|Home | About | Journals | Submit | Contact Us | Français|
Hematopoietic stem cell (HSC) self-renewal is regulated by both intrinsic and extrinsic signals. Although some of the pathways that regulate HSC self-renewal have been uncovered, it remains largely unknown whether these pathways can be triggered by deliverable growth factors to induce HSC growth or regeneration. Here we show that pleiotrophin, a neurite outgrowth factor with no known function in hematopoiesis, efficiently promotes HSC expansion in vitro and HSC regeneration in vivo. Treatment of mouse bone marrow HSCs with pleiotrophin caused a marked increase in long-term repopulating HSC counts in culture, as measured in competitive repopulating assays. Treatment of human cord blood CD34+CDCD38−Lin− cells with pleiotrophin also substantially increased severe combined immunodeficient (SCID)-repopulating cell counts in culture, compared to input and cytokine-treated cultures. Systemic administration of pleiotrophin to irradiated mice caused a pronounced expansion of bone marrow stem and progenitor cells in vivo, indicating that pleiotrophin is a regenerative growth factor for HSCs. Mechanistically, pleiotrophin activated phosphoinositide 3-kinase (PI3K) signaling in HSCs; antagonism of PI3K or Notch signaling inhibited pleiotrophin-mediated expansion of HSCs in culture. We identify the secreted growth factor pleiotrophin as a new regulator of both HSC expansion and regeneration
HSCs possess the unique capacity to self-renew and give rise to all of the mature cell types within the blood and immune systems1–3. HSC self-renewal is regulated by both intrinsic and extrinsic signals1–9, but the mechanisms involved in the control of this process are incompletely understood. Several growth factors have been identified whose action is associated with mouse. HSC self renewal, including Notch ligands4,5, Wnt3a6, angiopoietin-like proteins8 and prostaglandin E2 (ref. 9). Co-culture of HSCs with supportive stromal or endothelial cells10,11 or the enforced expression of the transcription factors homeobox protein B4 or homeobox protein A9 (refs. 1,12) can also cause robust expansions of HSCs in culture. However, strategies that require cell co-culture or genetic modification of HSCs are not readily translatable into the clinic13. Moreover, despite advances in understanding the biology of HSC self-renewal and differentiation, it has not yet been possible to induce HSC expansion or regeneration in a manner that can be translated to clinical practice. Here, we describe pleiotrophin, an18-kDa heparin-binding growth factor that is mitogenic for neurons17–19, has angiogenic and proto-oncogene activity30,31,32 as a potent, secreted regulator of HSC expansion and regeneration.
We have previously shown that human endothelial cells isolated from adults support the expansion of human HSCs in culture14,15. In contrast to co-culture studies with stromal cells16, which have a demonstrated requirement for cell-to-cell contact for HSC maintenance, we have shown that primary human brain endothelial cells (HUBECs) produce a soluble activity capable of inducing a ten-fold expansion of human HSCs ex vivo11,14. To identify the HUBEC-secreted proteins responsible for this HSC-amplifying activity, we performed genome-wide expression analysis of HUBECs as compared to nonbrain human endothelial cells that lack HSC-supportive activity (Fig. 1a). We identified 13 genes that were more than five-fold overexpressed in HUBECs and that were predicted to produce secreted gene products (Supplementary Table 1). We found that the expression of pleiotrophin, a heparin-binding growth factor with no known role in hematopoiesis17, was 25-fold higher in HUBECs versus nonbrain endothelial cells (Fig. 1b). Quantitative RT-PCR confirmed a >100-fold higher pleiotrophin expression in HUBECs, and analysis of HUBEC-conditioned medium by ELISA showed a higher concentration of pleiotrophin compared to nonbrain endothelial cell–conditioned medium (6.9 ± 0.3 pg ml−1 compared to 0.02 ± 0.01 pg ml−1, Fig. 1b).
We next examined whether bone marrow progenitor cells express one or more of the pleiotrophin receptors, receptor protein tyrosine phosphatase-β/ζ (RPTP-β/ζ), syndecan or anaplastic lymphoma kinase18,19. Both bone marrow mononuclear cells and bone marrow c-Kit+Sca-1+Lin− (KSL) progenitor cells expressed RPTP-β/ζ (n = 3 mice, mean 87.0% ± 8.8% and 89% of cells stain positive, respectively; Fig. 1c), whereas neither population expressed syndecan or anaplastic lymphoma kinase (data not shown). We next isolated bone marrow CD34−KSL cells from C57BL/6 mice by FACS; CD34−KSL cells are highly enriched for HSCs3,20. We placed the cells in liquid suspension culture with 20 ng ml−1 thrombopoietin, 120 ng ml−1 Stem Cell Factor (SCF) and 50 ng ml−1 Flt-3 ligand (TSF) without pleiotrophin or with 10, 100 or 1,000 ng ml−1 pleiotrophin for 7 days. We observed a dose-dependent increase in the number of total cells, the percentage of KSL cells and the number of total KSL cells in response to the addition of 10–100 ng ml−1 of pleiotrophin (Fig. 1d). In this experiment, the addition of 100 ng ml−1 pleiotrophin caused a 6.4-fold increase in total cell counts and a 17.7-fold increase in total KSL cell counts compared to TSF alone (P = 0.006 and P = 0.006, respectively, Fig. 1d). Similarly, cell cultures with TSF plus pleiotrophin contained twofold more ‘side population’ cells, which are highly enriched for primitive HSCs21, compared to cell cultures with TSF alone (mean 0.02% versus 0.01%, n = 4, P = 0.03, Fig. 1e). A 14-d replating study confirmed that pleiotrophin induced the expansion of CD34−KSL cells; at day 14, cell cultures with TSF plus pleiotrophin contained 3.2-fold more CD34−KSL cells compared to input (day 0) bone marrow CD34−KSL cells, whereas cell cultures with TSF alone contained a 1.6-fold increase in CD34−KSL cells compared to day 0 (data not shown). Notably, treatment of bone marrow CD34−KSL cells with 10–1,000 ng ml−1 pleiotrophin alone yielded no viable cells at day 7 of culture (n = 3 experiments), indicating that pleiotrophin cannot act alone as a survival factor for bone marrow HSCs and that other cytokines are required to maintain HSCs in culture.
To determine whether treatment with pleiotrophin could induce functional HSC expansion in culture, we performed competitive repopulating unit (CRU) assays with limiting dilutions of donor CD45.1+ bone marrow CD34−KSL cells transplanted into lethally irradiated CD45.2+ C57BL/6 mice. We collected peripheral blood from recipient mice at 4, 12 and 24 weeks after transplant to assess the engraftment of donor CD45.1+ cells. At 12 weeks, mice that had been transplanted with the progeny of CD34−KSL cells cultured with TSF plus pleiotrophin showed more than tenfold higher total CD45.1+ cell engraftment (means ± s.d., P = 0.006) and significantly higher B lymphoid (P = 0.003), myeloid (P = 0.03) and T cell engraftment (P = 0.006) at 12 weeks compared to mice transplanted with the same dose of day 0 bone marrow CD34−KSL cells or their progeny after culture with TSF alone (P = 0.007, 0.004, 0.04 and 0.007, respectively; one tailed t test, Fig. 2a,b). These data indicate that pleiotrophin promotes the expansion of HSCs in culture. Poisson statistical analysis (n = 75 mice) showed that the 12-week CRU frequency within day 0 bone marrow CD34−KSL cells was 1 in 39 cells (95% confidence interval: 1 in 21 to 1 in 70, Fig. 2c and Supplementary Table 2). The CRU frequency within the progeny of CD34−KSL cells after culture with TSF was reduced to 1 in 58 cells (95% confidence interval: 1 in 31 to 1 in 108). Conversely, the CRU frequency within the progeny of CD34−KSL cells cultured with TSF and pleiotrophin was 1 in 10 cells (95% confidence interval: 1 in 5 to 1 in 20). Therefore, the addition of pleiotrophin induced a fourfold increase in HSC frequency compared to input and a sixfold increase compared to the progeny of cells treated with TSF alone. Mice transplanted with CD34-KSL cells treated with TSF and pleiotrophin also demonstrated increased levels of CD45.1+ cells in the peripheral blood at all time points through 24 weeks compared to mice transplanted with day 0 CD34−KSL cells or their progeny after culture with TSF alone (Fig. 2d). This correlated with an increased CRU frequency in the pleiotrophin-treated CD34-KSL cells compared to day 0 CD34−KSL cells at all time points. At 4 weeks, the short-term CRU frequency was 6.4-fold higher in the progeny of CD34−KSL cells cultured with TSF and pleiotrophin compared to input CD34−KSL cells (1 in 5 cells (95% confidence interval: 1 in 2 to 1 in 10) versus 1 in 32 cells (95% confidence interval: 1 in 18 to 1 in 57)). At 24 weeks, the CRU frequency was fourfold increased in the pleiotrophin-treated progeny compared to day 0 CD34−KSL cells (1 in 13 (95% confidence interval: 1 in 6 to 1 in 30) versus 1 in 52 (95% confidence interval:1 in 25 to 1 in 106)).
To confirm that pleiotrophin caused an amplification of LT-HSCs with serial repopulating capacity, we performed secondary transplants. CD45.2+ mice transplanted with bone marrow collected at 24 weeks from the primary recipients of pleiotrophin plus TSF-treated CD34−KSL cells showed more than a tenfold higher CD45.1+ cell engraftment at 12 weeks after transplantation compared to mice transplanted with bone marrow from primary mice in the day 0 CD34−KSL cell group or in the TSF alone group (P = 0.003 and P = 0.02, respectively; Fig. 2e). Moreover, mice transplanted with bone marrow from primary mice that had been transplanted with pleiotrophin plus TSF-treated CD34−KSL cells showed normal multilineage differentiation at 12 weeks (Fig. 2e,f). These data demonstrate that treatment with pleiotrophin induces a substantial expansion of long-term repopulating HSCs in culture, and that this amplification does not alter their multilineage differentiation potential.
As altered homing capacity of HSCs can affect HSC frequency estimates in competitive repopulation assays, we compared the homing efficiency of mouse bone marrow progenitor cells with the progeny of bone marrow progenitor cells cultured with TSF or TSF plus pleiotrophin. We irradiated C57BL/6 mice with 950 cGy total body irradiation (TBI) and transplanted them with CD45.1+ Sca-1+Lin− bone marrow cells or their progeny after culture with TSF or TSF plus pleiotrophin, as previously described22. At 24 h after transplantation, we observed no significant differences in the degree of donor CD45.1+ cell engraftment between mice transplanted with day 0 bone marrow cells or with the progeny of TSF-alone or TSF plus pleiotrophin cultures (Fig. 2g). These results indicate that augmented homing capacity does not contribute to the increased HSC frequency estimate in pleiotrophin plus TSF-treated cultures compared to day 0 bone marrow cells or TSF alone-treated progeny.
Because pleiotrophin treatment increased the numbers of both short-term and long-term HSCs in culture, we sought to determine whether transplantation of pleiotrophin-treated bone marrow HSCs could accelerate neutrophil or platelet recovery in lethally irradiated mice. We irradiated C57BL/6 mice with 950 cGy and then transplanted them with a limiting dose (100 cells) of bone marrow CD34−KSL cells or their progeny after culture with TSF or TSF plus pleiotrophin. All mice (13 of 13) transplanted with day 0 bone marrow CD34−KSL cells had severe, persistent neutropenia and thrombocytopenia and died by day 18 (Supplementary Fig. 1). Conversely, mice transplanted with either the progeny of bone marrow CD34−KSL cells treated with TSF alone or TSF plus pleiotrophin survived through day 25 (Supplementary Fig. 1). Notably, mice transplanted with the progeny of TSF plus pleiotrophin cultures showed higher neutrophil counts at days 21 and 25 and higher platelet counts at day 21 compared to mice transplanted with the progeny of TSF-alone cultures (Supplementary Fig. 1). These data suggest that treatment with pleiotrophin produces a graft capable of accelerating neutrophil and platelet recovery compared to input CD34−KSL cells or treatment with TSF alone.
Because we identified pleiotrophin from a gene expression analysis of HUBECs, we also tested whether addition of a neutralizing pleiotrophin-specific antibody to HUBEC cultures could block HUBEC-mediated expansion of LT-HSCs11,14,15. Competitive repopulating assays confirmed that the progeny of non-contact co-cultures of CD34-KSL cells with HUBECs contained more than a threefold higher number of LT-HSCs compared to day 0 bone marrow CD34−KSL cells, and the addition of the pleiotrophin-specific antibody completely abolished this expansion (Supplementary Fig. 2).
In order to determine if pleiotrophin is a growth factor for human HSCs, we sorted human cord blood CD34+CD38−Lin− cells, which are enriched for HSCs14, and placed them in culture with TSF alone or TSF plus 100 or 500 ng ml−1 pleiotrophin for 7 d. The increase in total cell numbers was not significantly different between the TSF alone and TSF plus pleiotrophin groups, but the frequency of CD34+CD38−Lin− cells was significantly higher in cultures treated with TSF plus pleiotrophin versus TSF alone (Fig. 3a). Similarly, treatment of cord blood CD34+CD38−Lin− cells with TSF plus pleiotrophin for 7 d resulted in a fourfold increase in total colony-forming cell (CFC) content compared to TSF alone (Fig. 3b). Taken together, these data suggest that treatment with pleiotrophin induces the expansion of cord blood progenitor cells in culture.
To determine whether pleiotrophin treatment causes an increase in human HSCs in culture, we transplanted nonobese diabetic (NOD)-SCID mice with limiting doses of human cord blood CD34+CD38−Lin− cells or their progeny after culture with TSF or TSF plus pleiotrophin. First, we observed that mice transplanted with the progeny of cord blood CD34+CD38−Lin− cells cultured with TSF plus pleiotrophin had threefold higher human CD45+ cell engraftment at 4 weeks in the peripheral blood compared to mice transplanted with day 0 CD34+CD38−Lin− cells or with the progeny of cells cultured with TSF alone (Fig. 3c). These results indicate that treatment of human cord blood HSCs with pleiotrophin produces a graft capable of accelerated engraftment in transplanted mice. At 8 weeks after transplant using a dose of 500 cells, only one of nine mice (11%) transplanted with day 0 CD34+CD38−lin− cells showed human CD45+ cell engraftment (≥0.1% human CD45+ cells) in the bone marrow at 8 weeks, whereas 6 of 13 mice (46%) transplanted with the progeny of CD34+CD38−Lin− cells cultured with TSF showed human CD45+ cell engraftment (Fig. 3d). Conversely, 13 of 13 mice (100%) transplanted with the progeny of CD34+CD38−Lin− cells cultured with TSF plus pleiotrophin showed human hematopoietic cell engraftment at 8 weeks after transplant (Fig. 3d). At 8 weeks after transplant of a dose of 2,500 cells, four of seven mice (57%) transplanted with day 0 CD34+CD38−Lin− cells and 8 of 11 mice (72%) transplanted with the progeny of TSF-alone cultures showed human CD45+ cell engraftment. Conversely, 12 of 12 mice (100%) transplanted with the progeny of TSF plus pleiotrophin cultures showed human CD45+ cell engraftment (Fig. 3d). Mice transplanted with the progeny of TSF plus pleiotrophin cultures also showed normal multilineage differentiation in vivo(Fig. 3e), confirming that treatment with pleiotrophin does not alter the normal differentiation program of human HSCs.
To determine a potential mechanism through which pleiotrophin mediates HSC expansion, we examined whether pleiotrophin treatment alters signaling pathways known to be affected by RPTP-β/ζ17–19,23. Canonical pleiotrophin signaling occurs via binding and inactivation of RPTP-β/ζ17, which facilitates the tyrosine phosphorylation of several intracellular substrates, including Akt and β-catenin24,25. As pleiotrophin has been shown to mediate mitogenic effects outside the hematopoietic system via activation of the PI3K-Akt pathway24, we tested whether this pathway is involved in pleiotrophin-induced HSC amplification. We treated mouse bone marrow CD34−KSL cells with TSF with or without 100 ng ml−1 pleiotrophin in the presence or absence of 10 µM LY294002, a PI3K inhibitor23,24. The addition of LY294002 to TSF plus pleiotrophin caused an 88% decrease in total cell expansion and a 92% decrease in KSL cell expansion compared to cultures treated TSF plus pleiotrophin without LY294002 (P = 0.002 and P = 0.002, respectively; Fig. 4). These data suggest that PI3K signaling contributes to pleiotrophin-induced bone marrow stem/progenitor cell expansion. Since treatment with pleiotrophin induced a 2.4-fold increase (P=0.04, data not shown) in the expression of hairy and enhancer of split-1 (HES-1), a mediator of Notch signaling 27, we also examined the effect of pleiotrophin treatment of bone marrow CD34-KSL cells with and without a gamma secretase inhibitor, which inhibits Notch signaling 26 Treatment of bone marrow CD34−KSL bone marrow CD34−KSL cells with a γ-secretase inhibitor reduced the expansion of KSL cells by 72% in response to pleiotrophin, suggesting that pleiotrophin mediates effects on bone marrow stem and progenitor cell expansion via Notch signaling (Fig. 4). Since HES-1 has been shown to induce PI3K-Akt signaling in leukemogenesis28, it is plausible that treatment of CD34−KSL cells with pleiotrophin induced HES-1 expression which induced PI3K-Akt signaling in HSCs. Moreover, pleiotrophin treatment of bone marrow CD34−KSL cells downmodulated PTEN expression, a negative regulator of PI3K-Akt signaling29 (data not shown). Of note, bone marrow CD34−KSL cells treated with pleiotrophin showed no increase in the activated (non-phosphorylated) form of β-catenin (data not shown), which is a downstream target of RPTP-β/ζ and a positive regulator of HSC self-renewal7. Furthermore, we observed no difference in the ability of pleiotrophin to amplify KSL cells in cultures of bone marrow KSL cells isolated from mice bearing a deletion of β-catenin compared to cells isolated from wild-type littermate control mice (Supplementary Fig. 3). Taken together, these data suggest that activation of the PI3K and Notch signaling pathways contributes to pleiotrophin-induced HSC expansion.
We next tested whether pleiotrophin administration could augment bone marrow HSC regeneration in vivo after myelosuppression. For these experiments, we irradiated mice with 700 cGy TBI, which causes a 96% decrease in bone marrow HSC content20, and then treated the mice with 2 µg pleiotrophin, 2 µg granulocyte colony–stimulating factor (G-CSF) or saline intraperitoneally daily for 7 d. Pleiotrophin administration caused a significant increase in the number of total bone marrow cells at day 7 compared to G-CSF- or saline-treated mice (Fig. 5a; P = 0.02 and P = 0.04, respectively). Pleiotrophin treatment also caused a threefold increase in the number of bone marrow KSL cells at day 7 and a fourfold and 4.3-fold increase in the number of bone marrow KSL cells at day 14 compared to mice treated with saline or G-CSF, respectively (Fig. 5b). Both G-CSF- and pleiotrophin-treated mice showed significantly higher bone marrow CFC content after irradiation than mice treated with saline, and G-CSF-treated mice had the highest bone marrow CFC content overall over time (Fig. 5c). Notably, mice treated with pleiotrophin had an increase in the number of bone marrow long-term culture–initiating cells (LTC-ICs) by fourfold and >20-fold at days 7 and 14 after irradiation, respectively, compared to saline-treated and G-CSF-treated mice (Fig. 5d). Taken together, these results indicate that systemic treatment with pleiotrophin causes the selective regeneration of phenotypic and functional bone marrow stem/progenitor cells in vivo after myelosuppressive injury. As confirmation of the effect of pleiotrophin administration on bone marrow HSC regeneration in vivo, we also measured bone marrow HSC content via competitive repopulation assays using 5 × 105 bone marrow cells from irradiated, pleiotrophin-treated mice versus irradiated, saline-treated mice. Mice transplanted with bone marrow cells from pleiotrophin-treated mice showed significantly higher multilineage engraftment of donor CD45.1+ cells at 4, 8 and 12 weeks after transplant compared to mice transplanted with saline-treated bone marrow cells (Fig. 5e andSupplementary Fig. 4). These results confirm that systemic treatment with pleiotrophin induces the regeneration of short- and long-term HSCs in vivo after irradiation.
Of note, systemic administration of pleiotrophin to mice irradiated with lower-dose TBI (300 cGy) showed no consistent differences in the numbers of total bone marrow cells, bone marrow KSL cells or CFCs as compared to G-CSF-treated mice (Supplementary Fig. 5). However, nonirradiated mice treated with pleiotrophin for 7 d showed a significant expansion in the number of bone marrow KSL cells (P = 0.009) compared to saline-treated controls (Supplementary Fig. 5). These data suggest that systemic pleiotrophin administration can induce bone marrow stem/progenitor cell expansion in normal mice, an idea that will be tested in future studies.
Our results demonstrate that pleiotrophin is a secreted growth factor for HSCs, and the addition of pleiotrophin is sufficient to promote a substantial expansion of mouse LT-HSCs in culture, as shown using primary and secondary competitive repopulation assays. Furthermore, we show that treatment of human cord blood HSCs with pleiotrophin plus cytokines increases the short-term and long-term SCID-repopulating capacity of these cells compared to treatment with the cytokines alone. Pleiotrophin, therefore, has potential therapeutic application for an area of unmet clinical need: the production of cord blood grafts capable of accelerated engraftment in adult transplant candidates who lack a histocompatible adult donor. In addition, we show that systemic administration of pleiotrophin causes a substantial increase in the regeneration of both short- and long-term repopulating bone marrow HSCs in vivo after total body irradiation. Therefore, pleiotrophin regulates not only mouse and human HSC expansion in vitro but also HSC regeneration in vivo, a process that is largely uncharacterized.
Because bone marrow HSCs express RPTP-β/ζ and our in vitro studies showed a direct effect of pleiotrophin on HSCs, we propose that pleiotrophin acts directly on bone marrow HSCs to induce bone marrow HSC regeneration in vivo. However, it will be crucial to examine the effects of pleiotrophin administration on the bone marrow microenvironment. Pleiotrophin has been shown to have angiogenic activity30,31, and we and others have shown that bone marrow vascular endothelial cells can regulate hematopoietic reconstitution after injury20,33. Therefore, it is plausible that pleiotrophin might contribute indirectly to bone marrow HSC regeneration in vivo by promoting recovery of the bone marrow vascular niche. Given that little is known about the extrinsic or microenvironmental signals that regulate bone marrow HSC regeneration in vivo34, the demonstration that pleiotrophin induces bone marrow HSC regeneration in vivo provides a basis to begin to understand this process. Furthermore, as a soluble growth factor capable of inducing bone marrow HSC regeneration in vivo, pleiotrophin has unique translational potential compared to previously described pathways shown to regulate HSC expansion in vitro6,10,12.
As a possible mechanism of action, we showed that inhibition of PI3K signaling blocks pleiotrophin-induced expansion of bone marrow KSL cells in culture. Pleiotrophin also induced the expression of HES-1, a mediator of Notch signaling and a positive regulator of PI3K signaling27,28, suggesting the possibility that pleiotrophin promotes HSC amplification via activation of Notch signaling. Consistent with this hypothesis, we found that inhibition of Notch signaling with a γ-secretase inhibitor also blocked pleiotrophin-induced bone marrow KSL cell expansion. In future studies, we will test the role of HES-1 and Notch signaling in mediating pleiotrophin effects on HSC expansion using genetic strategies in mice. It has recently been reported that deletion of the gene encoding PTEN, a negative regulator of PI3K-Akt signaling, was associated with exhaustion of 12-week CRU in mice35; in addition, deletion of Foxo3a, a transcription factor that negatively regulates HSC cycling and is inhibited by Akt, has been associated with depletion of LT-HSCs in mice36. Therefore, it will be crucial to determine whether pleiotrophin-mediated expansion of HSCs is dependent upon PI3K-Akt signaling or whether pleiotrophin-mediated HSC expansion is independently caused by activation of alternative HSC regulatory pathways (for example, Notch signaling).
Much progress has been made in understanding the intrinsic and extrinsic pathways that regulate HSC self-renewal and differentiation1–3,13,37. However, the successful development of soluble growth factors or cytokines capable of inducing human HSC expansion ex vivo or HSC regeneration in vivo has remained an elusive goal13,37. Here we show that the soluble growth factor pleiotrophin acts on both mouse and human HSCs and can induce LT-HSC expansion ex vivo and HSC regeneration in vivo. Pleiotrophin therefore has potential clinical utility for expanding human HSCs ex vivo and for accelerating hematopoietic recovery in vivo in patients after myelotoxic chemo- or radiotherapy.
We cultured individual primary human endothelial cells derived from uterine, umbilical, iliac, dermal, coronary and pulmonary arteries (Lonza) according to the manufacturer’s guidelines. We generated primary HUBECs and cultured them in complete endothelial cell culture medium (M199, Invitrogen; Endothelial Growth Supplement, Sigma; Fetal Bovine Serum, Hyclone; Heparin, Sigma; L-glutamine, Invitrogen) as previously described11,14. We amplified RNA from n = 6 sources of HUBECs and n = 8 sources of nonbrain endothelial cells to a human oligonucleotide spotted microarray (Operon). We analyzed the microarray data with an unsupervised hierarchical cluster analysis (reference 38 in original ms; Dressman H et al. PLoS Medicine 4;e106), and we screened the gene list for annotated soluble proteins. We processed and hybridized samples to Operon Human Arrays (Operon) as previously described38.
We performed all studies with mice under a protocol approved by the Duke University Animal Care and Use Committee. We isolated purified bone marrow CD34−KSL cells from C57BL/6 and B6.SJL mice (Jackson Laboratory) by flow cytometric cell sorting as previously described6,20. We supplemented liquid suspension cultures of bone marrow CD34−KSL cells with Iscove's modified Dulbecco's medium plus 10% FBS, 1% penicillin-streptomycin, 20 ng ml−1 thrombopoietin, 125 ng ml−1 stem cell factor and 50 ng ml−1 Flt-3 ligand (TSF) with or without recombinant (human) pleiotrophin (R&D Systems). We set up noncontact HUBEC cultures using 0.4-µm transwell inserts (Corning) and supplemented them with TSF medium with or without goat antibody to pleiotrophin (AF-252-PB, R&D Systems, Minneapolis, MN) or isotype control antibody (AB-108-C, R&D Systems). We analyzed for surface marker expression of c-kit, sca-1 and lineage markers as previously described15,20.
We isolated bone marrow CD34−KSL cells from CD45.1+ Bl6.SJL mice for injection into recipient mice and injected them without culturing or cultured them with TSF alone, TSF plus pleiotrophin, TSF plus HUBECs plus goat IgG, or TSF plus HUBECs plus goat pleiotrophin-specific antibody. C57BL/6 mice (CD45.2+) recipient mice were irradiated recipient with 950 cGy TBI and were injected via the tail vein with limiting doses of bone marrow CD34−KSL cells or their progeny after culture. 1 × 105 host bone marrow mononuclear cells from C57Bl6 mice were used as competitor cells. We measured multilineage hematologic reconstitution in the peripheral blood by flow cytometry over time after transplantation as previously described6,20. We considered the mice to be engrafted if the donor CD45.1+ cells were present at ≥1% in the peripheral blood11,14,39. We made CRU estimates with L-Calc software (Stem Cell Technologies) as previously described6,15,39.
We performed secondary competitive transplant assays with whole bone marrow collected from primary CD45.2+ mice at 24 weeks after transplantation with either CD45.1+ bone marrow CD34−KSL cells or the progeny of CD34−KSL cells after culture with TSF alone or TSF plus pleiotrophin. By volume, 75% of the primary BM cells wre transplanted into recipient mice along with 1×105 host competitor BM cells from the C57Bl6 mice. We irradiated secondary recipient CD45.2+ C57BL/6 mice with 950 cGy TBI and analyzed peripheral blood of donor cell engraftments at 12 weeks after transplantation in secondary mice.
We isolated human CD34+CD38−Lin− cells from cord blood units obtained from the Carolinas Cord Blood Bank. We cultured 5 × 103 cord blood CD34+CD38−Lin− cells with TSF in Iscove's modified Dulbecco's medium containing 10% FBS and 1% penicillin-streptomycin. We added recombinant pleiotrophin at 100–500 ng ml−1. We performed phenotypic analyses and 14-d CFC progenitor cell counts (number of colonies per 1 × 103 cells plated × fold expansion) as previously described11. We transplanted limiting doses (0.5 × 103–2.5 × 103) of day 0 cord blood CD34+CD38−Lin− cells or their progeny after 7 d of culture with TSF alone or TSF plus 500 ng ml−1 pleiotrophin into NOD-SCID mice conditioned with 300 cGy TBI. We considered the mice as positively engrafted if we detected ≥ 0.1% human CD45+ cells in the bone marrow at 8 weeks after transplant, as previously described40. We measured multilineage human cell repopulation at 8 weeks, as previously described11,14.
We gave adult B6.SJL mice a single fraction of 700 cGy TBI and then treated them with either with PBS (saline), 2 µg GCSF or 2 µg pleiotrophin intraperitoneally daily for 7 d (beginning 4 h after irradiation). At days 4, 7 and 14 after irradiation, we killed the mice and quantified total viable bone marrow cells. We performed flow cytometric analysis to estimate the percentage of bone marrow KSL cells in each femur15,20. We performed CFC assays with MethoCult M3434 medium (Stem Cell Technologies) as previously described15,20. We performed LTC-IC assays as follows: we plated mouse M2-10B4 (American Type Culture Collection CRL-1972) bone marrow stromal cells in a 24-well dish and irradiated them with 1,500 cGy. We prepared limiting dilutions (45,000, 90,000 and 180,000) of bone marrow mononuclear cells from irradiated mice that had been treated with either pleiotrophin or PBS, added these cells to the stromal cell layers and maintained the layers in MyeloCult M5300 medium (Stem Cell Technologies) with weekly half-medium changes for 4 weeks. We then collected the nonadherent and adherent cells (15,000 cells per dish) and plated them into 3 × 35 mm dishes (MethoCult, StemCell Technologies). After 2 weeks, we counted and scored hematopoietic colonies. We also performed CRU assays as previously described20 with a limiting dose of bone marrow cells from mice that had been irradiated with 700 cGy TBI and then treated with saline or pleiotrophin for 7 d to compare bone marrow HSC content within these mice.
Data are expressed as the means ± s.d. or s.e.m. We analyzed simple pair-wise comparisons with the Student’s t test (one-tailed distribution with unequal variance). For the comparison of donor engraftment in the secondary transplant assays, we used a Mann-Whitney test. For competitive repopulating assays, we performed limiting-dilution assays and calculated CRU frequency with the maximum likelihood estimator for the single-hit Poisson model11.
Detailed methodology is described in the Supplementary Methods.
We acknowledge J. Whitesides for assistance with cell sorting procedures. This work was supported in part by US National Institutes of Health grant AI067798 to J.P.C. H.A.H. is supported by a post-doctoral training grant from the Center for Biomolecular and Tissue Engineering, US National Institute of Biomedical Imaging and Bioengineering.
Accession codes. Microarray data have been deposited in the Gene Expression Omnibus (GEO) Database with accession code GSE20243.
Note: Supplementary information is available on the Nature Medicine website.
AUTHOR CONTRIBUTIONSH.A.H. designed and performed experiments, analyzed data and wrote the paper; G.G.M., P.D., S.K.M., J.L.R, P.D., A.B.S. and W.E.L. performed experiments; J.-T.C. guided the microarray analysis; T.R. and N.C. analyzed data and wrote the paper; J.P.C. designed the experiments, analyzed the data and wrote the paper.
COMPETING FINANCIAL INTERESTS
The authors declare no competing financial interests.