PRL-3, an oncogenic dual-specificity phosphatase, is overexpressed in 50% of acute myelogenous leukemia (AML) and associated with poor survival. We found that stable expression of PRL-3 confers cytokine independence and growth advantage of AML cells. However, how PRL-3 mediates these functions in AML is not known. To comprehensively screen for PRL3-regulated proteins in AML, we performed SILAC-based quantitative proteomics analysis and discovered 398 significantly perturbed proteins after PRL-3 overexpression. We show that Leo1, a component of RNA polymerase II–associated factor (PAF) complex, is a novel and important mediator of PRL-3 oncogenic activities in AML. We described a novel mechanism where elevated PRL-3 protein increases JMJD2C histone demethylase occupancy on Leo1 promoter, thereby reducing the H3K9me3 repressive signals and promoting Leo1 gene expression. Furthermore, PRL-3 and Leo1 levels were positively associated in AML patient samples (N = 24; P < 0.01). On the other hand, inhibition of Leo1 reverses PRL-3 oncogenic phenotypes in AML. Loss of Leo1 leads to destabilization of the PAF complex and downregulation of SOX2 and SOX4, potent oncogenes in myeloid transformation. In conclusion, we identify an important and novel mechanism by which PRL-3 mediates its oncogenic function in AML. Cancer Res; 74(11); 3043–53. ©2014 AACR.

PRL-3, encoded by the PTP4A3 gene, is a small dual-specificity phosphatase characterized by the conserved C(X5)R catalytic domain, and a unique C-terminal prenylation domain essential for its proper subcellular localization (1, 2). PRL-3 has been shown to promote cellular processes, such as cell motility, invasion, cell growth, and survival, through various mechanisms (2–7). PRL-3 was first linked to cancer when it was consistently found at elevated levels in colorectal cancer metastases, but at much lower levels in matched early-staged tumor and normal colorectal epithelium (8). Since then, elevated expression of PRL-3 has been implicated in the progression and metastasis of an array of cancer types, including gastric, ovarian, cervical, lung, liver, and breast, and PRL-3 protein was overexpressed in an average of 22.3% of 1,008 human carcinoma samples examined using immunohistochemistry (9, 10). Together with the fact that it has a highly restricted basal pattern of expression in adult tissues (11), PRL-3 is deemed as an attractive therapeutic target that spares normal tissues. The potential of this target has been demonstrated using PRL-3–specific antibodies in an in vivo model (12).

In recent years, accumulating evidence suggests that PRL-3 is also a novel therapeutic target and biomarker in leukemia (13, 14). We were the first to report that elevated PRL-3 protein expression occurs in about 47% of human acute myelogenous leukemia (AML) cases while absent from normal myeloid cells in bone marrow (13). In addition, a large-scale gene expression profiling study of 454 primary AML samples demonstrates that high PRL-3 levels is an independent negative prognostic factor in AML, both for overall survival and event-free survival (14). These reports collectively suggest that PRL-3 may be of biologic and clinical relevance in AML and warrants further investigation.

In the present study, we created a TF-1 AML cell line stably overexpressing PRL-3 (TF1-hPRL3) as a model to study the biologic relevance of PRL-3 in AML, and we use a quantitative proteomic strategy to profile on a global scale changes in protein expression induced by PRL-3 in AML. We found that PRL-3 has pro-oncogenic properties in AML, and Leo1, a component of the human RNA polymerase II–associated factor (PAF) complex, is one of the most differentially expressed proteins induced by PRL-3. Further, we showed that PRL-3 upregulates Leo1 through a novel epigenetic mechanism. On the other hand, knockdown of Leo1 significantly diminishes the oncogenic effects of PRL-3. Our current work implicates a pro-oncogenic role of PRL-3 in AML, and reveals Leo1 as a novel downstream molecule required for PRL-3 oncogenic function in leukemia.

Cell culture

HEK293T cells were cultured in Dulbecco's Modified Eagle Medium with 10% fetal calf serum (FCS). TF1-derived cell lines were cultured in RPMI 1640 medium containing 10% FCS (R10) supplemented with 5 ng/mL human interleukin (IL)-3 (Miltenyi Biotec). Molm-14, HEL, and HL-60 were cultured in R10. Human CD34+ cells were grown in StemSpan SFEM II medium supplemented with StemSpan CC100 cytokine cocktail (StemCell Technologies). Primary AML cells were grown in same conditions, with the addition of granulocyte macrophage colony-stimulating factor (20 ng/mL). Cell lines were obtained from American Type Culture Collection and authenticated. Plasmid details are available in Supplementary Methods.

SILAC-based mass spectrometry

TF-1 was cultured in “light” stable isotope labeling by amino acids in cell culture (SILAC) medium containing normal lysine and arginine amino acids, whereas the TF1-hPRL3 was grown in “heavy” SILAC medium with stable isotope-labeled 13C6 lysine (+6-Da shift) and 13C615N4 arginine (+10-Da shift; Thermo Scientific). The cellular lysates were combined and proteolytically digested by trypsin followed by tandem mass spectrometry (MS) identification. Peptides were subjected to target-decoy database search strategy with a false discovery rate (FDR) of less than 1% (<1%). Differential protein expression was quantified from the relative intensity ratios in the MS spectra between the “heavy” and “light” states.

Xenograft mode models

Six-week-old female nonobese diabetic/severe combined immunodeficient (NOD/SCID) mice were provided by Dr. Chan Shing-Leng (CSI Singapore, NUS). Exponentially growing TF1-pEGFP and TF1-hPRL3 cells (5 × 106) were subcutaneously injected into loose skin between the shoulder blades and the left and right front leg of NOD/SCID-recipient mice (3 mice total), respectively. The length (L) and width (W) of the tumor were measured with callipers, and tumor volume (TV) was calculated as TV = (L × W2)/2. The protocol is reviewed and approved by Institutional Animal Care and Use Committee in compliance to the guidelines on the care and use of animals for scientific purpose.

Transfection and lentiviral-mediated shRNA delivery

High-titre lentiviruses were produced in HEK293T cells by transfection using JetPrime (Polyplus) with packaging vectors psPAX2 and pMD2.G together with the respective short hairpin RNAs (shRNA) constructs. Virus was harvested, pooled, and concentrated by centrifugation (Amicon). Infection of leukemia cells was done by the spin-infection method.

RNA extraction and quantitative reverse transcription PCR

Detailed protocol is available in Supplementary Methods.

Western blotting

Cells were counted and lysed in radioimmunoprecipitation assay buffer. Anti-PRL3 (clone 318) was a kind gift from Dr. Zeng Qi [Institute of Molecular and Cell Biology, Agency for Science, Technology, and Research (A*STAR), Singapore]; anti-Leo1, anti-Paf1, and anti-Ctr9 antibodies were from Bethyl Laboratories; anti-actin, anti-hnRNPE1, anti-HSP90, anti-JMJD2C, anti-GFP, and anti-c-Myc (9E10) antibodies were from Santa Cruz Biotechnology; anti-Stathmin, anti-HDAC2 and anti-H3, anti-H3K9me2, anti-H3K27me2, anti-H3K27me3, anti-H3K4me1, anti-H3K4me2, anti-H3K4me3, and anti-H3K79me2 antibodies were from Cell Signaling Technology; anti-H3K9me3 antibody was from Active Motif.

Chromatin immunoprecipitation

Chromatin immunoprecipitation (ChIP) assays were performed using the respective antibodies according to the manufacturer's instructions (Thermo Scientific). Quantitative PCR (qPCR) was performed using the eluted DNA (sample) and 1% input, with primers spanning Leo1 promoter region with respect to Leo1 transcriptional start site. The percentage of input was calculated as described previously (15). Primer sequences used for ChIP–qPCR are available in Supplementary Table S3.

Luciferase assay

Detailed protocol is available in Supplementary Methods.

Cell proliferation, apoptosis, and CFA

Detailed protocol is available in Supplementary Methods.

PRL-3 promotes cytokine-independent growth, colony-forming capacity of AML cells in vitro and tumorigenecity in vivo

To assess the roles of PRL-3 in pathogenesis of AML, we developed a pair of stable, isogenic cell lines, TF1-pEGFP and TF1-hPRL3 by transfecting pEGFP (vector control) and pEGFP-hPRL-3 vectors into TF-1 cells, respectively, followed by G418 selection and fluorescence-activated cell sorting (Fig. 1A). TF-1 is a cytokine-dependent AML cell line. Quantitative reverse transcription PCR (qRT-PCR) and Western blot validated the overexpression of PRL-3 on both mRNA and protein levels in the TF1-hPRL3 cells relative to TF1-pEGFP cells (Fig. 1B). In the absence of cytokine (human IL-3), the majority of TF1-pEGFP cells became apoptotic after 72 hours. In contrast, most of TF1-hPRL3 cells were viable (Fig. 1C). TF1-hPRL3 cells not only resisted cytokine deprivation-induced apoptosis, but also proliferated well without additional cytokines (Fig. 1C). Furthermore, methylcellulose assay showed only TF1-hPRL3 cells, but not TF1-pEGFP cells, formed colonies (Fig. 1D). To determine the tumorigenic role of PRL-3 in vivo, we used the pair of TF1 cells in a subcutaneous mouse xenograft model. Strikingly, only TF1-hPRL3 cells, but not TF1-pEGFP cells, formed tumor mass in immunodeficient mice (Fig. 1E). The average TF1-hPRL3 tumor volume was 813 mm3 at 4 weeks after cell inoculation (Fig. 1E). Overall, these data suggest that PRL-3 confers important oncogenic function in AML cells in vitro and in vivo.

Figure 1.

Establishment of stable PRL-3 cell line. A, images of TF1-pEGFP and TF1-hPRL3 cells in bright field and GFP channel under an invert microscopy. B, qRT-PCR and Western blot analysis of PRL-3 mRNA and protein expression in TF1-pEGFP and TF1-hPRL3 cells. β-Actin was used as a loading control in the protein analysis. C, after IL-3 starvation, TF1-pEGFP and TF1-hPRL3 cells were stained and assessed with Trypan Blue Exclusion methods on a hemacytometer. The percentages of viable cells at 72 hours were calculated relative to before IL-3 withdrawal (0 hour). To construct the cell growth curve, 20 × 105 TF1-hPRL3 cells were initially seeded in standard medium without additional human cytokines. Cells were counted and reseeded in fresh medium every 2 days for up to 6 days. In both figures, three independent replicates were conducted (mean ± SD). D, colony-forming assay of TF1-pEGFP and TF1-hPRL3. The experiments were duplicated and representative pictures are presented. E, mouse xenograft models of TF1-pEGFP and TF1-hPRL3. Five million TF1-pEGFP and TF1-hPRL3 cells were subcutaneously injected into loose skin between the shoulder blades and the left and right front leg of NOD/SCID recipient mice, respectively. At 4 weeks after injection, tumors were harvested and measurement of tumor volume was taken. The tumor volume is shown as the mean ± SD (mm3). Arrows, the TF1-hPRL3 tumor mass at right and no tumor at left (TF1-pEGFP cell injected site) in one mouse. **, P < 0.005.

Figure 1.

Establishment of stable PRL-3 cell line. A, images of TF1-pEGFP and TF1-hPRL3 cells in bright field and GFP channel under an invert microscopy. B, qRT-PCR and Western blot analysis of PRL-3 mRNA and protein expression in TF1-pEGFP and TF1-hPRL3 cells. β-Actin was used as a loading control in the protein analysis. C, after IL-3 starvation, TF1-pEGFP and TF1-hPRL3 cells were stained and assessed with Trypan Blue Exclusion methods on a hemacytometer. The percentages of viable cells at 72 hours were calculated relative to before IL-3 withdrawal (0 hour). To construct the cell growth curve, 20 × 105 TF1-hPRL3 cells were initially seeded in standard medium without additional human cytokines. Cells were counted and reseeded in fresh medium every 2 days for up to 6 days. In both figures, three independent replicates were conducted (mean ± SD). D, colony-forming assay of TF1-pEGFP and TF1-hPRL3. The experiments were duplicated and representative pictures are presented. E, mouse xenograft models of TF1-pEGFP and TF1-hPRL3. Five million TF1-pEGFP and TF1-hPRL3 cells were subcutaneously injected into loose skin between the shoulder blades and the left and right front leg of NOD/SCID recipient mice, respectively. At 4 weeks after injection, tumors were harvested and measurement of tumor volume was taken. The tumor volume is shown as the mean ± SD (mm3). Arrows, the TF1-hPRL3 tumor mass at right and no tumor at left (TF1-pEGFP cell injected site) in one mouse. **, P < 0.005.

Close modal

SILAC quantitative proteomic profiling of PRL-3 overexpressed cell

To understand downstream pathways induced by overexpressing PRL-3, we performed a quantitative proteomic study by combining metabolic labeling using SILAC and high-resolution MS approach (Supplementary Fig. S1A). We identified 398 proteins whose expression was significantly perturbed by PRL-3 overexpression, with 331 proteins upregulated and 67 proteins downregulated (>1.5-fold; Supplementary Fig. S1B).

To gain a functional understanding of the dataset, we subjected the differentially expressed proteins (up- or downregulated) to gene ontology analysis. Upregulated proteins are enriched for processes such as nucleic acid metabolism (19.7%), protein metabolism (15.1%), and pre-mRNA processing (5.7%), which are novel biologic processes associated with PRL-3, as well as known involvement in cell-cycle regulation (7.1%) and cell motility (6%) as depicted in Supplementary Fig. S1C. Amongst the differentially regulated proteins, 45.5% of them are nuclear proteins, whereas 54.5% are cytoplasmic proteins (Supplementary Fig. S1D).

Validation of MS results

Expression of several representative proteins, including Leo1, stathmin, HSP-90, hnRNPE1, and HDAC2, was compared between TF1-pEGFP and TF1-hPRL3 cells (Fig. 2A), and concordance was observed between the Western blot and MS results. Unexpectedly, we did not detect PRL-3 in our liquid chromatography/tandem mass spectrometry (LC/MS-MS) result. This is likely due to the trypsin cleavage sites distribution in the PRL-3 that generates peptides either too long or too short for detection by LC-MS/MS. Nonetheless, PRL-3 overexpression was confirmed at both the mRNA (data not shown) and protein levels by Western blot (Fig. 2A).

Figure 2.

Western blot validation of candidate proteins identified by MS. A, expression of several representative upregulated proteins is compared between TF-1 and TF1-hPRL3. B, knockdown of endogenous PRL-3 in AML cell lines Molm-14 and HEL. The same panel of proteins is examined by Western blot. C, 293T cells were stably transfected with an inducible Tet-on myc-tagged PRL-3 construct, and increasing concentration of doxycyclin was used to induce PRL-3 expression. D, 293T cells were transfected with specified doses of PRL-3 siRNA for 48 hours before analysis with Western blot.

Figure 2.

Western blot validation of candidate proteins identified by MS. A, expression of several representative upregulated proteins is compared between TF-1 and TF1-hPRL3. B, knockdown of endogenous PRL-3 in AML cell lines Molm-14 and HEL. The same panel of proteins is examined by Western blot. C, 293T cells were stably transfected with an inducible Tet-on myc-tagged PRL-3 construct, and increasing concentration of doxycyclin was used to induce PRL-3 expression. D, 293T cells were transfected with specified doses of PRL-3 siRNA for 48 hours before analysis with Western blot.

Close modal

We further performed lentiviral-based delivery of PRL-3 shRNA into two additional human AML cells, Molm-14 and HEL, which expresses high endogenous level of PRL-3. The knockdown of PRL-3 reduced the levels of Leo1, stathmin, HSP90, and hnRNPE1 (Fig. 2B). An exception was HDAC2, where although MS identifies it as one of the upregulated proteins by PRL-3, Western blot was not sensitive enough to detect the differences (Fig. 2B). These experiments confirmed the general validity of our proteomic analysis.

Leo1 is upregulated by PRL-3

Leo1, a component of the human PAF complex, was the top candidate from the SILAC LC-MS/MS analysis with the most significant increase in protein abundance. We therefore decided to focus on this protein in our subsequent experiments. To confirm that Leo1 is a downstream target of PRL-3, we introduced a Tet-On myc-tagged PRL-3 construct into 293T cells. Forty-eight hours after transfection, 293T cells were exposed to increasing concentrations of the tetracycline derivative, doxycycline, to induce PRL-3 expression. PRL-3 protein accumulated in a dose-dependent manner, and a corresponding increase in Leo1 protein levels was observed (Fig. 2C) and quantified (Supplementary Fig. S2). To complement the overexpression studies, we also performed knockdown by transfecting PRL-3 siRNA into 293T cells. The degree of PRL-3 inhibition correlated to the dose of siRNA added, and Leo1 was downregulated (Fig. 2D). Collectively, these observations corroborated the MS results showing that Leo1 expression is positively correlated with PRL-3 levels.

Abrogation of Leo1 reduces PRL3-mediated proliferation and cytokine independence of AML

To determine whether Leo1 contributes to PRL3-induced leukemic growth, we used lentiviral-based delivery of shRNAs to deplete Leo1 expression and looked at growth properties. qRT-PCR and immunoblotting analysis demonstrated a 5-fold reduction in Leo1 mRNA levels and a marked reduction in the Leo1 protein levels upon treatment with Leo1 shRNA (Fig. 3A). Notably, reduction in Leo1 levels significantly impaired growth of TF1-hPRL3 cells to growth rates comparable with TF-1 parental cells (Fig. 3B).

Figure 3.

Effects of stable Leo1 knockdown in TF1-hPRL3 cells. A, knockdown of Leo1 in TF1-hPRL3 cells were analyzed with qRT-PCR and Western blot. β-Actin was used as loading control. B, growth curve was determined by CellTitre-Glo luminescent assay. Cells were plated in 96-well format and grown for up to 4 days. Experiments were performed in triplicates. C, TF-1, PRL3-scrambled, and PRL3-Leo1 shRNA cells were seeded in medium without human IL-3 for 48 hours. The graph is plotted with the percentages of apoptotic cells defined by Annexin V and 7-AAD positivity. D, TF-1, PRL3-scrambled, and PRL3-Leo1 shRNA cells were evenly suspended in HSC-colony-forming unit medium (StemMACS) and incubated for 10 days for colonies to develop. **, P < 0.005.

Figure 3.

Effects of stable Leo1 knockdown in TF1-hPRL3 cells. A, knockdown of Leo1 in TF1-hPRL3 cells were analyzed with qRT-PCR and Western blot. β-Actin was used as loading control. B, growth curve was determined by CellTitre-Glo luminescent assay. Cells were plated in 96-well format and grown for up to 4 days. Experiments were performed in triplicates. C, TF-1, PRL3-scrambled, and PRL3-Leo1 shRNA cells were seeded in medium without human IL-3 for 48 hours. The graph is plotted with the percentages of apoptotic cells defined by Annexin V and 7-AAD positivity. D, TF-1, PRL3-scrambled, and PRL3-Leo1 shRNA cells were evenly suspended in HSC-colony-forming unit medium (StemMACS) and incubated for 10 days for colonies to develop. **, P < 0.005.

Close modal

To determine whether loss of Leo1 affects the survival of TF1-hPRL3, we plated the cells in the absence of human IL-3 for 72 hours, followed by flow cytometric analysis for apoptotic markers Annexin V-phycoerythrin and 7-aminoactinomycin D (7-AAD). Although approximately 90% of TF1-hPRL3 population was able to survive without supplementation of cytokine, Leo1 knockdown notably increased the apoptosis of TF1-hPRL3 induced by cytokine withdrawal (Fig. 3C), indicating that Leo1 abrogation removes the protective effect of PRL-3 toward cytokine deprivation. Leo1 knockdown in PRL-3 cells resulted in significantly fewer colonies, indicating a loss of clonogenic survival of these cells (Fig. 3D). These findings collectively indicate that Leo1 is an important mediator of the cell growth and survival of the TF1-hPRL3 cells, and loss of Leo1 is largely sufficient to reverse these properties conferred by PRL-3 overexpression.

Leo1 and PRL-3 levels are significantly associated in AML patient samples

Because we observed a strong association between Leo1 and PRL-3 levels in cell lines, we determined whether Leo1 is similarly upregulated in human primary AML cells. Western blot analysis of PRL-3 and Leo1 protein expression in an independent cohort of 24 primary AML samples found that 50% of the patient samples (Fig. 4A) were positive for PRL-3 protein. Of these, only one did not express Leo1. Therefore, a strong association exists between PRL-3 and Leo1 expression (χ2P value < 0.01). Leo1 was also expressed in five samples with no PRL-3 expression, suggesting that other mechanisms may be responsible in these cases. Our findings demonstrated that a significant proportion (50%) of AML cases seemed to coexpress PRL-3 and Leo1, suggesting that Leo1 is a likely downstream target of PRL-3.

Figure 4.

Leo1 and PRL-3 expression in human primary AML cells. A, Western blot was performed to show the expression levels of Leo1 and PRL-3 in a cohort of 24 primary AML samples. β-Actin was used as loading control. χ2 test indicates highly significant association (P < 0.01). B, analysis of PRL-3 mRNA expression in primary AML cells transfected with either control siRNA or PRL-3 siRNA for 24 hours. C, cell proliferation assay of primary AML cells transfected with either control siRNA or PRL-3 siRNA for 3 days. **, P < 0.01

Figure 4.

Leo1 and PRL-3 expression in human primary AML cells. A, Western blot was performed to show the expression levels of Leo1 and PRL-3 in a cohort of 24 primary AML samples. β-Actin was used as loading control. χ2 test indicates highly significant association (P < 0.01). B, analysis of PRL-3 mRNA expression in primary AML cells transfected with either control siRNA or PRL-3 siRNA for 24 hours. C, cell proliferation assay of primary AML cells transfected with either control siRNA or PRL-3 siRNA for 3 days. **, P < 0.01

Close modal

We also examined the basal expression levels of human PRL-3 and Leo1 in normal hematopoietic cell fractions, namely CD34+, T cell, B cell, and myeloid cell. Similarly, we found a relative correlation between PRL-3 and Leo1 levels. Furthermore, we observed that PRL-3 and Leo1 are expressed at higher levels in CD34+ and myeloid cells (Supplementary Fig. S3).

To validate the functional relevance of PRL-3 in human AML, we downregulate PRL-3 in primary AML cells (Fig. 4B). PRL-3 siRNA inhibited AML cell growth in a proliferation assay as compared with transfection with control siRNA (Fig. 4C). This result suggests that PRL-3 expression is also of biologic relevance in human AML and specific targeting of PRL-3 may have possible therapeutic applications in AML.

PRL-3 affects Leo1 mRNA levels through histone H3 lysine 9 (H3K9me3) demethylation

To examine whether PRL-3 regulates Leo1 at the transcriptional level, we perturbed PRL-3 with different shRNA constructs, which was accompanied by a concomitant reduction of Leo1 transcripts levels (Fig. 5A). To confirm that PRL-3 regulates Leo1 at the transcriptional level, we established a Leo1 reporter construct containing a 703-bp Leo1 promoter fragment cloned upstream of the firefly luciferase gene (pGL3-Leo1). Transfection of the reporter construct, together with vector expressing wild-type PRL-3, increased the luciferase activity in a dose-dependent manner (Fig. 5B).

Figure 5.

PRL-3 regulates Leo1 at the transcriptional level. A, Molm-14 cells were transfected with PRL-3 shRNA for 24 hours, and Leo1 and PRL-3 mRNA levels were measured. Data, mean ± SEM of triplicates. B, 293T cells were transfected with the respective luciferase reporter and expression vectors, and relative luciferase activity (Firefly/Renilla) was determined. Data, mean ± SEM of triplicates. C, nuclear extracts were obtained from TF1 or TF1-hPRL3 cells and was immunoblotted for respective histone marks. Histone H3 was used as the loading control. Intensity of bands was measured using ImageJ and provided as a fold increase using TF-1 as control. D, diagram of Leo1 promoter showing the position of four different primer pairs (P1–P4) used for qPCR. ChIP was performed in TF1 and TF1-hPRL3 cells with respective antibodies. Solid lines, ChIPs with the indicated antibodies; dotted lines, control rabbit IgG; TF1 cells, black; TF1-hPRL3 cells, gray. Data are plotted as percentage of input. Data, mean ± SEM of triplicates. *, P < 0.05; **, P < 0.005.

Figure 5.

PRL-3 regulates Leo1 at the transcriptional level. A, Molm-14 cells were transfected with PRL-3 shRNA for 24 hours, and Leo1 and PRL-3 mRNA levels were measured. Data, mean ± SEM of triplicates. B, 293T cells were transfected with the respective luciferase reporter and expression vectors, and relative luciferase activity (Firefly/Renilla) was determined. Data, mean ± SEM of triplicates. C, nuclear extracts were obtained from TF1 or TF1-hPRL3 cells and was immunoblotted for respective histone marks. Histone H3 was used as the loading control. Intensity of bands was measured using ImageJ and provided as a fold increase using TF-1 as control. D, diagram of Leo1 promoter showing the position of four different primer pairs (P1–P4) used for qPCR. ChIP was performed in TF1 and TF1-hPRL3 cells with respective antibodies. Solid lines, ChIPs with the indicated antibodies; dotted lines, control rabbit IgG; TF1 cells, black; TF1-hPRL3 cells, gray. Data are plotted as percentage of input. Data, mean ± SEM of triplicates. *, P < 0.05; **, P < 0.005.

Close modal

Given that PRL-3 is not a DNA-binding protein, the transcriptional involvement of PRL-3 is likely to be indirect. A recent report implicated an epigenetic role of PRL-3 in colorectal cancer through regulating histone-modifying enzymes involved in histone demethylation (16). To explore the possibility that PRL-3 activates Leo1 through histone methylation status in leukemic cells, we performed an unbiased screen of eight different permissive and repressive histone modifications in the TF-1 and TF1-hPRL3 cells. We found that PRL-3 overexpression specifically leads to a global decrease in di- and trimethylation at histone H3 lysine 9 (H3K9me2 and 3) residues (Fig. 5C). We further confirmed this observation in another AML cell line, Molm-14 (Supplementary Fig. S4).

To determine whether increased Leo1 expression is associated with local reductions in repressive histone methylation, we compared the basal levels of H3K9me2, H3K9me3, and H3K27me3 across Leo1 promoter region in the TF-1 and TF1-hPRL3 cells. In TF1-hPRL3 cells, a marked reduction in H3K9me3 and, to a lesser extent, in H3K9me2 was detected throughout the entire promoter region (P1–P4) as compared with the TF-1 cells (Fig. 5D). In contrast, H3K27me3 levels were very low on the promoter (Supplementary Fig. S5). Thus, Leo1 promoter seems to be predominantly modulated by the methylation levels of H3K9.

PRL-3 mediates H3K9me3 demethylation on Leo1 promoter through JMJD2C histone demethylase

Next, we sought to identify which histone demethylase is involved in mediating this process. The members of the JMJD2 family, including the JMJD2A, JMJD2B, JMJD2C, and JMJD2D, are known to exhibit specific lysine-trimethyl demethylation activities (17). We systematically knocked down these candidates and examined the impact on H3K9me3 levels and Leo1 expression. The JMJD2 shRNAs exhibited high specificity (Fig. 6A). We observed that JMJD2C depletion in TF1-hPRL3 cells was able to increase the global levels of H3K9me3, accompanied by a reduction in Leo1 (Fig. 6B), implying the specific involvement of JMJD2C in regulating H3K9me3 related to levels of Leo1.

Figure 6.

JMJD2C regulates H3K9me3 on the Leo1 promoter. A, shRNA specificity was tested in TF-1 cells using JMJD2A, 2B, 2C, and 2D primers. Graph is plotted relative to scrambled (Mock). Data, mean ± SEM of triplicates. B, TF1-hPRL3 transfected with Mock or JMJD2 shRNA and analyzed by Western blot using the indicated antibodies. C and D, ChIP was performed using antibody against JMJD2C (C) or H3K9me3 (D) on Leo1 promoter P2. E, ChIP was performed in TF-1 and TF1-hPRL3 cells using JMJD2C antibody on Leo1 promoter. Data shown as percentage of input. Data, mean ± SEM of triplicates. *, P < 0.05; **, P < 0.005.

Figure 6.

JMJD2C regulates H3K9me3 on the Leo1 promoter. A, shRNA specificity was tested in TF-1 cells using JMJD2A, 2B, 2C, and 2D primers. Graph is plotted relative to scrambled (Mock). Data, mean ± SEM of triplicates. B, TF1-hPRL3 transfected with Mock or JMJD2 shRNA and analyzed by Western blot using the indicated antibodies. C and D, ChIP was performed using antibody against JMJD2C (C) or H3K9me3 (D) on Leo1 promoter P2. E, ChIP was performed in TF-1 and TF1-hPRL3 cells using JMJD2C antibody on Leo1 promoter. Data shown as percentage of input. Data, mean ± SEM of triplicates. *, P < 0.05; **, P < 0.005.

Close modal

To demonstrate that JMJD2C directly regulates H3K9me3 on the Leo1 promoter, we examined the levels of H3K9me3 on region P2 on Leo1 promoter. Accordingly, knockdown of JMJD2C reduced the JMJD2C signal on the Leo1 promoter (Fig. 6C) and led to a 40% augmentation of the H3K9me3 signal (Fig. 6D). However, knockdown of JMJD2A, JMJD2B, and JMJD2D did not modulate H3K9me3 levels on Leo1 promoter (Supplementary Fig. S6A). Next, we examined how PRL-3 phosphatase regulates JMJD2C histone demethylase. JMJD2C protein expression did not increase in TF1-hPRL3 cells (Supplementary Fig. S6B). We also examined the gene expression levels of JMJD2A, JMJD2B, JMJD2C, and JMJD2D in TF1 and TF1-PRL3 cells, and found no significant changes between the isogenic cell lines (Supplementary Fig. S6C). Furthermore, PRL-3 and JMJD2C did not physically interact (Supplementary Fig. S6D). However, an increase in JMJD2C occupancy occurred on the promoter of Leo1 in the TF1-hPRL3 as compared with TF-1 cells (Fig. 6E). JMJD2C association to the P2 region was significantly abrogated in the PRL-3 knockdown cells, supporting the notion that JMJD2C binding to the Leo1 promoter region is dependent on the presence of the PRL-3 protein (Supplementary Fig. S6E).

Taken together, these results demonstrate that PRL-3 relieves the repressive H3K9me3 mark on Leo1 promoter through JMJD2C, resulting in a more transcriptionally permissive chromatin state, leading to the efficient promotion of Leo1 expression.

Leo1 suppression depletes the levels of PAF complex components and affects the expression of target genes

The PAF complex subunits have been shown to coimmunoprecipitate with each other. Therefore, we checked whether knockdown of Leo1 affects the interaction network of the PAF complex. Interestingly, we found that loss of Leo1 destabilized the PAF complex, and the mRNA levels of Paf1, Ctr9, and Ski8 were collectively downregulated (Fig. 7A). In addition, the protein levels of Paf1 and Ctr9 were also decreased with Leo1 depletion (Fig. 7B). To determine the functional relevance of Leo1-mediated destabilization of the PAF complex, we further examined the effects on several reported PAF complex target genes. A subset of these genes, namely Sox2, Sox4, and Tbx3, were significantly downregulated when Leo1 was silenced in Molm-14 cells (Fig. 7C).

Figure 7.

Leo1 knockdown destabilizes the PAF complex. A and B, Molm-14 cells were transfected with either Leo1 shRNA or control shRNA (Emp), and qRT-PCR (A) or Western blot (B) was done for the expression of the PAF complex components. Data, mean ± SEM of triplicates. C, qRT-PCR analyses of Sox2, Sox4, and Tbx3 after Leo1 knockdown in Molm-14 cells. Data, mean ± SEM of triplicates. D, Molm-14 was transfected with PRL-3 shRNA and analyzed for PAF components. Data, mean ± SEM of triplicates. E, Molm-14 was cotransfected with PRL-3 shRNA and Leo1 expression construct and analyzed for PAF components. Data, mean ± SEM of triplicates. *, P < 0.05; **, P < 0.005.

Figure 7.

Leo1 knockdown destabilizes the PAF complex. A and B, Molm-14 cells were transfected with either Leo1 shRNA or control shRNA (Emp), and qRT-PCR (A) or Western blot (B) was done for the expression of the PAF complex components. Data, mean ± SEM of triplicates. C, qRT-PCR analyses of Sox2, Sox4, and Tbx3 after Leo1 knockdown in Molm-14 cells. Data, mean ± SEM of triplicates. D, Molm-14 was transfected with PRL-3 shRNA and analyzed for PAF components. Data, mean ± SEM of triplicates. E, Molm-14 was cotransfected with PRL-3 shRNA and Leo1 expression construct and analyzed for PAF components. Data, mean ± SEM of triplicates. *, P < 0.05; **, P < 0.005.

Close modal

To determine whether Leo1 plays an independent role in mediating the oncogenic properties in AML or through PAF complex as a whole, we compared the effects of Leo1, Paf1, or Ctr9 knockdown in TF1-PRL3 cells. We confirmed the knockdown efficiency using Western Blot (Supplementary Fig. S7A). Interestingly, we found that knockdown of Leo1 impedes cell proliferation to a greater extent than Paf1 or Ctr9 knockdown (Supplementary Fig. S7B). In addition, Sox2 and Sox4 expression levels were similarly affected in Leo1, Paf1 of Ctr9 knockdown (Supplementary Fig. S7C). These observations suggest that the entire PAF complex is partly involved in mediating oncogenesis though SOX genes, but other mechanisms involving Leo1 alone are present. Furthermore, when we analyzed the expression levels of Leo1, Paf1, Cdc73, Ski8, and Ctr9 in TF1-PRL3 cells, only Leo1 showed an upregulation (Supplementary Fig. S7D).

Because we showed that knockdown of PRL-3 downregulated levels of Leo1, we further examined whether knockdown of PRL-3 similarly destabilizes the PAF complex. Interestingly, silencing PRL-3 downregulated all components of the PAF complex (Fig. 7D), and subsequent restoration of Leo1 levels rescued the levels of Paf1, Ski8, and Ctr9 (Fig. 7E). These results further confirm that Leo1 is functionally relevant and is an important downstream mediator of PRL-3 oncogenic functions.

Our present work represents the first large-scale quantitative survey of proteins regulated by PRL-3 in leukemia. We identified several known proteins downstream of PRL-3, including stathmin, nucleolin, and translationally controlled tumor protein (6, 18, 19), which endorsed the validity of our approach. We also discovered additional upregulated candidate oncogenes that were linked to carcinogenesis, such as hnRNPH1, TPI1, HSP90B1, prohibitin, as well as candidate tumor suppressors that were downregulated, such as POTE, nucleophosmin, and BTF3, which were previously not known to be modulated by PRL-3 (Supplementary Table S1). On the other hand, some reported substrates of PRL-3 in solid tumor cell lines, such as CDH22, ezrin, keratin8, integrin-α1, and EF-2, were not identified in our screen, suggesting that PRL-3 might activate both common and distinct networks in solid tumor as compared with leukemia.

From our comprehensive proteomic dataset, we gained several novel insights into the functions of PRL-3. First, we found that PRL-3 unexpectedly alters the expression levels of a significant number of nuclear proteins (50%), although it was reported to localize predominantly in the plasma membrane and cytosol due to the CAAX membrane-targeting signal (20). Despite this, PRL-3 was also found to be able to shuttle between the nucleus and cytoplasm in myeloma and AML cells (21), by an unclear mechanism. Second, gene ontology annotation by biologic process reveals a specific involvement of PRL-3 in pre-mRNA splicing and nucleic acid metabolism, as well as cell-cycle regulation and cell motility. Involvement in cell motility is consistent with published reports in which overexpression of PRL-3 in carcinomas has been linked to processes such as enhanced cell migration, invasion, and metastasis (20, 22). Notably, this is the first study to suggest that PRL-3 plays a role in the regulation of RNA-related processes. Overall, our unbiased approach enables us to implicate novel candidate proteins in the oncogenic function of PRL-3, which would fundamentally advance our understanding of how this phosphatase contributes to cancer progression.

In the recent years, several pivotal articles have proposed functions of the PAF complex in development and cancer, and the PAF complex was found to be substrates of another oncogenic phosphatase, SHP2 (23–26). In addition, the PAF complex is essential for leukemogenesis by mixed-lineage leukemia fusion proteins through inducing the expression of Hox genes, thus maintaining a haematopoietic stem cell (HSC)–like signature and resistance to differentiation (15). These led us to focus on the involvement of Leo1 and the PAF complex as the downstream mediators of PRL-3 activities. Although characterization of the individual subunits of the PAF complex revealed opposing roles in cancer (24), our current work suggests that Leo1 is a proto-oncogene in leukemia, and the levels of Ski8, Ctr9, and Paf1 subunits depended on Leo1 levels (Fig. 7A and B), supporting the notion of an interdependency within the PAF complex.

One important hallmark of AML is the acquisition of self-renewal capacity in the leukemic blast cells (27) through the disruption of self-renewal pathways such as Hox and Wnt/β-catenin. Interestingly, the PAF complex has been found to augment the expression of HOXA9 (15), and in another independent report, Leo1 has been found to associate directly with β-catenin (28). In addition, the PAF complex has also been reported to regulate directly key pluripotency genes in embryonic stem cells. Indeed, when we knocked down Leo1, several pluripotency genes, such as Sox2, Sox4, and Tbx3, were downregulated (Fig. 7C). The deregulation of Sox genes, especially Sox4, has been shown to be an important collaborating driver of myeloid leukemogenesis in several transgenic murine models of leukemia (29). Thus, our study suggests that PRL-3 could exert oncogenic properties in AML through Leo1 and PAF complex–mediated deregulation of critical pluripotency genes previously implicated in leukemogenesis.

There were few reports on how PRL-3 regulates its targets, in part due to the challenges in identifying the genuine targets of PRL-3. In our study, we revealed a novel mechanism of the transcriptional involvement of PRL-3 in regulating Leo1. Trimethylation on H3K9 is a strong repressive epigenetic mark generally associated with heterochromatin maintenance and transcriptional repression, and deregulation of H3K9me3 levels has been shown to cause aberrant gene expression, contributing to the development of cancer (30, 31). We found global demethylation of H3K9me3 in TF1-hPRL3 cells, suggesting that alterations in a histone-modifying enzyme lead to these widespread epigenetic changes (32). JMJD2C, originally identified as GASC1, is a gene located on chromosome 9p, and is frequently amplified in cancers (30). Our results identify JMJD2C as the main enzyme involved in the demethylation of H3K9me3 on the Leo1 promoter in TF1-hPRL3 cells, thereby enhancing the expression of Leo1. Interestingly, we noticed an increase in JMJD2C occupancy on the promoter region of Leo1, and this was dependent on PRL-3 expression levels (Fig. 6E). How PRL-3 affects the binding of JMJD2C or possibly other histone demethylase on the promoters of target genes remains to be determined. Importantly, our results demonstrated a possible interface between the activities of a phosphatase and a histone demethylase in gene regulation.

In conclusion, our study unravels downstream signaling pathways of PRL-3, identifying Leo1 as an important downstream target. Taken together, this study provides novel insights into the mechanisms by which PRL-3 promotes the malignant phenotype in AML.

No potential conflicts of interest were disclosed.

Conception and design: P.S.Y. Chong, J. Zhou, W.J. Chng

Development of methodology: P.S.Y. Chong, J. Zhou, S.K. Sze, W.J. Chng

Acquisition of data (provided animals, acquired and managed patients, provided facilities, etc.): J. Zhou, L.-L. Cheong, J. Qian, S.K. Sze, W.J. Chng

Analysis and interpretation of data (e.g., statistical analysis, biostatistics, computational analysis): P.S.Y. Chong, J. Zhou, J. Qian, T. Guo, S.K. Sze, W.J. Chng

Writing, review, and or revision of the manuscript: P.S.Y. Chong, J. Zhou, W.J. Chng

Administrative, technical, or material support (i.e., reporting or organizing data, constructing databases): L.-L. Cheong, S.-C. Liu, J. Qian, Q. Zeng, W.J. Chng

Study supervision: J. Zhou

The authors thank professors Daniel G. Tenen, H. Phillip Koeffler, and assistant professor Motomi Osato for helpful suggestions.

W.J. Chng is supported by a National Medical Research Council (NMRC) Clinician Scientist Investigator award. This work is partly supported by a Singapore Cancer Syndicate Grant. This research is also supported by the National Research Foundation Singapore and the Singapore Ministry of Education under the Research Centers of Excellence initiative. J. Zhou is supported by an NMRC Clinician-Scientist IRG grant CNIG11nov38.

The costs of publication of this article were defrayed in part by the payment of page charges. This article must therefore be hereby marked advertisement in accordance with 18 U.S.C. Section 1734 solely to indicate this fact.

1.
Zeng
Q
,
Si
X
,
Horstmann
H
,
Xu
Y
,
Hong
W
,
Pallen
CJ
. 
Prenylation-dependent association of protein-tyrosine phosphatases PRL-1, -2, and -3 with the plasma membrane and the early endosome
.
J Biol Chem
2000
;
275
:
21444
52
.
2.
Al-Aidaroos
AQ
,
Zeng
Q
. 
PRL-3 phosphatase and cancer metastasis
.
J Cell Biochem
2010
;
111
:
1087
98
.
3.
Kato
H
,
Semba
S
,
Miskad
UA
,
Seo
Y
,
Kasuga
M
,
Yokozaki
H
. 
High expression of PRL-3 promotes cancer cell motility and liver metastasis in human colorectal cancer: a predictive molecular marker of metachronous liver and lung metastases
.
Clin Cancer Res
2004
;
10
:
7318
28
.
4.
Polato
F
,
Codegoni
A
,
Fruscio
R
,
Perego
P
,
Mangioni
C
,
Saha
S
, et al
PRL-3 phosphatase is implicated in ovarian cancer growth
.
Clin Cancer Res
2005
;
11
:
6835
9
.
5.
Radke
I
,
Gotte
M
,
Kersting
C
,
Mattsson
B
,
Kiesel
L
,
Wulfing
P
. 
Expression and prognostic impact of the protein tyrosine phosphatases PRL-1, PRL-2, and PRL-3 in breast cancer
.
Br J Cancer
2006
;
95
:
347
54
.
6.
Chu
ZH
,
Liu
L
,
Zheng
CX
,
Lai
W
,
Li
SF
,
Wu
H
, et al
Proteomic analysis identifies translationally controlled tumor protein as a mediator of phosphatase of regenerating liver-3-promoted proliferation, migration and invasion in human colon cancer cells
.
Chin Med J (Engl)
2011
;
124
:
3778
85
.
7.
Jiang
Y
,
Liu
XQ
,
Rajput
A
,
Geng
L
,
Ongchin
M
,
Zeng
Q
, et al
Phosphatase PRL-3 is a direct regulatory target of TGFbeta in colon cancer metastasis
.
Cancer Res
2011
;
71
:
234
44
.
8.
Saha
S
,
Bardelli
A
,
Buckhaults
P
,
Velculescu
VE
,
Rago
C
,
Croix
BS
, et al
A phosphatase associated with metastasis of colorectal cancer
.
Science
2001
;
294
:
1343
6
.
9.
Bardelli
A
,
Saha
S
,
Sager
JA
,
Romans
KE
,
Xin
B
,
Markowitz
SD
, et al
PRL-3 expression in metastatic cancers
.
Clin Cancer Res
2003
;
9
:
5607
15
.
10.
Wang
H
,
Vardy
LA
,
Tan
CP
,
Loo
JM
,
Guo
K
,
Li
J
, et al
PCBP1 suppresses the translation of metastasis-associated PRL-3 phosphatase
.
Cancer Cell
2010
;
18
:
52
62
.
11.
Bessette
DC
,
Qiu
D
,
Pallen
CJ
. 
PRL PTPs: mediators and markers of cancer progression
.
Cancer Metastasis Rev
2008
;
27
:
231
52
.
12.
Guo
K
,
Li
J
,
Tang
JP
,
Tan
CBP
,
Hong
CW
,
Al-Aidaroos
AQ
, et al
Targeting intracellular oncoproteins with antibody therapy or vaccination
.
Sci Transl Med
2011
;
3
:
99ra85
.
13.
Zhou
J
,
Bi
C
,
Chng
WJ
,
Cheong
LL
,
Liu
SC
,
Mahara
S
, et al
PRL-3, a metastasis associated tyrosine phosphatase, is involved in FLT3-ITD signaling and implicated in anti-AML therapy
.
PLoS ONE
2011
;
6
:
e19798
.
14.
Beekman
R
,
Valkhof
M
,
Erkeland
SJ
,
Taskesen
E
,
Rockova
V
,
Peeters
JK
, et al
Retroviral integration mutagenesis in mice and comparative analysis in human AML identify reduced PTP4A3 expression as a prognostic indicator
.
PLoS ONE
2011
;
6
:
e26537
.
15.
Muntean
AG
,
Tan
J
,
Sitwala
K
,
Huang
Y
,
Bronstein
J
,
Connelly
JA
, et al
The PAF complex synergizes with MLL fusion proteins at HOX loci to promote leukemogenesis
.
Cancer Cell
2010
;
17
:
609
21
.
16.
Liu
Y
,
Zheng
P
,
Liu
Y
,
Ji
T
,
Liu
X
,
Yao
S
, et al
An epigenetic role for PRL-3 as a regulator of H3K9 methylation in colorectal cancer
.
Gut
2013
;
62
:
571
81
.
17.
Shin
S
,
Janknecht
R
. 
Diversity within the JMJD2 histone demethylase family
.
Biochem Biophys Res Commun
2007
;
353
:
973
7
.
18.
Semba
S
,
Mizuuchi
E
,
Yokozaki
H
. 
Requirement of phosphatase of regenerating liver-3 for the nucleolar localization of nucleolin during the progression of colorectal carcinoma
.
Cancer Sci
2010
;
101
:
2254
61
.
19.
Zheng
P
,
Liu
YX
,
Chen
L
,
Liu
XH
,
Xiao
ZQ
,
Zhao
L
, et al
Stathmin, a new target of PRL-3 identified by proteomic methods, plays a key role in progression and metastasis of colorectal cancer
.
J Proteome Res
2010
;
9
:
4897
905
.
20.
Zeng
Q
,
Dong
JM
,
Guo
K
,
Li
J
,
Tan
HX
,
Koh
V
, et al
PRL-3 and PRL-1 promote cell migration, invasion, and metastasis
.
Cancer Res
2003
;
63
:
2716
22
.
21.
Fagerli
UM
,
Holt
RU
,
Holien
T
,
Vaatsveen
TK
,
Zhan
F
,
Egeberg
KW
, et al
Overexpression and involvement in migration by the metastasis-associated phosphatase PRL-3 in human myeloma cells
.
Blood
2008
;
111
:
806
15
.
22.
Fiordalisi
JJ
,
Keller
PJ
,
Cox
AD
. 
PRL tyrosine phosphatases regulate rho family GTPases to promote invasion and motility
.
Cancer Res
2006
;
66
:
3153
61
.
23.
Squazzo
SL
,
Costa
PJ
,
Lindstrom
DL
,
Kumer
KE
,
Simic
R
,
Jennings
JL
, et al
The Paf1 complex physically and functionally associates with transcription elongation factors in vivo
.
EMBO J
2002
;
21
:
1764
74
.
24.
Chaudhary
K
,
Deb
S
,
Moniaux
N
,
Ponnusamy
MP
,
Batra
SK
. 
Human RNA polymerase II-associated factor complex: dysregulation in cancer
.
Oncogene
2007
;
26
:
7499
507
.
25.
Wang
P
,
Bowl
MR
,
Bender
S
,
Peng
J
,
Farber
L
,
Chen
J
, et al
Parafibromin, a component of the human PAF complex, regulates growth factors and is required for embryonic development and survival in adult mice
.
Mol Cell Biol
2008
;
28
:
2930
40
.
26.
Takahashi
A
,
Tsutsumi
R
,
Kikuchi
I
,
Obuse
C
,
Saito
Y
,
Seidi
A
, et al
SHP2 tyrosine phosphatase converts parafibromin/Cdc73 from a tumor suppressor to an oncogenic driver
.
Mol Cell
2011
;
43
:
45
56
.
27.
Tenen
DG
. 
Disruption of differentiation in human cancer: AML shows the way
.
Nat Rev Cancer
2003
;
3
:
89
101
.
28.
Mosimann
C
,
Hausmann
G
,
Basler
K
. 
Parafibromin/Hyrax activates Wnt/Wg target gene transcription by direct association with beta-catenin/Armadillo
.
Cell
2006
;
125
:
327
41
.
29.
Sandoval
S
,
Kraus
C
,
Cho
EC
,
Cho
M
,
Bies
J
,
Manara
E
, et al
Sox4 cooperates with CREB in myeloid transformation
.
Blood
2012
;
120
:
155
65
.
30.
Cloos
PA
,
Christensen
J
,
Agger
K
,
Maiolica
A
,
Rappsilber
J
,
Antal
T
, et al
The putative oncogene GASC1 demethylates tri- and dimethylated lysine 9 on histone H3
.
Nature
2006
;
442
:
307
11
.
31.
Bannister
AJ
,
Kouzarides
T
. 
Regulation of chromatin by histone modifications
.
Cell Res
2011
;
21
:
381
95
.
32.
Fullgrabe
J
,
Kavanagh
E
,
Joseph
B
. 
Histone onco-modifications
.
Oncogene
2011
;
30
:
3391
403
.