Lysophosphatidic Acid Induces Erythropoiesis through Activating Lysophosphatidic Acid Receptor 3§


  • Chi-Ling Chiang,

    1. Institute of Zoology, National Taiwan University, Taipei, Taiwan, Republic of China
    Search for more papers by this author
  • Swey-Shen Alex Chen,

    1. Institute of Zoology, National Taiwan University, Taipei, Taiwan, Republic of China
    Search for more papers by this author
  • Shyh Jye Lee,

    1. Institute of Zoology, National Taiwan University, Taipei, Taiwan, Republic of China
    2. Department of Life Science, National Taiwan University, Taipei, Taiwan, Republic of China
    Search for more papers by this author
  • Ku-Chi Tsao,

    1. Institute of Zoology, National Taiwan University, Taipei, Taiwan, Republic of China
    Search for more papers by this author
  • Pei-Lun Chu,

    1. Department of Chemical Engineering and Materials Science, Yuan-Ze University, Chung-Li, Taiwan, Republic of China
    Search for more papers by this author
  • Cheng-Hao Wen,

    1. Bioresource Collection and Research Center, Food Industry Research and Development Institute, Hsinchu, Taiwan, Republic of China
    Search for more papers by this author
  • Shiaw-Min Hwang,

    1. Bioresource Collection and Research Center, Food Industry Research and Development Institute, Hsinchu, Taiwan, Republic of China
    Search for more papers by this author
  • Chao-Ling Yao,

    Corresponding author
    1. Department of Chemical Engineering and Materials Science, Yuan-Ze University, Chung-Li, Taiwan, Republic of China
    • Department of Chemical Engineering and Materials Science, Yuan Ze University, 135, Yuan-Tung road, Chung-Li city 32003, Taiwan, Republic of China
    Search for more papers by this author
    • Telephone: 886-3-4638800; Fax: 886-3-4559373

  • Hsinyu Lee

    Corresponding author
    1. Institute of Zoology, National Taiwan University, Taipei, Taiwan, Republic of China
    2. Department of Life Science, National Taiwan University, Taipei, Taiwan, Republic of China
    3. Center for Biotechnology, National Taiwan University, Taipei, Taiwan, Republic of China
    4. Angiogenesis Research Center, National Taiwan University, Taipei, Taiwan, Republic of China
    5. Research Center for Developmental Biology and Regenerative Medicine, National Taiwan University, Taipei, Taiwan, Republic of China
    • Department of Life Science and Institute of Zoology, National Taiwan University, 1 Roosevelt Road, Sec. 4, Taipei 106, Taiwan, Republic of China
    Search for more papers by this author
    • Telephone: 8862-3366-2499; Fax: 8862-2363-6837

  • Author contributions: C.-L.C.: designed the research, performed experiments, and wrote the manuscript; S.-S.A.C.: performed experiments; S.J.L.: designed and performed experiments; K.-C.T.: performed zebrafish experiments; P.-L.C. prepared human hematopoietic stem cells; C.-H.W., performed experiments; S.-M.H., provided stem cells; C.-L Y.: designed the research, prepared human hematopoietic stem cells, performed experiments, and wrote the manuscript; H.L. designed the research, and wrote the manuscript. C.-L Y. and H.L. contributed equally to this article.

  • Disclosure of potential conflicts of interest is found at the end of this article.

  • §

    First published online in STEM CELLSEXPRESS September 13, 2011.


Lysophosphatidic acid (LPA), an extracellular lipid mediator, exerts multiple bioactivities through activating G protein-coupled receptors. LPA receptor 3 (LPA3) is a member of the endothelial differentiation gene family, which regulates differentiation and development of the circulation system. However, the relationship among the LPA receptors (LPARs) and erythropoiesis is still not clear. In this study, we found that erythroblasts expressed both LPA1 and LPA3, and erythropoietic defects were observed in zLPA3 antisense morpholino oligonucleotide-injected zebrafish embryos. In human model, our results showed that LPA enhanced the erythropoiesis in the cord blood-derived human hematopoietic stem cells (hHSCs) with erythropoietin (EPO) addition in the plasma-free culture. When hHSCs were treated with Ki16425, an antagonist of LPA1 and LPA3, erythropoietic process of hHSCs was also blocked, as detected by mRNA and protein expressions of CD71 and GlyA. In the knockdown study, we further demonstrated that specific knockdown of LPA3, not LPA1, blocked the erythropoiesis. The translocation of β-catenin into the nucleus, a downstream response of LPAR activation, was blocked by Ki16425 treatment. In addition, upregulation of erythropoiesis by LPA was also blocked by quercetin, an inhibitor of the β-catenin/T-cell factor pathway. Furthermore, the enhancement of LPA on erythropoiesis was diminished by blocking c-Jun-activated kinase/signal transducer and activator of transcription and phosphatidylinositol 3-kinase/AKT activation, the downstream signaling pathways of EPO receptor, suggested that LPA might play a synergistic role with EPO to regulate erythropoietic process. In conclusion, we first reported that LPA participates in EPO-dependent erythropoiesis through activating LPA3. STEM CELLS 2011;29:1763–1773


Erythropoiesis, one of the main processes of hematopoiesis, depends on the unique ability of hematopoietic stem cells (HSCs) to proliferate and differentiate into progenitors of red blood cells. Recently, CD34+ stem cells derived from umbilical cord blood have provided a promising alternative to bone marrow for differentiation research [1]. In the multistep process of erythropoiesis, HSCs can be induced into progenitor cells and erythroblasts and terminally differentiate into red blood cells with nuclear condensation and extrusion [2]. The expansion and differentiation of HSCs were reported in liquid culture using a variety of recombinant human cytokine combinations with stem cell factor (SCF) and erythropoietin (EPO), which play essential roles in erythropoiesis [3]. Several signaling pathways are known to regulate hematopoiesis and erythropoiesis. The binding of EPO and its receptor (EPO receptor [EPOR]) activates c-Jun-activated kinase (JAK), which stimulates intracellular pathways such as, Ras/mitogen-activated protein kinase, phosphatidylinositol 3-kinase (PI3K), and signal transducer and activator of transcription (STAT) transcription factors, to regulate erythroid differentiation and survival. GATA-1 is an essential transcription factor modulated by the EPOR/PI3K/AKT pathway and it mediates specific target genes influencing the development of an erythroid progenitor in apoptosis, proliferation, and terminal differentiation [4]. In addition, in the early stages of erythropoiesis, Wnt/β-catenin pathways are also critical for determining the differentiation fate of CD34+ HSCs [5, 6]. Several studies on cancer cells suggested that activation of G protein-coupled receptors (GPCRs), including lysophosphatidic acid receptor 2 (LPA2) and LPA3, can turn on the Wnt/β-catenin pathway [7].

LPA is a simple phospholipid mediator which stimulates cell proliferation, migration, and survival of multiple cell types [8]. LPA receptors (LPARs), formerly designated endothelial differentiation gene receptors, are GPCRs and widely exist in tissues and exert diverse bioactivities. To date, up to five LPARs were identified [9]. Among these receptors, LPA1, LPA2, and LPA3 are widely expressed in most tissues and react with Gi, Gq, G12/13, or Gs [10]. LPA4 is expressed in specific organs, such as the pancreas, ovaries, and thymus; it raises intracellular Ca2+ and cAMP levels through activating Gq and G12/13 [11]. LPA5, expressed at low levels in multiple tissues, reacts with Gq and G12/13 and increases intracellular cAMP levels [12]. The specific functions of each receptor were elucidated by studies using LPAR gene-deficient mice. LPA1-deficient mice exhibited reduced suckling, attributed to olfactory defects, and developmental abnormalities in the neurological system, whereas LPA2 knockouts had no obvious phenotype [13]. In addition, LPA3-knockout mice had delayed uterine implantation, altered embryo spacing, and reduced litter sizes [14]. LPA4 deletion did not display a noticeable phenotype. However, embryonic fibroblasts from LPA4-knockout mouse responded with hypersensitivity to LPA-induced cell migration [15]. Recent studies also showed that stem cell differentiation is regulated through LPARs [16]. LPA accelerates proliferation and differentiation of human mast cells derived from cord blood through LPARs and peroxisome proliferator-activated receptor γ (PPARγ)–dependent pathways [17]. In addition, LPA induces osteoblastic differentiation from telomerase reverse transcriptase–overexpressed human mesenchymal stem cells through interplay of LPA1 and LPA4 [18]. In contrast, LPA inhibits neuronal differentiation of neural progenitor cells derived from human embryonic stem cells [19]. However, the roles of LPA during the erythropoiesis process remain unclear.

In an in vivo study, we used an antisense zLPA3-morpholino oligonucleotide (MO) to knockdown LPA3 in zebrafish, and hematopoietic defects were observed in MO-injected embryos. Furthermore, we cultured cord blood-derived human (h)HSCs in vitro and detected that LPA1 and LPA3 were expressed in the progenitor and erythroblast stages. Erythropoietic differentiation was abolished by Ki16425, a pharmacological blocker of LPA1 and LPA3, as evaluated by mRNA and protein expressions of CD71 and GlyA. Furthermore, using electrical nucleofection, we found that knockdown of LPA3, but not of LPA1, specifically inhibited the erythropoietic process. In addition, LPA also induced the expression of these erythroid biomarkers in cultured human hematopoietic stem cells (hHSCs) under plasma-free conditions. Translocation of β-catenin into the nucleus, a downstream response of LPAR activation, was blocked by Ki16425 treatment. Moreover, quercetin, an inhibitor of the β-catenin/T-cell factor (TCF) pathway, blocked LPA-enhanced hematopoiesis. In conclusion, our study indicated that LPA induced an EPO-dependent erythropoietic process through activating LPA3, and β-catenin translocation may participate in this process. This study may provide a novel treatment for promoting erythropoiesis.


Zebrafish Maintenance and Embryo Collection

Wild-type zebrafish (Danio rerio) obtained from the Zebrafish Information Research Center (Eugene, OR) were inbred at 27°C–28°C. Embryos were collected from natural spawning, washed, and incubated in 0.3× Danieau's buffer at 28°C until being used. Each stage of fish embryos was determined from their morphology. Embryos at 12 hours postfertilization (hpf) were treated with 0.2 mM 1-phenyl-2-thiourea (Sigma, St. Louis, MO) in 0.3× Danieau's buffer to suppress melanization. Dechorionation of embryos was carried out with 0.01 g/ml protease (Sigma).

Morpholino Antisense Oligonucleotides

Glass capillary tubes were pulled by a horizontal puller (P-97, Sutter Instrument, Navato, CA). Two zLPA3 translation-blocking (t)MOs (Gene Tools, Philomath, OR) were used as follows: tMO1: 5′-CACTGTCCTAGTGGGTTTTGTCCAA-3′ (−27 to −3) and tMO2: 5′-ATGGCCAGGCACAACATCTGCTA-3′ (+1 to +23). The MOs were prepared at 1 mM in sterile H2O as a stock solution, and further diluted into desired working concentrations in 1× Danieau's buffer with 0.5% phenol red. Fish embryos were fixed on 1% injection agar plates and 2.3 nl of solution was injected into the yolk sac and blastodisc interface with a nanoliter injector. (World Precision Instruments, Sarasota, FL). The injected embryos were incubated in 0.3× Danieau's buffer at 28.5°C for further experiments.

O-Dianisidine Blood Cell Staining

Embryos from designated stages were incubated in freshly prepared o-dianisidine solution (40% ethanol, 0.01 M sodium acetate, 0.65% H2O2, and 0.6 mg/ml o-dianisidine Σ) for 15 minutes in the dark. Embryos were observed under a stereomicroscope for red blood cells with red o-dianisidine staining.

Cultures of Hematopoietic Stem Cells

CD133+ stem cells were isolated from human umbilical cord blood by magnetic microbead selection using Direct CD34 progenitor isolation beads (Miltenyi Biotech, Bergisch Gladbach, Germany) and MACS LS-columns (Miltenyi Biotech), and were subsequently cultured. In stage I, 105 CD133+ cells were cultured in 10 ml of embryo development medium (EDM) (104 cells/ml), containing human SCF (50 ng/ml), human EPO (6 IU/ml), human interleukin (IL)-3 (10 ng/ml), human vascular endothelial growth factor (10 ng/ml), and human insulin-like growth factor-II (250 ng/ml) for 6 days. In stage II, cells were cultured in 10 ml of EDM (3 × 104 cells/ml) in the presence of SCF (50 ng/ml) and EPO (6 IU/ml) for 4 days. In stage III, cells were cultured in 10 ml of EDM (5 × 104 cells/ml) in the presence of SCF (50 ng/ml) and EPO (2 IU/ml) for 6 days. Cultured cells underwent 30-, 20-, and 10-fold expansions in stages I, II, and III, respectively. All cell cultures were incubated at 37°C in 5% CO2 under humidified conditions. LPA (Sigma) was prepared in chloroform and methanol (1:9) and stored at −20°C. LPA was added to serum-free medium with 0.005% fatty acid-free bovine serum albumin (BSA) as a carrier.

Treatment with Inhibitors

Ki16425 (Cayman Chemical, Ann Arbor, MI) powder was diluted in dimethyl sulfoxide (DMSO) at a concentration of 40 mM as the stock. At the end of stages I and II, HSCs were pretreated with Ki16425 in serum-free medium for 1 hour. Then, cells were cultured in media of stages II and III containing Ki16425 at a final concentration of 20 μM. Quercetin (Sigma), LY294002 (Sigma), and WP1066 (Merck, Germany) was dissolved in DMSO at proper concentration as the stock and stored at −20°C. Each inhibitor was diluted into each final concentration by culture media for treatment.

ShRNA Nucleofection

Approximately 5 × 105 HSCs at the ends of stages I and II were prepared for transfection. The short-hairpin RNA (shRNA) plasmid against LPA1 and LPA3 was transiently transfected into HSCs using Amaxa nucleofection (Lonza, Cologne, Germany) according to the manufacturer's protocol for CD34+ stem cells. Each target plasmid was constructed with a green fluorescent protein (GFP) gene to confirm that the transfection efficiency was >70%. After 48 hours, GFP-positive cells were sorted out by BD FACSAria II Cell Sorter (BD Bioscience, San Diego, CA) and cultured in media for further stages.

mRNA Expression

Total RNA extraction was isolated using Trizol Reagent (Invitrogen, Carlsbad, CA, USA). Complementary (c)DNA was synthesized from 1 μg total RNA with a reverse-transcription polymerase chain reaction (RT-PCR). Primer sequences were: GATA-1, 5′-CAAGAAGCGCCTGATTGTCAG-3′ (forward) and 5′-AGTG TCGTGGTGGTCGTCTG-3′ (reverse); LPA1, 5′-TGTCTCGGCA TAGTTCTGGACCC-3′ (forward) and 5′-CATTTCTTTGTCGC GGTAGGAG-3′ (reverse); LPA3, 5′-TGACTGTCTTAGGGGC GTTT-3′ (forward) and 5′-TTCTCCTGAGAGAAGCAGCA-3′ (reverse); CD71, 5′-GGATAAAGCGGTTCTTGGTACC A-3′ (forward), 5′-CCAGTAACCGGATGCTTCACA-3′ (reverse); and GlyA, 5′-ACAGACAAAT GATACGCACAAACGGG-3′ (forward) and 5′-GGGCTTTTCTTTATCAGTCGGCGA-3′ (reverse).

A real-time PCR with the mixture reagent SYBR-Green I (Thermo Scientific, San Diego, CA) was carried out on an iCycler iQ real-time detection system (Bio-Rad, Hercules, CA). The specificity of the primers was confirmed from a single peak of the melting curve. Each target mRNA level was evaluated from the real-time threshold cycle and compared with the glyceraldehyde 3-phosphate dehydrogenase amount as an internal control.

Colony-Forming Cell Assay

After stage I expanding culture, cells were plated in semisolid culture (MethoCult GF SF H4236, StemCell Technologies) following the manufacturer's instruction for colony-forming unit assay. The cells were seeded at suitable concentration (to give approximately 100 colonies per 1 mL culture) in the methylcellulose-based media with (or without) addition of cytokines (50 ng/ml SCF and 6 IU/ml EPO) and (or) LPA (5 μM). Methylcellulose-based media and cells were aliquotted in 35 mm petri dishes and incubated at 37°C in an atmosphere of 5% CO2 and humidified incubator. After 14 days of culture, burst forming unit-erythroid (BFU-E) and colony-forming unit-erythroid (CFU-E) were scored under inverted microscope.

Immunofluorescence Analysis

Harvested cells were fixed on cover slides in phosphate-buffered saline (PBS) containing 4% formaldehyde for 10 minutes and permeated through both the plasma and nuclear membrane by 0.02% Triton X-100 for 10 minutes. After blocking with 5% BSA for 1 hours, samples were incubated with mouse immunoglobulin (IgG) or anti-β-catenin primary antibody (BD Bioscience) overnight. Finally, cells were stained with 4,6-diamidino-2-phenylindole (DAPI, Sigma) for 5 minutes and a goat anti-mouse IgG FITC secondary antibody for 1 hour. Slides were mounted with Fluoromount-GTM (Emsdiasum, Fort Washington, PA) and visualized using a DeltaVision OMX Core fluorescence microscope (Applied Precision, Issaquah, WA).

Flow Cytometry

HSC cultures were collected, washed, and stained either with anti-GlyA-PE and anti-CD71-PE/Cy5 for 30 minutes (BD Biosciences Pharmingen, San Diego, CA). Stained cells were washed and resuspended in PBS. Nonstained and single-stained samples were prepared for fluorescent compensation. For all experiments, analysis of cells used a FACSCalibur instrument (BD Biosciences) and FCS Express software (De Novo, Los Angeles, CA).

Statistical Analysis

Experiments were assessed using Excel computer software. Differences between groups were assessed using a nonparametric analysis of variance. p <.05 was considered significant.


zLPA3 tMO Affects Red Blood Cell Formation in Zebrafish

To investigate the effects of zLPA3, two translational-blocking MOs, tMO1, and tMO2 were synthesized and injected into one-cell stage zebrafish embryos. The specificity, potency, and efficacy of tMO1 and tMO2 in blocking zLPA3 translation were confirmed (Supporting Information Fig. S1). By using o-dianisidine, which binds erythrocyte hemoglobin, we found that LPA3 knockdown caused erythropoietic defects. O-dianisidine staining was observed in the yolk sphere of control embryos, but had notably decreased in LPA3 knockdown embryos at 48 hpf. The degrees of erythrocyte formation were classified into normal, moderate, and severe conditions. A dose-dependent inhibition on erythropoiesis was observed in both zLPA3 tMO1 and tMO2-injected embryos at 0, 2.5, and 5 ng (ncontrol = 295, ntMO1, 2.5 ng = 86, ntMO1, 5 ng = 89; ncontrol = 86, ntMO2, 2.5 ng = 86, ntMO2, 5 ng = 89) (Fig. 1). In addition, similar erythropoiesis defect was observed in 60 hpf embryos (Supporting Information Fig. S2).

Figure 1.

Hematopoietic defects in zLPA3-knockdown zebrafish. Specific knockdown of zLPA3 inhibited the hematopoietic development in a dose-dependent manner. Zebrafish embryos were injected with 2.5–5 ng translation-blocking morpholino oligonucleotide (tMO), cultured, and photographed at 48 hours postfertilization. The amounts of red blood cells were measured by staining of hemoglobin with o-dianisidine. (A): Zebrafish embryos were classified by characterization of the embryonic blood phenotype into three groups: normal, moderate, and severe. Moderate embryos manifested slight hematopoietic defects; while severe group lacked hemoglobinized erythrocytes (B): The quantitative plots demonstrated the statistic results of embryos injected with zLPA3 tMO1 and tMO2. The numbers of moderate and severe phenotype raised in the embryos injected with tMO1and tMO2. The hematopoietic defects were significantly higher in zLPA3-knockdown zebrafish (ncontrol = 295, ntMO1, 2.5 ng = 86, ntMO1, 5 ng = 89; ncontrol = 86, ntMO2, 2.5 ng = 86, ntMO2, 5 ng = 89). Mean ± SEM; *p < 0.05, **p < 0.01, ***p < 0.001. Abbreviation: tMO1, translation-blocking morpholino oligonucleotide 1.

In Vitro Hematopoiesis Using Human CD133+ Stem Cells

To confirm the observation that LPA3 may be involved in regulating hematopoiesis, we established an erythropoiesis system using cord-blood-derived hHSCs as described above (Fig. 2A). In these procedures, human CD133+ stem cells were induced by growth factors to become mature erythrocytes. Human HSCs underwent both proliferation and differentiation in four stages (Fig. 2A). In stage I, hHSCs were expanded by approximately 30-fold into myeloid progenitors. Sequentially, progenitor cells were induced to go through erythropoietic differentiation by SCF and EPO treatments during stages II and III. Progenitor cells were further expanded by 20- and 10-fold, and generated a greater number of erythroid cells expressing specific biomarkers: CD71 and GlyA. In stage IV, erythroblasts were stimulated to enucleate and become mature erythrocytes [3]. The expression patterns of LPA1 and LPA3 during different stages were determined by a real-time PCR. The mRNAs of LPA1 and LPA3 were detected in both stages II and III. In contrast, expression levels of LPA1 and LPA3 were higher at the end of stage I and decreased as cells entered stages II and III (Fig. 2B).

Figure 2.

In vitro hematopoiesis using CD133+ human hematopoietic stem cells (hHSCs). Human CD133+ HSCs were isolated from umbilical cord blood and induced into mature anucleate erythrocytes by four-stage process. (A): Human HSCs underwent both proliferation and differentiation in this process. In stage I, hHSCs were expanded approximately 30-fold into progenitor cells. Subsequently, erythroid progenitor cells were stimulated to go through erythropoietic differentiation by stem cell factor and erythropoietin treatments during stages II and III. The progenitors expanded individually by 20- and 10-fold in stages II and III. (B): The reverse-transcription polymerase chain reaction analysis of mRNA encoding lysophosphatidic acid receptor 1 (LPA1), LPA3 from hHSCs developed from stages II and III. Expression of LPA1 (black) and LPA3 (gray) were varied with the different stages of hematopoietic differentiation. Abbreviations: hHSC, human hematopoietic stem cell; LPA1/3, lysophosphatidic acid receptor 1/3.

LPAR Antagonist Blocks Erythropoietic Processes

To investigate the roles of LPA1 and LPA3 in erythropoietic differentiation, we used Ki16425, an LPA1- and LPA3-specific antagonist, to block receptor-activated signaling processes. Cells harvested from the end of stage I were incubated in serum-free media with DMSO or Ki16425 for 1 hour and allowed to develop to stage II differentiation. In stages II and III, the expansion of erythroblasts was significantly lower under Ki16425 treatment compared with the control group (Fig. 3A). In addition, relative mRNA levels of CD71 and GlyA both decreased with Ki16425 treatment (Fig. 3B). The long-term effects of LPA1 and LPA3 inhibition with Ki16425 treatment were also monitored. Human HSCs were treated with Ki16425 in stages II, III, or both. Treated cells were harvested and analyzed after 10 days. The mRNA expression levels of CD71 and GlyA were affected under constant Ki16425 treatment (Fig. 3C). Ki16425 treatment in stage II only caused greater inhibition than in stage III, especially in terms of GlyA mRNA levels. Furthermore, Ki16425's effects on CD71 and GlyA protein levels were also monitored by flow cytometry. In stage III, high percentages of cells exhibiting surface CD71 and GlyA were observed after EPO treatment. However, with Ki16425, suppression of the expression of these two markers was observed. More than half of Ki16425-treated cells were trapped in an undifferentiated stage (CD71 GlyA population). Our data indicated that blocking LPA1 and LPA3 signaling inhibited the differentiation of erythrocytes.

Figure 3.

Erythropoietic inhibition caused by Ki16425, an antagonist of lysophosphatidic acid receptor 1 (LPA1) and LPA3. After the expansion of stage I, human hematopoietic stem cells were maintained in culture medium with dimethyl sulfoxide (DMSO) control or Ki16425 in stages II and III. (A): Seeding density for initiation in stages II and III contained 3 × 105 and 5 × 105 per 10 ml. The cell number at each stage was estimated by Trypan-blue staining and monitored by hemocytometer. Ki16425 treatment significantly inhibited the cell expansion in both stages II and III. (n = 3) (B): Human HSCs were harvested from stages II (black) and III (gray), and mRNA expressions of erythropoietic biomarkers, CD71 and GlyA, were measured by a real-time polymerase chain reaction. In stage II, blocking of LPA receptor signaling by Ki16425 reduced the mRNA level of CD71 and GlyA. Otherwise, continuous Ki16425 treatment from stages II to III demonstrated significant inhibition of erythropoietic differentiation. (n = 4) (C): Cells treated with the DMSO control (i), Ki treatment in stage II (ii), stage III (iii), and continually in stages II and III (iv) were stained with CD71 and GlyA monoclonal antibodies. The flow cytometric data were analyzed in quadrants. With continuous Ki16425 treatment from stages II to III, the amount of CD71+GlyA+ erythroid was substantially decreased. (n = 4). Mean ± SEM; *p < 0.05, **p < 0.01.

LPAR Knockdown Prevents Erythropoietic Processes

As Ki16425 blocks both LPA1 and LPA3, we further attempted to clarify which receptor is responsible for regulating erythropoiesis. Plasmids containing LPA1 and LPA3 shRNA or a scrambled sequence were transfected into hHSCs by electrical nucleofection procedures. After 24 hours, GFP-positive cells were isolated for further studies. The knockdown efficacies of LPA1 and LPA3 were individually determined (Fig. 4A). Messenger RNA levels of CD71 and GlyA were lower in LPA3-knockdown cells in stages II and III, compared with control and LPA1-knockdown cells (Fig. 4B). Similar results were observed for protein levels. LPA3-knockdown cells had the lowest percentage of CD71+ and GlyA+ double-positive erythroblasts, whereas LPA1-knockdown cells showed similar patterns as the control (Fig. 4C). We concluded that LPA-regulated erythropoiesis is mainly mediated through activating LPA3.

Figure 4.

Lysophosphatidic acid (LPA) receptors knockdown in erythropoietic processes. Human HSCs were transfected with green fluorescent protein (GFP) vectors of scrambled, LPA1, and LPA3 short-hairpin RNA (shRNA). After 24 hours, GFP-positive cells were isolated by flow sorter. (A): The knockdown efficacy of LPA1 and LPA3 shRNA was individually determined by a real-time polymerase chain reaction. (n = 3) (B): Messenger RNA levels of CD71 and GlyA in LPA1- and LPA3-knockdown cells were detected in stages II (black) and III (gray). In LPA3-knockdown cells, mRNA transcription of CD71 and GlyA was obviously decreased in both stages II and III. (C): CD71 and GlyA expressions of LPA1- and LPA3-knockdown cells were analyzed by flow cytometry in stage III. In comparison with scramble and LPA1 shRNA-transfected cells, the CD71+GlyA+ population in LPA3-knockdown erythroids was evidently diminished. (n = 3). Mean ± SEM; *p < 0.05, **p < 0.01, ***p < 0.001. Abbreviations: LPA1/3, lysophosphatidic acid receptor 1/3; shRNA, short-hairpin RNA.

Exogenously Added LPA Enhances Erythrocyte Differentiation

We further determined the direct effects of LPA on the development of hHSCs. A culture of cord blood progenitor cells was established in plasma-free medium to exclude the effect of endogenous serum-derived LPA. These cells were then cultured in the absence or presence of LPA 18:1 (1–10 μM) for 24 hours. Short-term culture under plasma-free conditions showed no obvious differences from the control. However, long-term plasma-free cultured cells had a lower survival rate and were poorly differentiated (data not shown). In stage II, 24-hour LPA treatment elevated mRNA levels of CD71, GlyA, and GATA-1 in concentration-dependent manners (Fig. 5A). At the end of stage II induction, protein expression levels were analyzed at 4 days using anti-CD71 PE.Cy5 and anti-GlyA PE. Cells treated with 5 and 10 μM LPA expressed higher levels of CD71, but not GlyA (Fig. 5B). We also tried to replace EPO effects in differentiation by LPA treatment alone. However, the differentiation of erythroid cells severely decreased in the absence of EPO, even in the presence of LPA (data not shown). These results suggest that LPA might enhance the EPO-dependent erythropoietic process.

Figure 5.

Enhancement of erythrocyte differentiation by lysophosphatidic acid (LPA) treatment. To avoid the effect of endogenous LPA in plasma, human hematopoietic stem cells (hHSCs) were cultured in plasma-free (PF) medium for LPA treatment. (A): In initial stage of erythropoietic differentiation, 24 hours of LPA treatment elevated mRNA levels of CD71, GlyA, and GATA-1. The enhancement of LPA on erythropoietic differentiation demonstrated concentration-dependent manners. (n = 3) (B): At the end of stage II induction, expression of CD71 protein levels was evaluated by the intensity of anti-CD71-Cy7. Comparing with the control (with plasma) and PF medium, LPA induced CD71 expression in dose-dependent manners. (n = 3) (C): Human HSCs were seeded in the Methylcellulose-based media with (or without) addition of cytokines (50 ng/ml stem cell factor and 6 IU/ml erythropoietin) and LPA (5 μM). After 14 days of culture, burst forming unit-erythroid and colony-forming unit-erythroid were estimated under microscope. Under LPA stimulation, hHSCs performed higher ability of colony formation in erythropoietic differentiation. (n = 6). Mean ± SEM; *p < 0.05, **p < 0.01, ***p < 0.001. Abbreviations: C, control; LPA lysophosphatidic acid; PF, plasma-free medium.

CFU-E is a necessary stage of erythroid development before proerythroblast stage, and colony formation assay is a general indicator for differentiation capacity of erythroid lineage in hHSCs. When comparing with control cultured condition, LPA treatments enhanced the differentiation activity and also potentiated erythroid colony formation. However, when comparing with cytokine stimulation, LPA treatment did not significantly increase the colony formation (Fig. 5C).

Regulation in the Early-Stage of Erythropoiesis

Translocation of β-catenin is known as a downstream signaling process activated by LPARs in cancer cells. We cultured hHSCs in the absence and presence of Ki16425 for 24 hours and then stained cells with an anti-β-catenin antibody and DAPI. Translocation of β-catenin into nuclei was obvious in early erythroid cells, while this was inhibited by Ki16425 (Fig. 6). The results suggest that the β-catenin/TCF pathway is involved in the early-stage of hematopoietic differentiation and inhibition by Ki16425 may occur through suppressing β-catenin translocation.

Figure 6.

Translocation of β-catenin in hematopoietic differentiation. The β-catenin was labeled with FITC, and the nuclear location was indicated by 4,6-diamidino-2-phenylindole. The merged image demonstrates the translocation of β-catenin in human hematopoietic stem cells (hHSCs) with/without Ki16425 treatment for 24 hours in stage II. (A): In end of stage I, most of β-catenin was located in cytosol of hHSCs. (B): As the hHSCs went through the erythropoietic differentiation in stage II, cytosolic β-catenin translocated into nuclear to trigger downstream signaling. (C): As hHSCs were treated by Ki16425 in stage II, translocation of β-catenin was inhibited. (n = 3). Abbreviation: DAPI, 4,6-diamidino-2-phenylindole.

To clarify the role of β-catenin in erythropoiesis, we used quercetin, an inhibitor of the β-catenin/TCF pathway, to treat HSCs in the early-stage. Because function of CD71 is essential for iron transport in early erythroid cells, we focused on CD71 expression. The increase in CD71 mRNA under LPA induction was suppressed by quercetin at different concentrations of 5, 20, and 50 μM, and similar pattern was also observed in the GATA-1 mRNA level (Fig. 7A). The protein expression of CD71 was detected by flow cytometry and the induction of LPA decreased under quercetin treatment (Fig. 7B). In addition, our results also demonstrated that Ki16425 treatment suppressed LPA enhancements for erythropoiesis (Fig. 7A).

Figure 7.

Suppression of lysophosphatidic acid (LPA)-enhanced differentiation by pharmalogical inhibitors. (A): Human HSCs were cultured in plasma-free (PF) medium with (or without) LPA (5 μM) added and treated with quercetin for 24 hours. The mRNA level of both CD71 and GATA-1 was increased with culture of LPA, but was inhibited by quercetin and Ki16425 treatment. Otherwise, quercetin (50 μM) did not suppress the basal mRNA level of CD71 and GATA-1 in plasma-free control without LPA stimulation. (n = 3). (B): At the end of stage II induction, the enhancement of CD71 protein level was suppressed by quercetin (20 μM). (n = 3). (C): Human HSCs were culture with the inhibitors of erythropoietin receptor signaling pathways: WP1066 (5 μM) and LY294002 (50 μM). WP1066 is an antagonist of c-Jun-activated kinase/signal transducer and activator of transcription pathway, and LY294002 is against phosphatidylinositol 3-kinase/AKT. At the end of stage II induction, the LPA-enhanced differentiation was inhibited by WP1066 and LY294002. (n = 3). Mean ± SEM; *p < 0.05, **p < 0.01. Abbreviations: LPA, lysophosphatidic acid; LY, LY294002; PF, plasma-free medium; WP, WP1066.

Furthermore, to clarify the role of EPO in LPA-induced erythropoiesis, we used pharmalogical inhibitors to block EPOR-activated signaling pathways, including JAK/STAT and PI3K/AKT. From flow cytometry analysis, the enhancement of LPA on CD71 expression was diminished by WP1066 treatment, an inhibitor for JAK/STAT pathway. The expression level of CD71 is lower than control as we blocked the PI3K/AKT signaling pathway by LY294002 (Fig. 7C). These results suggested that LPA might play a synergistic role with EPO to enhance erythropoietic process.


Lysophospholipids, including LPA and sphingosine-1-phosphate (S1P), were recently demonstrated to be regulators of cell fate determination in a variety of stem cells and their progenitors. LPA and S1P were shown to modulate proliferation, survival, differentiation, and migration of embryonic and neural stem cells [20]. In mouse hematopoietic progenitors, LPA1, LPA2, and S1P1–4, but not LPA3 or S1P5, were expressed in primitive Lin-Sca+ Kit+ cells isolated from bone marrow. In addition, LPA and S1P enhanced the chemotactic response in primitive HSCs stimulated by stromal-derived factor-1 [21]. However, in human hematopoietic progenitor cells, only a few studies demonstrated the effects of LPA. Our major contribution in this study was to clarify the regulatory roles of LPA in erythropoietic processes. Generally, in physiological conditions, LPA is constitutively present in human plasma at concentrations at 1–15 μM [22]. In our previous observation, the survival rate of hHSCs was significantly lower under long-term culture conditions without human plasma (Supporting Information Fig. S1). Furthermore, as we used Ki126425 to block the LPAR signaling, the erythroid differentiation and proliferation were significantly inhibited (Fig. 3). These results suggested that LPA present in human plasma may play a critical role in both the survival of HSCs and their differentiation into mature erythrocytes. Then we induced exogenous LPA into plasma-free culture system. We observed that exogenously added LPA increased the expression of CD71 and GlyA in hHSCs. In addition, expression level of GATA-1, one of the essential mediators of the EPO receptor (EPOR) that affected and promoted erythroid maturation, was also enhanced by LPA treatment (Fig. 5) [23]. Our studies strongly indicated that LPA might act as an enhancer of erythropoietic processes.

LPARs were G protein-couple receptors and formerly considered to regulate endothelial and hematological differentiation. More than five LPARs were identified with their multiple bioactivities [9]. In the previous literatures, LPA inhibits cell adhesion of the TF-1 erythroblastic progenitor cell line through a Rho-dependent pathway presumably via G12/13. However, the actual receptors involved were not determined [24]. Furthermore, LPA stimulation accelerated the development of mast cells derived from human umbilical cord blood. This stimulation was mediated through LPAR and PPARγ-dependent pathways to enhance proliferation and differentiation of human mast cells [17]. In addition, LPA was also reported to enhance osteogenic differentiation of human mesenchymal stem cells. LPA-induced osteogenic differentiation is controlled by the interplay between LPA1 and LPA4: activation of LPA1 leads to an increase in intracellular Ca2+ and induces osteogenesis, whereas activation of LPA4 results in increased cAMP production and inhibits osteogenic differentiation [18]. However, the detail regulative mechanisms of LPARs in stem cell differentiation remain unclear, especially in HSCs. In initiation of this study, we observed that LPA3-knockdown zebrafish embryos showed severe erythropoiesis defects (Fig. 1). In human model, we clearly demonstrated that erythropoiesis from hHSCs was affected by blocking of LPA3 signaling (Figs. 3 and 4). These results provided a novel model for LPA and its receptors in regulating erythroid differentiation. It is helpful to further clarify the roles of LPA in stem cell differentiation.

In the present studies of HSCs, Wnt/β-catenin signaling was suggested to be critical for cellular survival and development [25, 26]. In stem cells, Wnt signaling pathways have diverse functions [27]. Wnt10b has an inhibitory effect on BFU-E differentiation. In contrast, Wnt5a and Wnt2b act as activators of this process [5]. β-Catenin is a crucial factor in maintaining murine embryonic stem cells in an undifferentiated status. Activation of β-catenin promotes the growth of HSCs in vitro and maintains the self-renewal of HSCs under long-term culture [28]. In Xenopus embryos, Wnt4/β-catenin expressed in the mesoderm of the ventral blood island is essential for the expression of hematopoietic and erythroid marker genes [29]. In addition, constitutive β-catenin activation caused defects of HSCs and blocked multilineage differentiation in mouse blood cells [30]. In contrast, another report suggested that inactivation of the β-catenin gene is dispensable for hematopoiesis and lymphopoiesis in bone-marrow progenitors [31]. Furthermore, the β-catenin pathway is activated by G protein-coupled LPARs, which affect several cellular functions [32]. In this study, we confirmed that activation of LPAR induced β-catenin activation at the early-stage of erythropoiesis (Fig. 6). Erythropoiesis enhanced by LPA was blocked by quercetin treatment (Fig. 7). These results suggest the role of LPARs and β-catenin in human erythropoiesis and translocation of β-catenin is important for stem cell differentiation. However, quercetin also blocks the PI3K/AKT pathway which modulates the early-stage of EPO-dependent erythroid development [33, 34]. In addition, activation of LPA3 has been suggested to regulate proliferation and migration in several cell types via PI3K/AKT pathway [35, 36]. Therefore, details signaling of how LPARs regulate early-stage of erythropoiesis though β-catenin/TCF, PI3K/AKT, and JAK/STAT pathway remain to be determined.

It was known that EPO is the main regulator of erythropoietic processes, and the EPOR is primarily expressed in hematopoietic progenitor cells. Upon EPO binding, dimerization of EPOR activates the JAK/STAT signaling pathway to modulate erythropoiesis. In addition, the GATA-1 transcription factor is also essential for normal erythroid cell development [37, 38]. DNA binding and transactivation of GATA-1 are required for both primitive and definitive stages of hematopoiesis in zebrafish [39]. Several studies pointed out that the interaction between the GATA-1 and EPOR is responsible for the stability of erythropoiesis [40]. GATA-1 transactivated the gene expression of the EPOR and GATA-1 itself to regulate erythroid differentiation [41]. In fetal liver erythroid progenitor cells, EPOR signaling stimulated phosphorylation of GATA-1 to regulate the maturation of erythroids [4]. Our results (Fig. 5) showed that LPA enhanced the expression of the erythroid biomarkers, CD71 and GlyA, during erythroid differentiation. In addition, GATA-1 gene expression was also increased under LPA treatment. These results demonstrated an enhancing effect of LPA on maturation of erythroid cells in our system. We also tried to culture hHSCs in an EPO-free condition to clarify the effect of LPA alone. However, even in the presence of LPA in culture, the differentiation of erythroid cells was severely hampered (data not shown). Furthermore, as we blocked the EPOR-relative signaling pathway: PI3K/AKT and JAK/STAT, the LPA-induced enhancement of erythropoietic differentiation was also diminished. These results suggested that LPA induction of erythropoiesis may depend on the presence of EPOR signaling, which activates the transcription factor, GATA-1, to trigger expression of other essential proteins for erythroid differentiation.

In clinical, EPO is a major treatment for anemia to restore the production of erythrocytes and hemoglobin. Anemia patients are usually injected with EPO two or three times a week. One of the common side effects of EPO treatment is hypertension observed in anemic hemodialysis patients [42]. In addition, recent reports suggested that treatment with erythropoiesis-stimulating agents may contribute to tumor progression [43–45]. In in vitro studies, EPO and activation of the EPOR enhanced cell proliferation, survival, and growth of several human cancer cell lines [46–48]. In a breast cancer cell line, EPO activated the JAK2/STAT5, PI3K/AKT, and Ras/ERK pathways and promoted malignant cell behaviors [49]. Furthermore, clinical therapies with recombinant human EPO also affected the sensitivity of radiotherapeutic and chemotherapeutic outcomes in cancer patients.5050 Therefore, the safety of long-term EPO treatment remains uncertain, and the influence of EPO reagents on cardiovascular side effects and tumor progression needs to be determined. Our results suggest that in the presence of EPO, LPA may enhance the erythropoietic process. Therefore, we propose that with appropriate activation of LPA signaling, erythroid differentiation may take place with a lower dependence on EPO. We also speculated that this application may provide a novel clinical alternative to decrease the requirement for EPO and lower the risks of side effects from long-term EPO therapy.


This work was supported by National Taiwan University, the grants NSC 99-2120-M-002-004, NSC100-2325-B-002-045, and NSC97-2311-B-002-002-MY3 from the National Science Council, Taiwan, Republic of China (to H.L.). The human hematopoietic stem cell culture system was supported by the grant NSC99-2221-E-155-086-MY2 from the National Science Council, Taiwan, Republic of China (to C.-L.Y.).


The authors indicate no potential conflicts of interest.