Negative regulation of AMPKα1 by PIM2 promotes aerobic glycolysis and tumorigenesis in endometrial cancer

Xue Han1 ● Chune Ren1 ● Tingting Yang1 ● Pengyun Qiao1 ● Li Wang1 ● Aifang Jiang1 ● Yuhan Meng1 ● Zhijun Liu2 ●
Yu Du2 ● Zhenhai Yu1

Received: 17 August 2018 / Revised: 9 December 2018 / Accepted: 16 April 2019
© The Author(s), under exclusive licence to Springer Nature Limited 2019

Abstract
Endometrial cancer (EC) is one of the most common gynecologic malignancies. However, the molecular mechanisms underlying the development and progression of EC remain unclear. Here, we demonstrated that the protein proviral insertion in murine lymphomas 2 (PIM2) was necessary for maintaining EC tumorigenesis in vivo and in vitro, and could inhibit AMPKα1 kinase activity in EC cells. Specifically, we found that PIM2 bound to AMPKα1, and directly phosphorylated it onThr467. Phosphorylation of AMPKα1 by PIM2 led to decreasing AMPKα1 kinase activity, which in turn promoted aerobic
glycolysis and tumor growth. In addition, PIM2 expression positively correlated with AMPKα1 Thr467 phosphorylation in EC tissues. Further, treatment with a combination of the PIM2 inhibitor SMI-4a and the AMPKα1 activator AICAR could effectively inhibit tumor growth. Thus, our findings provide insight into the role of PIM2 and AMPKα1 in EC and suggest that combination targeting of these proteins may represent a new strategy for EC treatment.

Introduction
Proviral insertion in murine lymphomas 2 (PIM2) belongs to a family of serine/threonine kinase encoded by proto- oncogenes, and which also includes the highly homologous proteins, PIM1 and PIM3 [1]. PIM2 is highly expressed in multiple cancers, including blood cancer, breast cancer, prostate cancer, liver cancer, and colon carcinoma [2]. This protein phosphorylates many proteins that are essential for tumor progression [3], including the important cell cycle regulators, p21 and p27 [4, 5]. PIM2 also activates the

These authors contributed equally: Xue Han, Chune Ren, Tingting Yang
mTOR-C1 pathway via inhibition of tuberous sclerosis complex 2 (TSC2), which maintains cell growth in multiple myelomas [6]. Moreover, phosphorylation of pyruvate kinase M2 and Hexokinase 2 (HK2) by PIM2 could enhance glycolysis and promote tumor growth [7–9]. Based on these observations, a several small molecule inhibitors for PIM2 kinase activity have been developed for the treatment of tumors [10–12]. PIM2 kinase also plays a prominent role in suppressing T-cell responses, thus pro- viding a strong rationale for targeting PIM2 in cancer immunotherapy [13]. In addition, PIM2 has other functions that are distinct from, and independent of, its kinase activity. For example, it can act as a co-factor of hypoxia inducible factor-1α (HIF-1α) [14] and plays a role in the degradation of the RNA-binding protein, tristetraprolin (TTP) [15]. Endometrial cancer (EC) is the fourth most

Supplementary information The online version of this article (https:// doi.org/10.1038/s41388-019-0898-z) contains supplementary material, which is available to authorized users. common malignancy among women, which is a major cause of women cancer-related morbidity and mortality
[16]. In China, the EC incidence has been increasing in
* Zhenhai Yu [email protected]

1 Department of Reproductive Medicine, Affiliated Hospital of Weifang Medical University, Weifang, Shandong Province, PR China
2 Department of Medical Microbiology, Weifang Medical University, Weifang, Shandong Province, PR China
recent years [17], which highlights the importance of developing improved treatments for this condition. How- ever, the precise mechanisms underlying the role of PIM2 in EC are not well understood.
The metabolic regulator protein, 5′ AMP-activated pro-
tein kinase (AMPK), is a tumor suppressor that consists of a catalytic α subunit, as well as regulatory β and γ subunits

The α catalytic subunit activation loop is phosphory- lated at T172 by liver kinase B1 (LKB1), which is required for AMPK kinase activity [19]. Under starvation conditions,AMP is produced via the activity of adenylate kinase when ATP is converted to ADP, and this increased cellular AMP binds to the γ regulatory subunit of AMPK, inducing the allosteric activation of this protein [18]. In addition, increased levels of ADP can also bind to the γ regulatorysubunit to promote AMPK kinase activity [18]. The phos-phorylation of AMPK regulatory subunits also plays an essential role in regulating its kinase activity [20]. For example, glycogen synthase kinase 3 constitutively interacts with the AMPK heterotrimeric kinase complex and inhibits AMPK kinase activity under anabolic conditions [21]. In addition, p70S6 kinase phosphorylates AMPK on S491 to mediate the effect of leptin on food intake [22], and phos- phorylation of AMPK by AKT promotes cell growth and proliferation [23]. Notably, a previous study also reported that PIM protein kinases could inhibit activation of AMPK [24], but the molecular mechanisms underlying the regula- tion of AMPK by PIM family proteins have not been elucidated.

Here, we found PIM2 played an essential role in the invasion and proliferation of EC cells. Specifically, PIM2
could constitutively interact with AMPKα1, and phos- phorylated it on Thr467. Phosphorylation of AMPKα1 by PIM2 directly inhibited AMPKα1 kinase activity and pro- moted tumor growth in vivo. Moreover, immunohisto-
chemical analysis of tumors from EC patients showed that PIM2 and pT467-AMPKα1 expressions were positively correlated with tumor size and TNM stage. In addition, a combination of the PIM2 inhibitor SMI-4a and the
AMPKα1 activator AICAR effectively inhibited tumor growth in vivo. Thus, based on these data, we propose that PIM2-induced inhibition of AMPKα1 may represent a potential target for the development of improved treatments
for EC.

Results

PIM2 is essential for EC cell migration, invasion, and proliferation

To determine if PIM2 contributes to tumor progression of EC, we stably knocked down PIM2 using shRNA and validated it by western blot in HEC-1B and Ishikawa cells (Fig. 1a). We then performed Cell Counting Kit-8 (CCK-8) assays and found that knockdown of PIM2 significantly suppressed both cell proliferation and viability (Fig. 1b). Colony formation assays further revealed that EC cells containing shRNA targeting PIM2 (shPIM2) produced smaller and fewer colonies than cells containing the control
RNA (shNC group) (Fig. 1c). To further determine whether PIM2 may also be involved in cell migration and invasion, we performed wound-healing and transwell assays in EC cells containing shPIM2. PIM2 knockdown significantly inhibited cell migration and invasion in EC cells, as com- pared with shNC control (Fig. 1d, e). Collectively, these data suggest a role for PIM2 in EC cell proliferation, as well as in cell migration and invasion.

We then used a xenograft mouse model to directly investigate the role of PIM2 in tumor growth in vivo. Specifically, stable shPIM2 Ishikawa cells, or those expressing the shNC control, were subcutaneously implanted into 4-week-old nude mice, and tumor volume was measured every 7 days. Tumor nodule size was examined (Fig. 1f), and growth curves were plotted for each group (Fig. 1g). Tumor weight was significantly decreased in the shPIM2 group, as compared with the control group (Fig. 1h). Tumor samples were further processed for hematoxylin and eosin (H/E) and Ki67 staining, revealing a significant inhibitory effect of PIM2 knockdown on the proliferation of Ishikawa cells (Fig. 1i). Thus, these results suggest that PIM2 plays a critical role in EC cell pro- liferation in vivo.

PIM2 inhibits AMPKα1 kinase activity

A previous study reported that PIM protein kinases could inhibit activation of AMPK, a key regulator of cell meta- bolism [24]. Phosphorylation of AMPK at T172 by LKB1 could reflect its kinase activity [25]. Therefore, to further assess the role of PIM2 in AMPKα1 kinase activity reg- ulation in EC, we overexpressed Flag-tagged PIM2 in EC cells and found that PIM2 overexpression could inhibit AMPK phosphorylation at pT172 and promote decreased AMPK activity (Fig. 2a, b). Conversely, shRNA-mediated knockdown of PIM2 resulted in increased AMPKα1 kinase activity (Fig. 2c, d). Consistent with this observation,treatment of EC cells with the PIM2 kinase inhibitor, SMI- 4a, also promoted increased AMPKα1 kinase activity (Fig. 2e, f).
We then used an AMPK kinase assay kit to directly measure the effect of PIM2 on AMPKα1 kinase activity in vivo and found that PIM2 could directly inhibit AMPKα1 kinase activity in EC cells (Fig. 2g–i). Collec- tively, these data provide evidence in support of the hypothesis that PIM2 negatively regulates AMPKα1 kinase activity in EC cells.

PIM2 interacts with AMPKα1

To further determine whether PIM2 directly interacts with AMPKα1, we performed immunoprecipitations and immunoblot analyses on extracts from HEK293T cells PIM2 is essential for EC cell migration, invasion and pro- liferation. a shRNA-PIM2 was stably expressed in EC cells followed by western blotting using indicated antibodies. b Cell proliferation was tested by the CCK-8 assay in shRNA-PIM2 EC cells. c Plate colony formation assay was used to measure cell proliferation of shRNA- PIM2 EC cells. d Wound-healing assay was performed to measure cell migration in shRNA-PIM2 EC cells. e Transwell assay was performed to test cell invasion in shRNA-PIM2 EC cells. f Photographs of tumors
excised 3 weeks after inoculation of shRNA-PIM2 EC cells into nude mice. g Showing the comparative growth rate of tumors formed from shRNA-PIM2 EC cells. One week after injection the size of the tumor was measured every week. h After 3 weeks, the nude mice were killed, tumor weights were measured. i Nude mice tumor tissues were examined by hematoxylin and eosin (H/E), Ki67, and PIM2 (scale bar, 20 µm). (All data represent mean ± SEM n ≥ 3), *p < 0.05 transfected with Flag-tagged PIM2 and HA-tagged AMPKα1. Flag-tagged PIM2 protein was detected in immunoprecipitations of HA-tagged AMPKα1 and vice versa (Fig. 3a, b). Further, anti-AMPKα1 antibodies were capable of immunoprecipitating PIM2 protein from Ishi- kawa cell extracts (Fig. 3c), and anti-PIM2 antibodies could immunoprecipitate AMPKα1 protein as well (Fig. 3d). Consistent with these data, GST pull-down assays demon-
strated that recombinant His-tagged PIM2 could interact with GST-tagged AMPKα1, but did not interact with the GST tag alone (Fig. 3e). Overall, these results indicate that PIM2 binds directly to AMPKα1.

We then performed immunofluorescence confocal microscopy and found that PIM2 colocalized with AMPKα1 mainly in the cytoplasm, whereas only a small amount was detected in the nucleus (Fig. 3f). To determine which domain(s) of PIM2 or AMPKα1 are responsible for this interaction, we generated three truncated forms of both PIM2 and AMPKα1 as GFP fusion proteins and measured their interactions with Co-IP assays (Fig. 3g, h). AMPKα1 bound to the kinase domain of PIM2, whereas PIM2 interacted with an internal region of AMPKα1 (aa residues 281–340) (Fig. 3i, j), further supporting a direct interaction between PIM2 and AMPKα1.

PIM2 inhibits AMPKα1 kinase activity. a, b Flag-tagged PIM2 or empty vector control was expressed in EC cells. The cell lysates were analyzed by immunoblotting. c, d Western blot analysis to determine the indicated protein levels in stably knockdown PIM2 EC cells. e, f PIM2 inhibitor SMI-4a (10 μM) was added in EC cells for 24 h, and the cell lysates were analyzed by immunoblotting. g AMPKα1 proteins were enriched by the IP assay in Flag-tagged PIM2 overexpressing Ishikawa cells. AMPKα1 kinase activity was determined by kit. (Data represent mean ± SEM n = 3), *p < 0.05. h AMPKα1 proteins were enriched by the IP assay in stable shRNA- PIM2 Ishikawa cells. AMPKα1 kinase activity was determined by kit. (Data represent mean ± SEM n = 3), *p < 0.05. i AMPKα1 proteins were enriched by IP the assay in Ishikawa cells which were cultured with SMI-4a (10 μM) for 24 h. AMPKα1 kinase activity was deter- mined by kit. (Data represent mean ± SEM n = 3), *p < 0.05

PIM2 phosphorylates AMPKα1 on Thr467 AMPKα1 kinase activity has previously been shown to be regulated by phosphorylation on serine or threonine resi- dues [26]. We therefore hypothesized that PIM2 might similarly regulate AMPKα1 kinase activity through phos-
phorylation. To test this possibility, exogenous HA-tagged AMPKα1 protein was immunoprecipitated from HEK293T cells overexpressing either wild-type (WT) PIM2 or the K61A kinase-dead mutant, both with a Flag-tag. WT PIM2 could promote increased levels of pThr-AMPKα1, but not pSer-AMPKα1 (Fig. 4a, b). Moreover, we identified a PIM2 consensus site at Thr467 in the AMPKα1 protein sequence that could be phosphorylated by PIM2 (Fig. 4c). Therefore, to assess the potential for PIM2-mediated mod-ification of AMPKα1 at Thr467, we generated the AMPKα1 mutant, T467A. This mutation abolished PIM2-induced threonine phosphorylation on AMPKα1, suggesting that the Thr467 residue contributes to the PIM2-dependent AMPKα1 phosphorylation signal (Fig. 4d).
Next, to more precisely monitor the phosphorylation status of AMPKα1 at Thr467, we generated a site-specific antibody (pT467-AMPKα1), the specificity of which was validated using peptide competition assays in Ishikawa cells

PIM2 interacts with AMPKα1. a, b HEK293T cells were cotransfected with HA-tagged AMPKα1 and Flag-tagged PIM2. Immunoprecipitation was performed with anti-HA agarose or anti-Flag agarose (IgG as a control). Immunoprecipitates and the cell extracts were analyzed by immunoblotting for the indicated proteins. c, d Lysates of Ishikawa cells were immunoprecipitated with the antibodies against PIM2 or AMPKα1. Immunoprecipitates and the cell extracts were analyzed by immunoblotting for the indicated proteins. e GST- AMPKα1 and His-PIM2 were expressed in bacteria. The GST fusion proteins were immobilized on glutathione sepharose beads and then incubated with His-PIM2 protein. Samples were analyzed by immunoblotting for the indicated proteins. f Ishikawa cells were cotrans- fected with HA-tagged AMPKα1 and GFP-tagged PIM2.

Immunofluorescence was performed with anti-HA antibody. Cover- slips were examined by confocal microscope. g Schematic repre- sentation of PIM2 truncation mutants. h Schematic representation of AMPKα1 truncation mutants. i HEK293T cells were transfected with
HA-tagged AMPKα1 and GFP fusion proteins of PIM2 truncation mutants. Immunoprecipitation was performed with anti-HA agarose.
Immunoprecipitates and the cell extracts were analyzed by immuno- blotting for the indicated proteins. j HEK293T cells were transfected with HA-tagged PIM2 and GFP fusion proteins of AMPKα1 trunca- tion mutants. Immunoprecipitation was performed with anti-HA agarose. Immunoprecipitates and the cell extracts were analyzed by immunoblotting for the indicated proteins and EC tissues (Supplementary Fig. S1). We then utilized this antibody to determine

whether PIM2 could directly phosphorylate AMPKα1 at Thr467. Using in vitro kinase assays, we found that recombinant His-tagged PIM2 was
able to phosphorylate GST-tagged AMPKα1 on Thr467 (Fig. 4e). However, treatment with the PIM2 kinase inhi- bitor SMI-4a abrogated the effects of PIM2 on AMPKα1 phosphorylation at Thr467 (Supplementary Fig. S2a). Moreover, we found nutrient of glucose could increase
AMPKα1 T467 phosphorylation level via inducing PIM2 expression in Ishikawa cells (Supplementary Fig. S2b). To determine whether phosphorylation at Thr467 specifically affects AMPKα1 kinase activity, we immunoprecipitated WT and T467A mutant HA-tagged AMPKα1 with Flag- tagged PIM2, using Co-IP assays. Consistent with our previous results, mutant T467A AMPKα1 showed increased kinase activity compared with WT protein (Fig. 4f). Lastly, to further confirm that AMPKα1 kinase activity

PIM2 phosphorylates AMPKα1 on Thr467. a, b HEK293T cells were transfected with HA-tagged AMPKα1 and Flag- tagged PIM2 (empty vector, K61A or WT) for 48 h. The cell lysateswere immunoprecipitated with HA antibody, and analyzed by immu- noblotting for the indicated antibodies. c A putative PIM2 substrate motif was identified in AMPKα1. d HEK293T cells were over-expressed HA-tagged AMPKα1 (T467A) and Flag-tagged PIM2 (WTor K61A) proteins. The cell lysates were immunoprecipitated with HA antibody, and analyzed by immunoblotting for the indicated anti- bodies. e Purified GST-tagged AMPKα1 was incubated with the indicated purified His-tagged PIM2 protein. An in vitro kinase assay was performed. f HEK293T cells were overexpressed HA-tagged AMPKα1 (WT, T467A or T467D). The cell lysates were immuno- precipitated with HA antibody, and analyzed by CycLex® AMPK
Kinase Assay Kit. (Data represent mean ± SEM n = 3), *p < 0.05 is regulated by phosphorylation at T467, we constructed Ishikawa/AMPKα1 shRNA knockdown cells, with recon- stituted expression of either WT, T467A, or T467D AMPKα1 (Supplementary Fig. S2c, d). The AMPKα1 T467A mutant displayed significantly increased levels of phosphorylation at pT172, which suggests that phosphor- ylation of AMPKα1 on Thr467 decreases the kinase activity of this protein by inhibiting phosphorylation at T172 (Supplementary Fig. S2e). Thus, our data suggest that PIM2 can directly phosphorylate AMPKα1 on Thr467, which functions to downregulate the kinase activity of this protein.

AMPKα1 pT467 mediates aerobic glycolysis and tumor growth Because AMPKα1 is known to act as a sensor for glucose metabolism [27], we next knocked down AMPKα1 using shRNA in Ishikawa cells and measured glucose metabo- lism. As predicted, knockdown of AMPKα1 increased glucose uptake and lactate production (Supplementary FigS3a, b). To then determine whether AMPKα1 is required for EC tumor growth in vivo, we established a mouse xenograft model by subcutaneously injecting 4-week-old nude mice with stable shAMPKα1 Ishikawa cells or cells
expressing the shNC control. Notably, AMPKα1 knockdown increased the rate of tumor growth and resulted itumors with higher weights than those from vector control. These data suggest that AMPKα1 functions as a critical regulator of tumor growth in EC cells (Supplementary Fig.
S3c–f). To further investigate the effects of PIM2 knock- down on pT467-AMPKα1 phosphorylation, we performed a rescue experiment in PIM2 shRNA knockdown cells expressing WT or mutant AMPKα1. Rescue of the AMPKα1 T467D mutant significantly increased cell pro- liferation and migration of Ishikawa cells (Supplementary Fig. S4). As a negative regulator of the Warburg effect, AMPKα1 is known to suppress tumor growth in EC cells. We therefore asked whether AMPKα1 phosphorylation at pT467 could regulate tumor cell glycolysis. To address this
question, we measured glycolysis in EC cells expressing WT and mutant AMPKα1 and found that the T467A mutant showed significantly decreased glucose consumption, lac- tate production, and ATP production (Fig. 5a–c). Moreover, the reconstituted expression of AMPKα1 T467A in stable shAMPKα1 Ishikawa cells inhibited cell proliferation (Fig. 5d).

To further measure the effect of AMPKα1 phosphor- ylation at Thr467 in vivo, we subcutaneously injected 4- week-old nude mice with shAMPKα1 Ishikawa cells dis- playing reconstituted expression of WT, T467A, or T467D AMPKα1. Mice were dissected 3 weeks after injection, revealing that tumors arising from cells expressing either AMPKα1 pT467 mediates aerobic glycolysis and tumor growth. a–c AMPKα1-depleted Ishikawa cells were reconstituted with the indicated protein expression. Media were collected for analysis of glucose consumption (a), lactate production (b), or ATP production (c), which was normalized by total proteins. d CCK-8 assay was performed todetermine cell proliferation of AMPKα1-depleted Ishi- kawa cells with reconstituted expression of the indicated proteins. Representative tumors extracted from tumor-bearing mice established by subcutaneously inoculating with xenografts of Ishikawa cells. f Tumors weight in subcutaneous tumor bearing mice model. g Tumors volume in subcutaneous tumor bearing mice model. h Representative images of H/E staining and Ki67 staining of tumor samples (scale bar, 20 μm). (All data represent mean ± SEM n ≥ 3), *p < 0.05

WT or T467D AMPKα1 showed more rapid growth than those arising from T467A AMPKα1 (Fig. 5e). Similarly, assessment of tumor volume and weight in each group of animals indicated that phosphorylation of AMPKα1 at T467 promotes tumor growth in vivo (Fig. 5f, g). Consistent with
this interpretation, Ki67 staining demonstrated that expres- sion of AMPKα1 with the T467A mutation has a significant inhibitory effect on proliferation of Ishikawa cells (Fig. 5h).Collectively, these results highlight the significance of PIM2-dependent AMPKα1 T467 phosphorylation at Thr467 for EC growth and development in vivo.

Combination of SMI-4a and AICAR to treatment of EC is efficacious in vivo Our data are consistent with a model whereby PIM2 phosphorylates AMPKα1 on residue Thr467, which func- tions to inhibit the kinase activity of AMPKα1 and induce Combination of SMI-4a and AICAR to treatment of EC is efficacious in vivo. a Ishikawa cells were treated with PBS (control), 0.5 mM AICAR, 10 μM SMI-4a, or 0.5 mM AICAR combined with 10 μM SMI-4a for 2 days. Cell
proliferation rates were measured by cell counting. b Ishikawa cells (1 × 107 each) were injected into the flanks of BALB/c nude mice. After
1 week, mice were randomly subjected to control, SMI-4a, AICAR or a combination of the drugs given every 2–3 days at the same dose. Three weeks after
implantation, tumors were isolated from the treated mice. Pictures of removed tumors are shown. c Tumor volume was assessed after injection of indicated inhibitors. d Tumor weight was weighted and represented. e Representative IHC images. Xenograft tumors from mice treated with control, SMI-4a, AICAR or the combination, respectively, were assessed using H/E staining and Ki67 antibodies (scale bar,
20 μm). (All data represent mean ± SEM n ≥ 3), *p < 0.05cell proliferation. To further confirm that PIM2-mediated phosphorylation of AMPKα1 can promote cell proliferation in vitro and in vivo; we utilized the PIM2 inhibitor, SMI-4a, and the AMPKα1 activator, AICAR, in in vitro cell pro- liferation assays and in our mouse xenograft model.

Cell proliferation was reduced in Ishikawa cells treated with each agent alone or in those treated with a combination of both compounds (Fig. 6a). Consistent with this, we observed reduced tumor growth in mice treated with each agent alone. In addition, tumor growth was completely arrested in animals treated with both compounds, and tumor size was significantly smaller in this group than in other groups of mice (Fig. 6b–d). We then performed Ki67 staining on tumors from each of the four groups of Ishikawa xenograft tumors. In groups treated with SMI-4a, AICAR, or SMI-4a + AICAR, Ki67 expression levels were significantly lower than in controls (Fig. 6e). These data further indicate that tumors treated with both SMI-4a and AICAR are less proliferative than those treated with either single agent. Thus, our results suggest that AICAR exerts a substantial chemotherapeutic effect to potentiate the effects SMI-4a on xenograft growth in mice.

PIM2 expression is positively correlated with pT467- AMPKα1 in EC tissues
Lastly, to measure expression of PIM2 and pT467- AMPKα1 in normal endometrium and EC tissue, we . 7 PIM2 expression is positively correlated with pT467-AMPKα1 in EC tissues. a Immunochemistry of PIM2 in endometrial tissues (scale bar, 20 μm). b Semiquantitative immunohistochemical analysis of 10 normal and 40 EC tissues for PIM2. *p < 0.05. c Immunochemistry of pT467- AMPKα1 in endometrial tissues (scale bar, 20 μm). d Semiquantitative
immunohistochemical analysis of 10 normal and 40 EC tissues for pT467-AMPKα1. *p < 0.05.
e Pearson correlative analysis of semiquantitative staining scores for PIM2 and pT467-AMPKα1. The standard curve was drawn by linear regression of the
correlation scores. Correlation is shown using r and significance was determined using a Spearman correlation. f Schematic diagram of the proposed PIM2-
AMPKα1 signaling pathway performed immunohistochemical analysis with anti-PIM2 and pT467-AMPKα1 antibodies. These data revealed that PIM2 and pT467-AMPKα1 are strongly expressed in the cytoplasm and nucleus of most cells in EC but show markedly lower expression in normal endometrium tissue (Fig. 7a, c). Quantification of these data confirmed that PIM2 and pT467-AMPKα1 levels are significantly higher in tumor tissues than in normal tissues (Fig. 7b, d). Statis-tical analyses further indicate that both PIM2 and pT467- AMPKα1 levels are significantly correlated with histologic grade (P = 0.001 and P = 0.016, respectively) and TNM stage (P = 0.002 and P = 0.043, respectively) (Table 1).

We also identified a positive correlation between PIM2 and pT467-AMPKα1 expression (Fig. 7e). These results support a role for PIM2 and pT467-AMPKα1 in the clinical beha- vior of human EC tumors and reveal a relationship between pT467-AMPKα1 and clinical tumor aggressiveness.

Discussion

As an oncogene, PIM2 kinase plays an essential role in regulating multiple signals in tumor progression, and has many substrates, including MDM2, p21, p27, TSC2, p65, c- MYC, and AR [2]. However, although PIM2 upregulation has been reported in a number of cancers, to our knowledge, there are no published reports describing PIM2 expression in EC tissues. Here, we performed immunohistochemical staining with anti-PIM2 antibodies and found that PIM2 expression is significantly upregulated in EC tissues, as compared with normal endometrium. Knockdown of PIM2 with shRNA also decreases EC cell proliferation both in vitro and in vivo, suggesting that this protein plays a significant role in the progression of human EC tumors and that targeting PIM2 may represent a new strategy for EC treatment.

Although PIM2 is a serine/threonine kinase, it also has cellular functions that are independent of its kinase activity.
For example, PIM2 acts as a cofactor of HIF-1α by binding to the transactivation domain of this protein to enhance
HIF-1α activity in response to hypoxia [14]. In addition, PIM2 interacts with TTP, and inhibits its activity, inde- pendent of PIM2 kinase activity, which promotes breast cancer tumorigenesis [15]. However, despite these obser- vations, most functions of PIM2 are dependent on its kinase activity. In our study, use of an inhibitor to block the PIM2 kinase activity upregulated AMPKα1 kinase activity,implying that PIM2 kinase could inhibit AMPKα1 function.

Further, PIM2 directly interacted with AMPKα1 and phosphorylated it on Thr467. Overall, our data suggest that
phosphorylation of AMPKα1 by PIM2 downregulates AMPKα1 activity and promotes cell proliferation in vivo and in vitro. This further serves to illuminate the mechan- isms by which PIM2 promotes tumorigenesis of EC. AMPKα1 is a tumor suppressor that plays critical reg-
ulatory roles in cancer cell growth and tumorigenesis, and this protein is often downregulated in a variety of human cancers [28]. For example, AMPKα1 decreases the growth of human colon cancer and hepatocellular carcinomas [29, 30]. Here, AMPKα1 kinase activity was upregulated by the PIM2 inhibitor SMI-4a in human EC cells, resulting in the inhibition of EC cell proliferation. AICAR, an activator of AMPKα1, could also effectively activate its kinase activity, which has previously been shown to inhibit cancerobtained from ATCC (Manassas, VA, USA). HEK293T cells were cultured in DMEM medium supple- mented with 10% FBS, and HEC-1B and Ishikawa cells were grown in DMEM/F12 supplemented with 10% FBS, 100 U/ml penicillin and 100 μg/ml Streptomycin. All cell lines were maintained at 37 °C in ahumidified incubator containing 5% CO2. PCR-amplified human genes used in this study were cloned into pcDNA3.0/HA, pFlag-CMV4, pEGFP-C1, pQCXIH, PET28a-His, or pGEX-4T-1. The mutants were generated by using overlap PCR.
Rabbit anti-PIM2 and mouse anti-phosphoserine anti- bodies were purchased from Abcam. Mouse anti-AMPKα1 antibody was from Novus. Rabbit anti-phosphothreonine and anti-pT172 AMPKα antibodies were obtained from Cell Signaling. Hygromycin B, rabbit, or mouse anti-HA,
-GFP, -Flag, or β-actin antibodies were from Sigma. Rabbit anti-pT467-AMPKα1 antibody was generated by a small peptide (RKNPVpTSTYSKC). Puromycin was from GBICO. Rabbit IgG and mouse IgG were from Santa Cruz Biotechnology. SMI-4a was obtained from Calbiochem. AICAR was obtained from Medchem express (MCE).

Immunoprecipitation and western blotting

The cells were lysed in 0.5% NP40 buffer containing 150 mM NaCl, 50 mM Tris-HCl, pH 7.5 and multiple protease inhibitors (Sigma-Aldrich). The IP assay and western blotting were performed as described previously [9].

GST pull-down assay

BL21 (DE3) were used to express GST-tagged AMPKα1 and His-tagged PIM2 (WT or K61A). The fusion proteins were purified with Glutathione Sepharose 4B beads (GEHealthcare) or Ni affinity resins (GE Healthcare) according to manufacturer’s instructions. GST pull-down assay was performed as described previously [9].

Confocal immunofluorescence microscopy
Ishikawa cells were cotransfected HA-tagged AMPKα1 and GFP-tagged PIM2. After 2 days of transfection, confocal immunofluorescence microscopy was performed as described previously [9].

In vitro kinase assay
The assay was performed as described previously [9].

AMPKα1 kinase activity

AMPKα1 proteins were accumulated by Co-IP assays. The enriched proteins were washed five times by PBS, and AMPKα1 kinase activity was measured by CycLex® AMPK Kinase Assay Kit (MBL, Japan) according to
manufacturing instructions.

Glucose consumption, lactate production, and ATP production

The assay was performed as described previously [9]. ATP production was measured by ATP Colorimetric/Fluorometric Assay Kit (Sigma) according to manufacturing instructions.

Stable cell lines

AMPKα1 siRNA1 [32] was 5′-GAGGAGAGCUAUUUGA UUA-3′. AMPKα1 siRNA2 [33] was 5′-GCUUGAU
GCACACAUGAAU-3′. The PIM2 shRNA [6] was gener- ated with oligonucleotide 5′-CTCGAAGTCGCACTGCT AT-3. The AMPKα1 shRNA [32] was generated with oli- gonucleotide 5′-GAGGAGAGCTATTTGATTA-3′, and the plasmids were constructed by GenePharma company. The
shRNA plasmids were cotransfected with vectors expres- sing gag and vsvg genes into HEK293T cells. The viruses were harvested and applied to EC cells. The knockdown cells were selected with puromycin for more than 2 weeks.
AMPKα1 (WT, T467A or T467D) was cloned into the retroviral vector (pQCXIH), and 5′-GTGGTGAACTT TTCGACTA-3′ was also included to generate a mismatch with the shRNA to prevent retargeting of the modified locus. After retroviral production, shAMPKα1 stable cells
were infected with the indicated virus for 2 days and screened using hygromycin for at least 2 weeks.

CCK-8, clone formation, cell invasion, and wound- healing assays

The assays were performed as described previously [9, 15, 34].

Immunohistochemistry (IHC)

The study was approved by the Human Assurance Com- mittee of Affiliated Hospital of Weifang Medical Uni- versity. Tissues were probed using human PIM2 and pT467-AMPKα1 antibodies. In addition, IHC for Ki67 in
mouse tumors from each group was performed. The
immunostained tissues were scored by the intensity (0–3) and the percentage of cells with score of 0 (0–5%), 1 (6–25%), 2 (26–50%), 3 (51–75%) and 4 (76–100%). The
staining grade was stratified as absent (score 0), weak (score 1–3), moderate (score 4–8) or strong (score 9–12). The cutoff value for the low expression is set to <4.

xenografts

A total of 1 × 107 stable indicated expressing Ishikawa cells were injected subcutaneously into the shoulder sides of BALB/c nude mice (4-week-old female) with matrigel gel. Tumor volumes were measured during the tumor growth for
3 or 4 weeks. Tumor volume was calculated using the following formula: volume = (length × width2)/2. After 3 or 4 weeks, the mice were killed, and tumors were weighed prior to further histological evaluation. All protocols involving live mice were approved by the Animal Care and Use Committee of Weifang Medical University.
For the inhibitor group, Ishikawa cells (1 × 107 each) were injected into the flanks of BALB/c nude mice (4- week-old female) with matrigel gel. After 1 week, mice were randomly subjected to control, SMI-4a (60 mg/kg), AICAR (500 mg/kg) or a combination of the drugs given intraperitoneally every 2 days at the same dose. After 2 weeks, the tumors were obtained and removed. Tumor volume was calculated as volume = (length × width2)/2.

Statistical analysis

The data analysis was performed by using statistical program SPSS software version 17.0 (Chicago, IL). Pearson correlation analysis was used to evaluate the relationship between two variables. Two-tailed t-test was used to analyze the difference between two groups. P values < 0.05 were regarded as statistically significant.

Acknowledgements The study was supported by research grants from Innovation fund of National Natural Science Foundation of China (Grant no. 81602440, 81602301, 81501275 and 81471048), Shandong
Province College Science and Technology Plan Project (Grant no. J16LL08 and J17KA254), Projects of medical and health technology development program in Shandong province (Grant no. 2016WS0688 and 2017WS398).

Author contributions ZY designed research. XH, CR, TY, QP, and LW performed research. AJ, YM, YD, and ZL contributed new reagents/analytic tools. ZY and XH analyzed data. ZY wrote and revised the paper.

Compliance with ethical standards

Conflict of interest The authors declare that they have no conflict of interest.

Publisher’s note: Springer Nature remains neutral with regard to jurisdictional claims in published maps and institutional affiliations.

References

1. Warfel NA, Kraft AS. PIM kinase (and Akt) biology and signaling in tumors. Pharmacol Ther. 2015;151:41–49.
2. Narlik-Grassow M, Blanco-Aparicio C, Carnero A. The PIM
family of serine/threonine kinases in cancer. Med Res Rev. 2014;34:136–59.
3. Hospital MA, Jacquel A, Mazed F, Saland E, Larrue C, Mondesir
J, et al. RSK2 is a new Pim2 target with pro-survival functions in FLT3-ITD-positive acute myeloid leukemia. Leukemia. 2018;32:597–605.
4. Wang Z, Zhang Y, Gu JJ, Davitt C, Reeves R, Magnuson NS.
Pim-2 phosphorylation ofp21(Cip1/WAF1) enhances its stability and inhibits cell proliferation in HCT116 cells. Int J Biochem Cell Biol. 2010;42:1030–8.
5. Morishita D, Katayama R, Sekimizu K, Tsuruo T, Fujita N. Pim
kinases promote cell cycle progression by phosphorylating and down-regulating p27Kip1 at the SMI-4a transcriptional and post- transcriptional levels. Cancer Res. 2008;68:5076–85.
6. Lu J, Zavorotinskaya T, Dai Y, Niu XH, Castillo J, Sim J, et al. Pim2
is required for maintaining multiple myeloma cell growth through modulating TSC2 phosphorylation. Blood. 2013;122:1610–20.
7. Yu Z, Zhao X, Huang L, Zhang T, Yang F, Xie L, et al. Proviral
insertion in murine lymphomas 2 (PIM2) oncogene phosphor- ylates pyruvate kinase M2 (PKM2) and promotes glycolysis in cancer cells. J Biol Chem. 2013;288:35406–16.
8. Yu Z, Huang L, Qiao P, Jiang A, Wang L, Yang T, et al. PKM2
Thr454 phosphorylation increases its nuclear translocation and promotes xenograft tumor growth in A549 human lung cancer cells. Biochem Biophys Res Commun. 2016;473:953–8.
9. Yang T, Ren C, Qiao P, Han X, Wang L, Lv S, et al. PIM2-
mediated phosphorylation of hexokinase 2 is critical for tumor growth and paclitaxel resistance in breast cancer. Oncogene. 2018;37:5997–6009.
10. Nair JR, Caserta J, Belko K, Howell T, Fetterly G, Baldino C,
et al. Novel inhibition of PIM2 kinase has significant anti-tumor efficacy in multiple myeloma. Leukemia. 2017;31:1715–26.
11. Zhao YQ, Yin YQ, Liu J, Wang GH, Huang J, Zhu LJ, et al.
Characterization of HJ-PI01 as a novel Pim-2 inhibitor that induces apoptosis and autophagic cell death in triple-negative human breast cancer. Acta Pharmacol Sin. 2016;37:1237–50.
12. Kreuz S, Holmes KB, Tooze RM, Lefevre PF. Loss of PIM2
enhances the anti-proliferative effect of the pan-PIM kinase
inhibitor AZD1208 in non-Hodgkin lymphomas. Mol Cancer. 2015;14:205.
13. Daenthanasanmak A, Wu Y, Iamsawat S, Nguyen HD, Bastian D, Zhang M, et al. PIM-2 protein kinase negatively regulates T cell responses in transplantation and tumor immunity. J Clin Investig. 2018;128:2787–2801.
14. Yu Z, Zhao X, Ge Y, Zhang T, Huang L, Zhou X, et al. A
regulatory feedback loop between HIF-1alpha and PIM2 in HepG2 cells. PLoS ONE. 2014;9:e88301.
15. Ren C, Yang T, Qiao P, Wang L, Han X, Lv S, et al. PIM2 interacts with tristetraprolin and promotes breast cancer tumor- igenesis. Mol Oncol. 2018;12:690–704.
16. Onstad MA, Schmandt RE, Lu KH. Addressing the Role of
Obesity in Endometrial Cancer Risk, Prevention, and Treatment. J Clin Oncol. 2016;34:4225–30.
17. Ye S, Wen H, Jiang Z, Wu X. The effect of visceral obesity on
clinicopathological features in patients with endometrial cancer: a retrospective analysis of 200 Chinese patients. BMC Cancer. 2016;16:209.
18. Herzig S, Shaw RJ. AMPK: guardian of metabolism and mito- chondrial homeostasis. Nat Rev Mol Cell Biol. 2018;19:121–35.
19. Jansen M, Ten Klooster JP, Offerhaus GJ, Clevers H. LKB1 and
AMPK family signaling: the intimate link between cell polarity and energy metabolism. Physiol Rev. 2009;89:777–98.
20. Garcia D, Shaw RJ. AMPK: mechanisms of cellular energy sensing and restoration of metabolic balance. Mol Cell. 2017;66:789–800.
21. Suzuki T, Bridges D, Nakada D, Skiniotis G, Morrison SJ, Lin JD,
et al. Inhibition of AMPK catabolic action by GSK3. Mol Cell. 2013;50:407–19.
22. Dagon Y, Hur E, Zheng B, Wellenstein K, Cantley LC, Kahn BB.
p70S6 kinase phosphorylates AMPK on serine 491 to mediate leptin’s effect on food intake. Cell Metab. 2012;16:104–12.
23. Hawley SA, Ross FA, Gowans GJ, Tibarewal P, Leslie NR,
Hardie DG. Phosphorylation by Akt within the ST loop of AMPK-alpha1 down-regulates its activation in tumour cells. Biochem J. 2014;459:275–87.
24. Beharry Z, Mahajan S, Zemskova M, Lin YW, Tholanikunnel
BG, Xia Z, et al. The Pim protein kinases regulate energy meta- bolism and cell growth. Proc Natl Acad Sci USA. 2011;108:528–33.
25. Hawley SA, Davison M, Woods A, Davies SP, Beri RK, Carling
D, et al. Characterization of the AMP-activated protein kinase kinase from rat liver and identification of threonine 172 as the major site at which it phosphorylates AMP-activated protein kinase. J Biol Chem. 1996;271:27879–87.
26. Lin SC, Hardie DG. AMPK: sensing glucose as well as cellular
energy status. Cell Metab. 2018;27:299–313.
27. Faubert B, Boily G, Izreig S, Griss T, Samborska B, Dong Z, et al.
AMPK is a negative regulator of the Warburg effect and sup- presses tumor growth in vivo. Cell Metab. 2013;17:113–24.
28. Hardie DG, Schaffer BE, Brunet A. AMPK: an energy-sensing
pathway with multiple inputs and outputs. Trends Cell Biol. 2016;26:190–201.
29. Shukla K, Sonowal H, Saxena A, Ramana KV, Srivastava SK.
Aldose reductase inhibitor, fidarestat regulates mitochondrial biogenesis via Nrf2/HO-1/AMPK pathway in colon cancer cells. Cancer Lett. 2017;411:57–63.
30. Zhou X, Chen J, Chen L, Feng X, Liu Z, Hu, et al. Negative
regulation of Sirtuin 1 by AMP-activated protein kinase promotes metformin-induced senescence in hepatocellular carcinoma xenografts. Cancer Lett. 2017;411:1–11.
31. Hall DT, Griss T, Ma JF, Sanchez BJ, Sadek J, Tremblay AMK,
et al. The AMPK agonist 5-aminoimidazole-4-carboxamide ribo- nucleotide (AICAR), but not metformin, prevents inflammation- associated cachectic muscle wasting. EMBO Mol Med. 2018;10: pii: e8307.
32. Bungard D, Fuerth BJ, Zeng PY, Faubert B, Maas NL, Viollet B, et al. Signaling kinase AMPK activates stress-promoted transcription via SMI-4a histone H2B phosphorylation. Science. 2010;329:1201–5.
33. Yuan J, Ng WH, Yap J, Chia B, Huang X, Wang M, et al. The
AMPK inhibitor overcomes the paradoxical effect of RAF
inhibitors through blocking phospho-Ser-621 in the C terminus of CRAF. J Biol Chem. 2018;293:14276–84.
34. Yu Z, Ge Y, Xie L, Zhang T, Huang L, Zhao X, et al. Using a
yeast two-hybrid system to identify FTCD as a new regulator for HIF-1alpha in HepG2 cells. Cell Signal. 2014;26:1560–6.