice are demonstrated hypoplasia. Anp32b−/− mouse embryo fibroblast cell has a slower proliferation, even after oncogenic immortalization. ANP32B knockdown also significantly inhibits in vitro and in vivo growth of cancer cells by inducing G1 arrest. In line with this, ANP32B protein has higher expression in malignant tissues than adjacent normal tissues from a cohort of breast cancer patients, and its expression level positively correlates with their histopathological grades. Moreover, ANP32B deficiency downregulates AKT phosphorylation, which involves its regulating effect on cell growth. Collectively, our findings suggest that ANP32B is an oncogene and a potential therapeutic target for breast cancer treatment.

The acidic leucine-rich nuclear phosphoprotein 32 kDa (ANP32) protein family are characterized by a N-terminal leucine-rich repeat domain and a C-terminal low-complexity acidic region. 1 In mammals, the ANP32 family has at least three members named ANP32A, ANP32B and ANP32E, and they regulate a wide spectrum of biological processes including chromatin regulation, 2-6 caspase activation, [7][8][9] protein phosphatase inhibition [10][11][12] and intracellular transport. 13,14 Although early investigations suggested that three ANP32 members functionally overlap, 10 they are reported to have diverse roles in cancer progression. ANP32A was shown to inhibit cell transformation [15][16][17] and has reduced expression in prostate and breast cancer. 18,19 ANP32E was reported to have enhanced expression in gastric cancer, 20 and a high expression of ANP32E was associated with better survival rate in follicular lymphoma. 21 Previously we reported that ANP32B, also designated as PHAPI2 or SSP29, is a negative prognostic indicator for human breast cancer. 22 Full analysis of the expression and functional role of ANP32B in cancer progression has still not been undertaken.
Knockout mouse studies demonstrated that loss of Anp32b, but not Anp32a and Anp32e, caused a high degree of perinatal lethality and reduced body weight, [22][23][24][25] indicating a greater importance of Anp32b in normal development. In addition, gene expression analysis indicates that elevated ANP32B mRNA expression correlates with highly proliferative tissues. 22 We also showed that ANP32B acts as a negative regulator of leukemic cell apoptosis, 8 and inhibits all-trans retinoic acid induced leukemic cell differentiation. 26,27 Although these studies strongly suggested ANP32B as a master regulator of cell fate determination, its cellular and molecular mechanisms are still not understood. Considering that some physiological and pathological processes share many common molecular regulators, 28 and ANP32B mRNA expression is a marker for aggressive breast cancer, 22 we proposed that ANP32B also functions in breast cancer. Here, we used Anp32b-knockout mice, multiple breast cancer cell lines and clinical patient samples to uncover the potential role for ANP32B in cell proliferation of both mouse embryo fibroblasts (MEFs) and breast cancer cells, and find that loss of ANP32B by knockout or RNAi silencing reduced rates of cell proliferation. We also show that RNAi silencing induces an extended G 1 -phase of the cell cycle. In addition, phosphorylation of AKT, an upstream regulator of cell cycle-associated proteins, is lower coincident with reduced ANP32B upon silencing and in both mouse and human cancers.

Results
Anp32b − / − MEFs are impaired in cell proliferation and oncogenic transformation. As seen in mixed-bred homozygous Anp32b − / − mice, 22 the Balb/c-congenic Anp32b − / − mice also had a statistically significant reduction of body weight compared with wild-type (Anp32b +/+ ) and heterozygous (Anp32b +/ − ) mice at 4 weeks after birth ( Figure 1a). Here we also found the decreased weigh of the Anp32b − / − mice was accompanied by the reduced size of organs such as the heart, liver, spleen and kidney (Figures 1b Figure 1 Anp32b deficiency impairs normal cell proliferation and oncogenic transformation. (a) The body weight of 22 Anp32b +/+ , 30 Anp32b +/ − and 8 Anp32b − / − mice at 4 weeks of age. Data were analyzed using Mann-Whitney U-test. **Po0.01. (b) Photographs of appearance and organs of Anp32b − / − mice and Anp32b +/+ littermate at 4 months of age. (c) The weight of organs from Anp32b +/+ and Anp32b − / − mice at 4 months of age. Data are presented as mean ± S.D. and significance is *Po0.05 (n = 4). (d) Western blots for the indicated protein in primary Anp32b +/+ and Anp32b − / − MEF cells. (e) Proliferation of primary Anp32b +/+ and Anp32b − / − MEFs was monitored for the indicated times. N t /N 0 represents the cell number at a given time normalized to the cell number at day 0. Data are presented as mean ± S.D. The experiment is representative of three separate experiments. (f) Western blots for the indicated protein in immortalized Anp32b +/+ and Anp32b − / − MEF cells. (g) Proliferation of immortalized Anp32b +/+ and Anp32b − / − MEFs was monitored for the indicated times. Data are presented as mean ± S.D. (h) Immortalized MEF cells were subcutaneously injected into nude mice and the size of masses were analyzed. Data are presented as mean ± S.D. To determine whether this is due to reduced cell volumes or hypoplasia, histological H&E staining and flow cytometric evaluation were performed. The results showed that there was no obvious cell size difference between these organs in Anp32b − / − and Anp32b +/+ mice (Supplementary Figure S1A and B). However, cell numbers in organs such as spleen and thymus were dramatically decreased in Anp32b − / − mice (Supplementary Figure S1C), supporting that Anp32b deficiency causes a hypoplastic phenotype in multiple organs.
To functionally characterize the role of Anp32b in normal cell proliferation, we isolated MEFs from Anp32b +/+ and Anp32b − / − mice. The ANP32B protein was totally knocked out in Anp32b − / − MEF cells (Figure 1d). Cell proliferation assay showed that primary Anp32b − / − MEF cells divided significantly more slowly than Anp32b +/+ MEF cells (Figure 1e). Given the effect of Anp32b on cell proliferation, we set out to assess whether Anp32b deficiency could inhibit oncogenic transformation. To this end, MEFs were immortalized by infection with a retrovirus encoding two oncogenes, adenovirus 5 E1A and constitutively active form of H-RasV12 ( Figure 1f). The results showed that the immortalized Anp32b − / − MEF cells also presented lower cell proliferation than Anp32b +/+ MEF cells (Figure 1g). We also subcutaneously injected immortalized Anp32b +/+ and Anp32b − / − MEF cells into nude mice. Mice xenografted with Anp32b − / − MEF cells developed significant smaller tumors compared with the mice injected with Anp32b +/+ MEF cells (Figure 1h). All these data suggest a functional role of Anp32b in the proliferation of normal and transformed cells.
ANP32B knockdown inhibits breast cancer cell proliferation in vitro. We further investigated whether ANP32B regulates cancer cell proliferation with breast cancer cells as models. For this purpose, we used two pairs of shRNAs (sh32b#1 and sh32b#2) specifically against ANP32B to generate stable ANP32B knockdown along with a control shRNA transfectant (shNC) in BT549, MCF7 and MDA-231-D3H2LN breast cancer cell lines. These two specific shRNAs could effectively knockdown ANP32B but not its closely related ANP32A expression in these breast cancer cell lines (Figure 2a and Supplementary Figure S2A). Then, we examined the effect of ANP32B knockdown on breast cancer cell proliferation. As shown in Figures 2b and c, ANP32B knockdown significantly inhibited the growth of BT549 cells with no effect on their viability. Similar effects could also be seen in MDA-231-D3H2LN ( Figure 2b) and MCF7 cells (Supplementary Figure S2B and C). Compared with the control cells, in addition, BT549 and MCF7 cells with ANP32B silencing showed markedly decreased colony formation ability with reduced colony number and size (Po0.05; Figure 2d and Supplementary Figure S2D and E). To demonstrate that cell growth inhibition is specifically due to the silencing of ANP32B, we re-introduced GFP-tagged ANP32B into sh32b#2-transfected MDA-231-D3H2LN cells, and found that re-expression of ANP32B could reverse ANP32B knockdown-induced cell growth inhibition (Figures 2e and g). Taken together, these data suggest that ANP32B may be closely associated with the proliferation of breast cancer cell lines.
ANP32B promotes cell cycle progression. To examine whether the effect of ANP32B on breast cancer cell proliferation is partly due to cell cycle arrest, we used a double thymidine block to synchronize cells at the G 1 /S border, followed by addition of nocodazole to block cells in G 2 /M. Flow cytometry analysis was then used to monitor the progression of cells from G 1 /S to G 2 /M. The results showed that after nocodazole treatment within 3 h 89.5% of cells in shNC BT549 cells entered the S phase, whereas cells at the S phase only had 59.9 and 63.8%, respectively, in sh32b#1 and sh32b#2-infected cells. After nocodazole treatment for 9 h, only 46.3 and 39.8% of total cells entered the G 2 /M phase in two sh32b-infected BT549 cells compared with 69.6% of total cells in shNC BT549 clones (Figure 3a). In addition, we analyzed the cell cycle regulatory proteins. As shown in Figure 3b, cyclin D1 protein level had no alteration between NC and sh32b-infected cells. However, cyclin D1 was time-dependently increased in shNC-infected BT549 cells upon nocodazole treatment, which was remarkably inhibited in sh32b-infected cells. Similarly, we re-introduced ANP32B into sh32B#2-transfected BT549 cells ( Figure 3c). Cell cycle analysis showed that complementation by ANP32B could rescue the cell cycle G 1 phase arrest in sh32b BT549 cells ( Figure 3d). Collectively, these results indicate that ANP32B promotes cell cycle progression at the G 1 phase.
Loss of ANP32B suppresses breast tumor growth in vivo. These in vitro results prompted us to examine whether ANP32B has some effects on breast tumor growth in vivo. Hence, shNC, sh32b#2 and sh32b#2/GFP-ANP32Binfected MDA-231-D3H2LN cell line (Figure 2e), which was derived from breast cancer cell line MDA-MB-231 with stable luciferase expression, were injected into the mammary fat pad of nude mice. Luciferase photon fluxes were monitored and the representative tumors are shown (Figures 4a and b). Consistent with the in vitro findings, reduction of ANP32B led to a significant inhibition of tumor size at the fourth week after injection, which could be partially reversed by re-expression of ANP32B (Figures 4a and b). Furthermore, ANP32B knockdown tumor cells showed obviously weaker Ki-67 staining compared with control tumor cells, which could be also reversed by re-expression of ANP32B, suggesting that ANP32B knockdown indeed decreased cell proliferation in vivo (Figure 4c). All these data strongly suggested that specific loss of ANP32B could significantly inhibit breast cancer growth in vivo.
ANP32B is highly expressed in human breast cancer. Previously, we examined the relationship between ANP32B mRNA expression and breast cancer patient prognosis using information from three available data sets and reported that patients whose tumors showed the highest ANP32B mRNA levels had significant shorter survival. 22 Here we performed Immunohistochemical staining (IHC) staining on 50 breast tumor tissues and the matched adjacent normal tissues, and found that breast tumor tissues presented higher ANP32B expression compared with adjacent normal tissues (Figures 5a and b). Moreover, an increase of ANP32B protein level in tumor tissues over adjacent normal tissues was also confirmed by western blot analysis in five paired clinical ANP32B deficiency suppresses proliferation and tumorigenesis S Yang et al breast cancer specimens (Figure 5c). These data indicate that ANP32B expression is enhanced in human breast cancer at the protein level.
We next evaluated the correlation between ANP32B expression and clinicopathological parameters. As presented in Supplementary Figure S3, there was no significant correction for ANP32B expression with age or clinical stage of breast cancer patients. However, ANP32B was associated significantly with histological grade. Higher levels of ANP32B was correlated with higher histological grade (I versus II; P = 0.0182, II versus III; P = 0.0231) (Figure 5d). Figure 5e depicts three representative IHC images respectively for low, medium and high ANP32B expressions of cancer tissues with different histological grade. These data suggest that elevated ANP32B protein expression in breast cancer is directly related with histological grade of cancer tissues.
ANP32B has positive correlation with p-AKT and regulates AKT activation. We analyzed the expressions of cyclins such as cyclin D1/3, cyclin-dependent kinases (CDKs) including CDK4, CDK6, CDK2, CDK inhibitor p27, as well as ERK and P38 in ANP32B silencing BT549 and MDA-231-D3H2LN cells. The results showed that knockdown of ANP32B failed to change all these protein levels (Supplementary Figure S4). More interestingly, ANP32B knockdown significantly reduced the phosphorylated AKT at Ser473 rather than AKT protein (Figure 6a). Of note, it did not change phosphorylated ERK and P38 (Supplementary (c) Cell viability after 6 days of growth was measured by trypan blue exclusion. Data are presented as mean ± S.D. of triplicate in an independent experiment, which was repeated for more than three times. (d) The morphology of shNC-and sh32b-infected BT549 cells under phase contrast microscopy (upper). Influence of ANP32B on colony formation of BT549 cells. Representative dishes are presented (middle). The number and size of clones were calculated for each well of six-well plates and shown in the y axis in the bottom panel. Data are presented as mean ± S.D. and significance is *Po0.05, **Po0.01, which was repeated for more than three times. (e) ShNC-and sh32binfected breast cancer MDA-231-D3H2LN cells were stably transfected with empty vector (EV) and GFP-tagged ANP32B, followed by immunoblots for the indicated proteins. (f) Cell counting of shNC/EV, sh32b/EV and sh32b/GFP-ANP32B MDA-231-D3H2LN cells after 3 days of growth. Data are presented as mean ± S.D. and significance is **Po0.01, which was repeated for more than three times. (g) Representative images from the morphology and colony formation of shNC/EV, sh32b/EV and sh32b/GFP-ANP32B MDA-231-D3H2LN cells ANP32B deficiency suppresses proliferation and tumorigenesis S Yang et al Figure S4). A similar impact of ANP32B on AKT phosphorylation was evident in Anp32b − / − MEF cells (Figure 6b). In line with this, a carcinogen 7,12-dimethylbenz(a)anthracene (DMBA)-induced mammary tumors 30 derived from Anp32b − / − but not Anp32b +/+ mice also displayed negative p-AKT staining expression (Figure 6c). We detected ANP32B and p-AKT expression by IHC staining in tumor tissues from breast cancer patients. Figure 6d depicts two representative IHC images for ANP32B and p-AKT expression. To better understand the correlation of ANP32B and p-AKT, we stratified the cohort into two groups based on ANP32B staining (ANP32B low and ANP32B high ), and P-AKT was scored using either median expression score (Figure 5e) or the percentage of low and high expression scores (Figure 5f). The results demonstrated a highly positive correlation between the ANP32B and P-AKT.
ANP32B regulates breast cancer cell proliferation through AKT activation. To further investigate the contribution of AKT signaling to the role of ANP32B in cell proliferation, we ectopically expressed AKT in ANP32B knockdown cells to evaluate whether it might overcome the suppression effect of ANP32B deficiency on cell proliferation. BT549 cells were stably co-transfected with shNC or sh32b#2 together with vector or flag-AKT, and the results showed that the p-AKT level was increased but still lower in sh32b cells compared with shNC BT549 cells ( Figure 6g). As expected, the ectopically expressed AKT could rescue ANP32B knockdown-induced cell growth inhibition in BT549 cells (Figure 6h). Considering that Akt overexpression-restored pAkt levels might be responsible for reversion of effects in ANP32B knockdown cells, the HA-myr-AKT with constitutive activation of AKT 31 was re-expressed in shNC and sh32b#2 BT549 cells (Figure 6i). Consistent with data in Figure 6h, the enforced expression of HA-myr-AKT could also rescue ANP32B knockdown-induced cell growth inhibition in BT549 cells (Figure 6j). All these results indicated that AKT signaling mediated ANP32B knockdown-induced cell growth inhibition.

Discussion
In this study, we used three models to examine the role of ANP32B in cell proliferation and oncogenesis. The knockout mouse model demonstrated that ANP32B has a broad impact on cell proliferation evidenced by hypoplasia in many organs, and that loss of ANP32B inhibits normal cell proliferation and Clinical data set analyses showed that ANP32B protein level is highly expressed in breast cancer patients and the elevated ANP32B protein expression is directly related with histological grade of breast cancer tissues. These data suggested that ANP32B acts as a predictive indicator in breast cancer treatment. However, owing to the limitation of patients sample information, relationships between ANP32B (e-f) Box plots of p-AKT scores (e) and the percentage of tumors with high and low p-AKTexpressions (f) in those with high and low ANP32B expressions. (g) ShNC-and sh32b-infected breast cancer BT549 cells were stably transfected with empty vector (EV) and Flag-AKT, followed by immunoblots for the indicated proteins. (h) ShNC-and sh32b-infected breast cancer BT549 cells were stably transfected with empty vector (EV) and HA-myr-AKT, followed by immunoblots for the indicated proteins. (i) Cell counting of EV-and Flag-AKT-transfected BT549 cells after 3 days of growth. Data are presented as mean ± S.D. and significance is **Po0.01, which was repeated for more than three times. (j) Cell counting of EV-and HA-myr-AKTtransfected BT549 cells after 3 days of growth. Data are presented as mean ± S.D. and significance is **Po0.01, which was repeated for more than three times overexpression and clinical prognosis were not fully analyzed. Increased ANP32B in tumors and knockdown models also correlated with high p-AKT expression, indicating a possible mechanism through which ANP32B exerts its effect on cell proliferation and tumor progression.
The activated AKT pathway has been demonstrated to have an essential role in normal cell and breast cancer cell proliferation. [32][33][34][35][36] We found that the p-AKT level was significantly decreased in ANP32B knockdown cells. Furthermore, the restoration of AKT or constitutively active AKT expression could rescue the inhibition of cell proliferation by ANP32B deficiency, suggesting the inhibition of cell proliferation by ANP32B deficiency is primarily mediated through AKT activation in breast cancer cells. How ANP32B might regulate the AKT activation is still unknown. Previous studies have reported that AKT activation could be regulated by many genes, including PH domain leucine-rich repeat protein phosphatase, 37 serine/threonine protein phosphatase 2A (PP2A), 38 pyruvate dehydrogenase kinase, isozyme 1 (ref. 39) and phosphatase and tensin homolog (PTEN), 40 but our preliminary experiments showed that ANP32B failed to interact with PP2A and PTEN (data not shown). So future experiments will be needed to investigate the detailed mechanism about how ANP32B regulates AKT activation.
Totally, our results concluded that ANP32B, through its positive regulation of p-AKT, serves as a master enforcer of cell proliferation. In the physiological context, knockout of ANP32B impedes the proper mammalian development, whereas in the pathological context, ANP32B deficiency functions as a suppressor of tumor growth and transformation. Notably, ANP32B has been highly detected in breast cancer patients, thus highlighting ANP32B as a potential therapeutic target for breast cancer treatment.

Materials and Methods
Cell lines and cell culture. Human breast cancer cell lines BT549 and MCF7 were obtained from the cell bank of the Chinese Academy of Sciences (Shanghai, China). MCF7 were cultured in Dulbecco's modified Eagle's medium (Hyclone, Logan, UT, USA) with 10% FBS and 0.01 mg/ml human recombinant insulin. BT549 was maintained in RPMI 1640 (Hyclone) with 10% FBS. The cell line MDA-231-D3H2LN (Xenogen, Alameda, CA, USA) was propagated in Minimum essential medium with Earl's balanced salts solution (Hyclone) medium supplemented with 10% FBS, 1% non-essential amino acids (Hyclone) and 1% sodium pyruvate (Hyclone). Primary MEFs were prepared from littermate Anp32b +/+ and Anp32b − / − E14.5 embroys. For transformed MEFs, primary MEFs were infected with a retrovirus generated from pLPC E1A/ras v12 using published techniques. 29 All cells were fostered in a humidified atmosphere of 5% CO 2 .
Patients. Fifty pairs of formalin-fixed and paraffin-embedded specimens of breast cancer and adjacent normal tissues were purchased from Shanghai Outdo Biotech Co (Shanghai, China). Detailed information is described in the Supplementary Table  S1. We obtained formalin-fixed and paraffin-embedded tumor specimens of breast cancer patients, which were histopathologically diagnosed during January 2003 and June 2010 in the Department of Surgery, Shanghai First People's Branch Hospital. All tumors were primary and were untreated before surgery. Detailed information is described in the Supplementary Table S2. In addition, we also collected five pairs of breast cancer and adjacent normal tissue specimens from Rui-Jin Hospital affiliated to Shanghai Jiao-Tong University School of Medicine for analyzing ANP32B protein expression. These studies were approved by the Medical Ethical Committee of the Affiliated Hospitals, Shanghai First People's Branch Hospital and Rui-Jin Hospital, respectively, and informed consent was obtained from all subjects or their relatives.
IHC. The protein expression levels of ANP32B and p-AKT were analyzed by IHC with anti-ANP32B and anti-p-AKT polyclonal antibody. All of the staining was assessed by pathologists who were blinded to the origin of the samples and subject outcome. Each specimen was assigned a score according to the intensity of the nucleus, cytoplasmic and/or membrane staining (no staining = 0; weak staining = 1, moderate staining = 2, strong staining = 3) and the area extent of stained cells (0% = 0, 1-24% = 1, 25-49% = 2, 50-74% = 3, 75-100% = 4). The final immunoreactive score (IRS) was determined by multiplying the intensity score with the extent score of stained cells, ranging from 0 (the minimum score) to 12 (the maximum score). Scores of ANP32B and p-AKT were divided into two classifications: low (IRS ⩽ 6) and high (IRS46). Detailed information of the two cohorts was shown in Supplementary Table S3. Cell proliferation and colony formation assay. MEF cells proliferation was evaluated by the CCK-8 assay (WST-8; Cell counting kit-8 from Dojindo, Kumamoto, Japan). In brief, each well was pulsed by addition of 10 μl WST-8 for 2 h. Absorbance readings at a wavelength of 450nm were taken on Synergy H4 Hybrid Microplate Reader. Breast cancer cells were plated on 6-cm dishes and were counted every 2 days. Cells were stained with a 0.4% trypan blue solution and counted with a hemacytometer. For colony formation assay, 500 cells were placed in complete growth media and allowed to grow until visible colonies formed in a fresh six-well plate (2 weeks). Cell colonies were fixed with cold methanol, stained with 0.1% crystal violet for 30 min, washed, air dried, photographed and counted.
Cell cycle analysis. For cell cycle analysis, all these BT549-transfected cells were pretreated with 2 mM thymidine twice to synchronize cells at G1/S border, and then treated with 100 ng/ml nocodazole to block cells in G2/M for indicated times. To analyze cellular DNA content by flow cytometry, 10 6 cells were collected, rinsed and fixed overnight with 75% cold ethanol at − 20°C. Cells were then treated with 100 μg/ml RNase A in Tris-HCl buffer (pH 7.4) and stained with 25 μg/ml propidium iodide. Samples were then subjected to the analysis by flow cytometry (FACSCalibur, BD Biosciences, San Jose, CA, USA) using CellQuest Pro software (BD Biosciences). Ten thousand cells were acquired and analyzed for the DNA content.
Animal experiments. Six-week-old female BALB/c nude mice were obtained from Shanghai SLAC Laboratory Animal Co., Shanghai, China. We subcutaneously injected 2 × 10 6 tumor cell lines into left abdominal mammary fat pad. Starting 4 weeks post injection, the tumor size was monitored weekly by bioluminescence imaging. Mice were anesthetized each time and given intraperitoneal injection of d-luciferin (150 μg/g body weight prepared in phosphate-buffered saline), and 10-15 min after the injection, bioluminescence images were captured with a chargecoupled device camera (IVIS; Xenogen). Mice were manipulated and housed according to protocols approved by Shanghai Medical Experimental Animal Care ANP32B deficiency suppresses proliferation and tumorigenesis S Yang et al