PRL-3 (PTP4A3), a metastasis-associated phosphatase, is also upregulated in patients with acute myeloid leukemia (AML) and is associated with poor prognosis, but the underlying molecular mechanism is unknown. Here, constitutive expression of PRL-3 in human AML cells sustains leukemogenesis in vitro and in vivo. Furthermore, PRL-3 phosphatase activity dependently upregulates LIN28B, a stem cell reprogramming factor, which in turn represses the let-7 mRNA family, inducing a stem cell–like transcriptional program. Notably, elevated levels of LIN28B protein independently associate with worse survival in AML patients. Thus, these results establish a novel signaling axis involving PRL-3/LIN28B/let-7, which confers stem cell–like properties to leukemia cells that is important for leukemogenesis.
Implications: The current study offers a rationale for targeting PRL-3 as a therapeutic approach for a subset of AML patients with poor prognosis. Mol Cancer Res; 15(3); 294–303. ©2016 AACR.
The 5-year overall survival of acute myeloid leukemia (AML) remains about 30% to 40% and significantly poorer in patients older than 65 years (1, 2). Leukemia stem cells (LSC), also called leukemia-initiating cells, a distinct subpopulation of AML cells with self-renewal capacity, initiate and sustain the development of the bulk leukemic population (3–5). LSC has been considered as the root source for the relapse and treatment failure of AML (6, 7). The identification of the molecular mechanisms underlying the transformation of LSCs offers a potential opportunity to eradicate leukemia (8).
Protein tyrosine phosphatase of regenerating liver 3 (PRL-3, encoded by PTP4A3 gene) is a member of the VH1-like protein tyrosine phosphatase (PTP) family with dual specificity (9). Elevated expression of PRL-3 was detected in a variety of metastatic and primary tumor tissues (10). PRL-3 activates the PI3K/AKT pathway (11) and Src-ERK1/2 pathways (12), thus promoting epithelial–mesenchymal transition (EMT) and tumor angiogenesis in solid tumors.
PRL-3 protein is detected in approximately 50% and 90% of bone marrow samples from patients with AML and multiple myeloma, respectively (13–16). Importantly, several studies have independently confirmed that the high expression of PRL-3 is associated with poor survival in AML (15–17). We previously demonstrated that LEO1, a component of RNA polymerase II–associated factor (PAF) complex, was induced by PRL-3 in AML (18). PAF complex plays an essential role in MLL-rearranged leukemia (19). These findings collectively indicate that dysregulated PRL-3 is involved in the development of AML. However, the molecular mechanisms underlying the role of PRL-3 in AML are not well understood.
Here, we report that LIN28B is a key oncogenic target in PRL-3–positive AML. Increased LIN28B in turn downregulates let-7 mRNA family. The PRL-3/LIN28B/let-7 axis takes part in the transformation of LSCs and leukemogenesis of AML. Antagonizing PRL-3 may be of clinical benefit in the treatment of AML and eradication of LSCs.
Materials and Methods
Cell lines and cell culture
AML cell lines MOLM-14, HL60, TF-1, and TF-1a were grown in RPMI1640 (Invitrogen) supplemented with 10% FBS (JRH Biosciences Inc.) at a density of 2–10 × 105 cells/ml in a humid incubator with 5% CO2 at 37°C. Additional human IL3 (PeproTech) was added into growth medium at 5 ng/mL to support TF-1 cell growth. Bone marrow blast cells (>90%) from newly diagnosed AML patients were obtained at National University Hospital in Singapore with informed consent. This study was approved by the Institutional Review Board of National University of Singapore.
Cell viability assays
Leukemic cells were seeded in 96-well culture plates at a density of 2 × 104 viable cells/100 μL/well in triplicates. CellTiter-Glo Luminescent Cell Viability Assay (CTG assay, Promega) was used to determine the cell growth and viability as described previously (20). Each experiment was in triplicate.
Affymetrix microarray and gene set enrichment analyses
TF1-pEGFP and TF1-hPRL-3 cells were harvested, and total RNA was extracted using the RNeasy Mini Kit, according to the manufacturer's instructions (Qiagen). Gene expression profiling was performed using Affymetrix U133 Plus 2.0 gene chip (Affymetrix) according to the manufacturer's protocol. The scanned data were processed using MicroArray Suite version 5.0 (Affymetrix). The detailed microarray analysis and subsequent gene set enrichment analysis (GSEA) are described in Supplementary Materials and Methods. Expression data (GSE64872) were deposited in the Gene Expression Omnibus.
Western blot analysis
Preparation of the cell lysate and immunoblotting were performed using standard techniques as described previously (21), except for PRL-3 analysis, and 16% gel was used. Anti-PRL3 antibody was clone 318 (1:1,000) as reported previously (22). Anti-LIN28B antibody was purchased from Santa Cruz Biotechnology (Cat. #sc-130802). Anti-β-actin (1:3,000, Sigma-Aldrich) was used as loading control.
Lentivirus LIN28B-shRNA infection
Scramble shRNA control and two human LIN28B specifically pLKO.1 lentiviral shRNAs were purchased from Open Biosystems. TF1-hPRL-3 or TF-1a cells (3 × 106) were mixed with concentrated viral supernatant and 8 μg/mL of polybrene (Millipore) and centrifuged at 2,500 rpm for 90 minutes at 30°C. After additional incubation at 37°C for 4 hours, the medium was changed to fresh complete medium. Two days later, cells were submitted for protein extraction, followed by Western blot analysis of LIN28B and PRL-3.
mRNA mimics and inhibitors
Predesigned miRCURY LNA microRNA Mimics to let-7a (#470408), let-7b (#470775), mRNA mimic negative control (#479903), specific miRCURY LNA microRNA Inhibitors to let-7a (#4101777), let-7b (#4100945), and control inhibitor (#199006) were purchased from Exiqon. Briefly, cells were transfected with 30 nmol/L mRNA mimics or mRNA inhibitors or their negative controls by electroporation with Neon Transfection System (Thermo Fisher Scientific) and were maintained in a humidified incubator at 37°C in 5% CO2. Cell proliferation assays were carried out at days 1, 2, 4, and 6 after transfection. For colony-forming assays, colony numbers were counted at day 7.
Colony formation assay and serial replating assay
TF1-pEGFP and TF1-hPRL-3 cells were washed twice with 1× PBS, and cell viability was determined by Typan blue exclusion method. After that, about 20,000 vial cells from each cell line were plated in MethoCult medium without cytokines (H4230, STEMCELL Technologies) in 6-well plates and cultured for 7 days at 37°C in a 5% CO2 incubator. Colonies consisting of more than 50 cells were counted under an inverted microscope. A total of five random 4 × 10 magnification fields were selected, and average number of colonies of each sample was calculated. The experiments were duplicated. In serial replating assay, the colony number was counted, and then subjected to replating every 7 days.
Limited serial dilution and bone marrow transplantation
The protocol was reviewed and approved by Institutional Animal Care and Use Committee in compliance with the Guidelines on the Care and Use of Animals for Scientific Purpose. Detailed information is provided in Supplementary Materials and Methods.
Student t test (two-tailed paired) was used for examining the statistical difference for in vitro cell line experiments, such as cell viability assay and colony-forming assay; P values of <0.05 were considered to be significant. Data were presented as mean ± SD. Multivariate survival analyses were conducted using R version 3.1.2 and survival package version 2.37-7. Protein expression of LIN28B and clinicopathologic parameters were first discretized into binary categories prior to Cox regression analysis. Kaplan–Meier analyses were conducted using GraphPad Prism version 5.04 (GraphPad Software). Statistical significance of the Kaplan–Meier analysis was calculated by log-rank test (P < 0.05).
PRL-3 promotes AML maintenance and progression in vivo
As reported in previous studies (15, 18), we established 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, and showed that PRL-3 promoted cytokine-independent growth of AML cells. Here, we further determined the in vivo function of PRL-3 on leukemogenesis by using this pair of TF1 cells in a bone marrow transplantation xenograft assay. NOD/SCID mice inoculated with TF1-hPRL3 cells via tail vein injection developed leukemia-like symptoms with enlarged spleen and liver and succumbed to the disease within 3 months. In contrast, mice implanted with TF1-pEGFP did not display any sign of sickness, and all survived till the end of the experimental period (Fig. 1A). We next performed serial bone marrow transplantation experiments. Bone marrow cells harvested from the primary recipients were transplanted into secondary recipients. Compared with primary leukemia, there was a significant shorter survival in the secondary recipients (Fig. 1A, P < 0.01). We then performed tertiary transplantation. Tertiary recipients succumbed to the disease even earlier than the secondary recipients (Fig. 1A, P < 0.001). Histologic analysis revealed that leukemic cells extensively infiltrated into the liver and spleen in diseased mice (Fig. 1B). Moreover, the expressions of PRL-3 and LIN28B were significantly elevated in engrafted mouse bone marrow cells harvested from secondary and tertiary recipients, while their expression of let-7 was significantly lower as compared with the primary generation of mouse bone marrow cells (Fig. 1C). These data bespeak a good correlation between the expressions of these three genes and aggressiveness of the disease.
Together, these studies demonstrate that PRL-3 promotes AML maintenance and progress in an animal model.
PRL-3 induces a “stemness” transcriptional program in AML cells
To gain insight into the role of PRL-3 in leukemogenesis, we conducted an mRNA expression microarray analysis of TF1-pEGFP and TF1-hPRL3 cell lines. Using 2-fold as the cut-off level, 89 genes showed increased expression and 24 genes had decreased expression in TF1-hPRL3 cells compared with TF1-pEGFP cells (Supplementary Table S1). The top 10 up- and downregulated genes were shown in Supplementary Table S2. It is noteworthy that PRL-3 induced expression of many LSC or cancer stem cell–related genes, including LIN28B, KIT, IGF2BP1 (IGF2 mRNA binding protein 1), LAPTM4B, FYN, EIF4E3, Caveolin 1, among others, while reducing the expression of cell surface markers of more differentiated cells, such as CD36 and CD38 (Supplementary Table S1). LIN28B ranked as the third most elevated gene after PTP4A3 (PRL-3) and MT1G in TF-hPRL3 cells (Supplementary Table S2). It has been well established that LIN28B is a stem cell reprogramming factor and a marker for cancer stem cell (23). qRT-PCR analysis validated the changes of selected genes, including LIN28B, LAPTM4B, IGF2BP1, FYN, RAB27B, KIT, CD36, and CD38 (Supplementary Fig. S1), and increased LIN28B in TF1-hPRL3 cells was confirmed by additional RT-PCR and Western blot assays (Fig. 2A). GSEA of this microarray result revealed a significant enrichment of genes identified in a previously defined gene expression signature of leukemia stem cell (LSC SIGNATURE_SAITO, P < 0.001) and hematopoietic stem cell (HSC SIGNATURE_EPPERT, P < 0.001; Fig. 2B). This differentially expressed gene list was uploaded to the MetaCore algorithm for gene cluster analysis, and the result showed LIN28B to act as a focal point in the regulation of many genes such as miRNA let-7 family (Supplementary Fig. S2).
Notably, the let-7 miRNA family members play a “gatekeeper” role of embryonic stem cell pluripotency and act as tumor suppressors in a wide range of human malignancies (24). LIN28 was identified to selectively repress the expression of let-7 miRNAs or directly bind mRNA (23, 25, 26). Therefore, we quantitated the level of let-7 miRNAs in TF1-pEGFP and TF1-hPRL3 cells. With the exception of let-7c, the other let-7 miRNA members, including let-7a, let-7b, let-7d, let-7f, let-7g, let-7i, and miR-98, were all significantly repressed in TF1-hPRL3 cells relative to TF1-pEGFP cells (Fig. 2C). To rule out the effect of cell type specificity, we overexpressed PRL-3 into other AML cell line, HL60, and measured the expression of LIN28B and let-7a miRNA. Similarly, we observed that HL60-hPRL3 cells had significant higher expression of PRL-3 and LIN28B, but lower let-7a as compared with HL60-pEGFP cells (Fig. 2D). In summary, our data indicate that PRL-3 enforces a stem-like transcription program in AML cells, driven by changes of LIN28B and let-7 miRNAs.
PRL-3–expressing AML cells have leukemia-initiating capabilities
It has been well established that leukemia-initiating cells are enriched in CD34+CD38− subpopulation of AML cells. Lineage analysis demonstrated that the percentage of CD34+CD38− cells was 8-fold higher in TF1-hPRL3 cell lines than in TF1-pEGFP cell lines (Fig. 3A; Supplementary Fig. S3A). We also found that PRL-3 was more highly expressed on CD34+CD38− AML than CD34+CD38+ AML cells, as well as CD34+CD38− cells from healthy controls, suggesting PRL-3 was upregulated in AML-initiating cells (Fig. 3B; Supplementary Fig. S3B and S3C). To assess the functional involvement of PRL-3 in leukemia-initiating cells, we used serial replating assay to assess another intrinsic characteristic of leukemia-initiating cells, self-renewal capacity, in TF1-hPRL3 cells. As expected, TF1-hPRL3 cells could be serially replated in basic methylcellulose without additional cytokines (Fig. 3C), whereas no colonies could be observed in TF1-pEGFP cells. To exclude possible artefacts accompanied by clone selection, we established another pair of isogenic TF1-FUGW and TF1-FUGW-PRL3 after transduced TF-1 cells with either FUGW control lentivirus or FUGW-PRL3 lentivirus and selection. In accordance with the previous report, TF1-FUGW-PRL3 cells gained cytokine-independent growth advantage, whereas TF1-FUGW cells failed to do (Supplementary Fig. S4). We next assessed the replating capacity of this pair of isogenic cells. In agreement with the results from TF1-pEGFP and TF1-hPRL3, TF1-FUGW-PRL3 cells could be serially replated, whereas TF1-FUGW control cells could not form colonies (Supplementary Fig. S4). To address the question of whether TF1-hPRL3 cells maintain leukemia-initiating cell properties in vivo, we conducted limiting dilution transplantation and serial BMT assays. The ability of TF1-hPRL3 cells to propagate leukemia in immunodeficient mice was evaluated using a 5-time cell dilution (50,000, 10,000, 2,000, and 400 cells). Leukemic symptoms development characterized by hepatosplenomegaly was observed in all mice injected with 50,000 and 10,000 cells (5 mice/group), and 4 of 5 mice receiving 2,000 cells (Supplementary Table S3; Supplementary Fig. S5A). Consistently, we also observed significantly higher expression of PRL-3 and LIN28B and decreased expression of let-7a miRNA in bone marrow cells harvested from diseased mice (Supplementary Fig. S5B). In contrast, none of the mice (n = 5) transplanted with 400 cells developed disease up to 6 months postinoculation (Supplementary Table S3; Supplementary Fig. S5A). So the estimated frequency of leukemia-initiating cell in TF1-hPRL3 cells could be 1 in 1,771 [95% confidence interval (CI), 712–4,410) cells according to the analysis with ELDA software. In contrast, mice (n = 5) receiving 50,000 TF1-pEGFP cells were all healthy at the end of experiments (Supplementary Table S3). Collectively, these results demonstrate a functional role for PRL-3 on transformation of leukemia-initiating cells in vitro and in vivo.
LIN28B is a key downstream target of PRL-3
Through our microarray study, LIN28B is one of the most significantly upregulated gene in TF1-hPRL3 compared with TF1-pEGFP. The C(X)5R motif within the P-loop of the PRL-3 protein possess the PTP enzymatic activity. To examine the importance of PRL-3 phosphatase activity in the regulation of LIN28B, we transfected a catalytic domain mutant of PRL-3 (C104S) into TF-1 cells and checked LIN28B protein level. Western blotting analysis showed that the substitution of C104 with serine effectively diminished the PRL-3–mediated upregulation of LIN28, suggesting a phosphatase activity–dependent mechanism of regulation of LIN28B (Supplementary Fig. S6A). To verify the relevance of LIN28B overexpression in clinical samples, we examined the expression of PRL-3 and LIN28B in CD34+CD38− cells isolated from 28 primary AML samples using quantitative RT-PCR. As shown in Fig. 4A, there was a significant positive correlation between the expression levels of PRL-3 and LIN28B (Pearson r = 0.51; two-tailed P = 0.007). Furthermore, as LIN28B plays an important role in repressing miRNAs and cellular reprogramming (27), we therefore focused on the biological significance and regulatory mechanisms of LIN28B in the transformation of LSCs induced by PRL-3. We used two independent LIN28B-shRNAs to selectively knockdown LIN28B expression in TF1-hPRL3 cells, and then performed serial replating assays. As shown in Fig. 4B, significant suppression of LIN28B protein was achieved by both shRNAs. The knockdown efficacy of shRNA2 was higher than shRNA1. Scrambled shRNA–treated TF1-hPRL3 cells could be replated and the numbers of the colonies increased with each round, whereas the colony number of LIN28B-shRNAs transduced cells was markedly diminished in each round of plating (Fig. 4C). Importantly, the significant reduction in the number of colonies correlated with the level of suppression of LIN28 expression (Fig. 4B), indicating the on-target effect from specific shRNAs. We used lentiviral transduction method to specifically knockdown PRL-3 in MOLM-14 cells and then performed serial replating assays. Next, stable shRNA-mediated knockdown led to stable reductions in PRL-3 expression by 80% (Supplementary Fig. S6B). These cells and cells transduced with scramble shRNA were then plated in methylcellulose for CFU in vitro assays. We observed significant reduced number of colonies in PRL-3 shRNA-expressing cells than scramble shRNA-expressing cells (Supplementary Fig. S6C). Collectively, these observations imply that inhibition of PRL-3 and LIN28B effectively diminishes the replating capacity of human AML cells.
Previously, a subclone of TF-1 cell line, TF-1a, was generated. TF-1a can proliferate continuously without cytokines and develop tumor in nude mice, in contrast to its parental line (28). Immunophenotypically, TF-1a cells are more immature and primitive (CD34+/CD38−) than the parental TF-1 cells (CD34+/CD38+; ref. 28). TF-1a cells and TF1-hPRL3 cells share similarities in these features, which are indicators of self-renewal and uncontrolled proliferative capacity. In accordance with results obtained from TF1-hPRL3 cells, high expression of PRL-3 and LIN28B protein was found in TF-1a cells compared with TF-1 cells (Fig. 4D). It is worth pointing out that the expression of PRL-3 and LIN28B in TF1-hPRL3 cells was comparable with their endogenous levels in TF-1a cells (Fig. 4D), suggesting our observations were not artificially produced. The different levels of PRL-3 and LIN28B proteins between TF-1a and TF-1 might be the underlying cause for the abovementioned different characteristics between these two cell lines. To exclude any possible artifacts caused by overexpression or nonspecific effect of the shRNA approach, we next conducted a loss-of-function together with a rescue experiment in the TF-1a cell line expressing endogenously high level of PRL-3 and LIN28B. TF-1a cells transduced with PRL-3-shRNAs or LIN28B-shRNAs produced significantly less number of colonies in the serial replating assay, while replating activity of PRL-3-shRNA transduced cells was restored to near normal level by the overexpression of LIN28B (Fig. 4E and F). In conclusion, these data support that LIN28B is essential for PRL-3–mediated replating capacity of leukemic cells.
Repression of let-7 miRNAs is required for the PRL-3/LIN28B-promoted leukemogenesis
The RNA-binding protein, LIN28B, functions to repress let-7 miRNAs via blocking the microprocessor complex (29). As it is well established that let-7 is an important tumor suppressor gene, which inhibits numerous oncogenes, such as RAS, c-MYC, NFκB, and HMGA (30, 31), we decided to examine whether the repression of let-7 miRNAs is required for PRL-3/LIN28B-induced leukemogenesis. Ectopic expression of let-7a and let-7b led to decreased cell colony-forming capacity of TF1-hPRL3 cells (Fig. 5A, P < 0.05) and proliferation (Fig. 5B, P < 0.01). Concordantly, antagomiRs to let-7a and let-7b significantly increased colony-forming capacity of TF1-hPRL3 cells (Fig. 5C, P < 0.05) and cell proliferation (Fig. 5D, P < 0.01). To further validate the important role of let-7 miRNAs, we repeated the same experiments by replacing TF1-hPRL3 cell line with MOLM-14, expressing high level of endogenous PRL-3 and LIN28B protein. Indeed, similar results were observed for MOLM-14 cells (Supplementary Fig. S7). Taken together, these evidences demonstrate that let-7 plays an essential role in the PRL-3/LIN28B-mediated leukemogenesis.
LIN28B overexpression independently predicts poor survival of patients with AML
To examine the clinical importance of LIN28B in AML, we used IHC to determine LIN28B protein expression in bone marrow samples from 159 patients with AML at diagnosis. LIN28B staining was calculated as the number of LIN28B-positive blast cells divided by the total number of blast cells in at least 10 fields and then expressed as a percentage. Scoring of the tissue microarray was completed independently by two clinically qualified pathologists who were blinded to the clinical information. We observed that LIN28B mainly localizes in nuclear of leukemia blast cells. We classified the LIN28B staining pattern as follows: negative, <20% leukemia cells with nuclear staining, 20%–<50% leukemia cells with nuclear staining, 50%–<75% leukemia cells with strong nuclear staining, and ≥75% leukemia cells with strong nuclear staining. The scoring results were shown in Supplementary Table S4, and representative LIN28B staining images were presented in Fig. 6A. Notably, the five normal bone marrow samples from healthy controls demonstrated very little LIN28B expression (Fig. 6A).
The survival data were available for 108 patients. We classified patients with LIN28B scoring 50% and above as the “LIN28B high” group and patients with negative or LIN28B scoring less than 50% as the “LIN28B low” group. The comparison of various clinical features between LIN28B high and low groups was summarized in Supplementary Table S5. We then compared event-free survival between these two groups of patients. On the basis of the Kaplan–Meier survival analysis, the group of patients who were LIN28B high had a much shorter event-free survival than patients who were LIN28B low (Fig. 6B, P = 0.0159). Multivariate analysis revealed that LIN28B protein expression was an independent prognostic factor of FLT3 status, NPM mutation, and cytogenetic risk. The HR of LIN28B expression was 3.96 (95% CI, 2.15–7.30; P < 0.001; Supplementary Table S6). Overall, high expression of LIN28B protein was a novel, independent predictor for poor survival in our cohort.
The phosphatase of regenerating liver (PRL) family contains three members, PRL-1, PRL-2, and PRL-3. Among the PRL family, PRL-3 has been most frequently studied in cancer, followed by PRL-2 and PRL-1. PRL-2 has been reported to increase EPO and IL3 response in hematopoietic cells (32) and play critical roles in regulating HSC self-renewal (33). A previous study reported that genetic disruption of PRL-3 reduces clonogenicity and growth of CD133+ mouse colon cancer cells in an in vitro culture system (34). CD133 is also a specific marker for human hematopoietic stem/progenitor cells (35, 36). The observations from these published studies indicate a possibility that aberrant expression of PRL family in hematopoietic cells may promote the transformation of LSC in AML. The association between PRL-3 expression and poor prognosis has been widely confirmed in a number of solid tumors and AML (10, 13, 15–17), arguing for the role of PRL-3 as a biomarker or a therapeutic target in stratifying patients for personalized therapy (9, 33, 37, 38). Thus, it is of clinical importance to further understand the molecular mechanism by which PRL-3 is regulated and contributes to leukemogenesis. In this study, we demonstrate that PRL-3 contributes to the leukemic phenotype of AML and plays an important role in the transformation of LSCs via the LIN28B/let-7 regulatory axis (graphic summary in Fig. 7).
Our global gene expression analysis showed the upregulation of several stemness factors, such as LIN28B, c-KIT, and IGF2BP1, and the downregulation of cell surface markers of more mature cells, such as CD36 and CD38, by PRL-3. c-KIT functions as an important regulator of HSC proliferation and self-renewal through binding to soluble steel factor (39). Gain-of-function mutations in c-KIT have been characterized in AML and other types of cancers (40). GSEA analysis further revealed that our gene expression profile significantly enriched with previously reported LSC and HSC gene signatures (41, 42), suggesting PRL-3 induces a stem cell-like transcriptional program in leukemia cells.
Serial transplantation in mouse model and serial replating assay provided functional evidence that PRL-3 confers stem cell–like properties in AML cells and mediates their transformation into leukemia-initiating cells. We further demonstrated that increased expression is dependent on the phosphatase activity of PRL-3, and LIN28B is essential for this transformation process. LIN28B is a stemness factor and an miRNA regulator, which inhibits the biogenesis of let-7 miRNA family (25, 26). A dysreguated LIN28B/let-7 circuit has been implicated in the development of many types of malignancies, including T-cell acute lymphoblastic leukemia (43), neuroblastoma (44), and liver cancer (45). Interestingly, MUC1-C, a transmembrane oncoprotein, induces activation of LIN28B and downregulation of let-7 in non–small cell lung cancer, thus stimulating EMT and stem cell self-renewal (46). Here, we demonstrate that the LIN28B/let-7 axis plays a key role in PRL-3–induced leukemia. The colony-forming capacity of leukemic cells is alleviated after knocking down LIN28B or upon ectopic expression of let-7 miRNAs, while overexpression of LIN28B partially rescues the decreased colony forming capacity caused by viral-mediated silencing of PRL-3 in serial replating assays. By studying a cohort of clinical samples, we showed the clinical significance of LIN28B expression in bone marrow samples of AML patients. The shared gene signatures of LSCs between our expression profile and others derived from clinical AML samples further underscores the clinical relevance of our findings (41, 42).
Leukemia-initiating or stem cells can arise from malignant transformation of normal hematopoietic stem cells or the deregulation of genes that induce stem cell–like properties in more mature progenitor cells. Our investigation has uncovered a novel and critical regulatory PRL-3/LIN28B/let-7 network, which plays an important role in transformation of LSCs and development of AML and identifies AML patients with poor outcomes. We have previously shown that PRL-3 is deregulated in myeloid malignancies by STAT activation by upstream oncogenic kinases, such as FLT3 (e.g., in FLT3-ITD AML; refs. 13, 21, 47). Our results therefore provide potential mechanistic insights to the high relapse rates and poor outcome of FLT3-ITD AML. This novel regulatory network may represent the “Achilles’ heel” of LSCs in AML, and its therapeutic disruption could lead to the elimination of LSCs in a molecular subset of AML in which PRL-3/LIN28B appears pivotal for the transformation.
Disclosure of Potential Conflicts of Interest
No potential conflicts of interest were disclosed.
Conception and design: J. Zhou, W.-J. Chng
Development of methodology: J. Zhou, J.-Y. Chooi, W.-J. Chng
Acquisition of data (provided animals, acquired and managed patients, provided facilities, etc.): J. Zhou, Z.-L. Chan, C. Bi, X. Lu, J.-Y. Chooi, S.-C. Liu, Y.Q. Ching, Y. Zhou, C.H. Ng, S.-B. Ng, S. Wang
Analysis and interpretation of data (e.g., statistical analysis, biostatistics, computational analysis): J. Zhou, Z.-L. Chan, C. Bi, J.-Y. Chooi, M. Osato, T.Z. Tan, W.-J. Chng
Writing, review, and/or revision of the manuscript: J. Zhou, W.-J. Chng
Administrative, technical, or material support (i.e., reporting or organizing data, constructing databases): C. Bi, P.S.Y. Chong, J.-Y. Chooi, L.-L. Cheong, Y.Q. Ching, Q. Zeng
Study supervision: W.-J. Chng
This work was supported by the Singapore National Research Foundation and the Ministry of Education under the Research Center of Excellence Program (to W.-J. Chng) and NMRC Clinician-Scientist IRG grant CNIG11nov38 (to J. Zhou). W.-J. Chng was also supported by NMRC Clinician Scientist Investigator Award.
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.
The authors thank the Singapore National Research Foundation and the Ministry of Education under the Research Center of Excellence Program and National Medical Research Council (NMRC) for financial support.
Note: Supplementary data for this article are available at Molecular Cancer Research Online (http://mcr.aacrjournals.org/).
- Received August 11, 2016.
- Revision received October 29, 2016.
- Accepted November 18, 2016.
- ©2016 American Association for Cancer Research.