The prognosis of patients undergoing hematopoietic stem cell transplantation (HSCT) depends on the rapid recovery and sustained life-long hematopoiesis. The activation of the fibrinolytic pathway promotes hematopoietic regeneration; however, the role of plasminogen activator inhibitor-1 (PAI-1), a negative regulator of the fibrinolytic pathway, has not yet been elucidated. We herein demonstrate that bone marrow (BM) stromal cells, especially osteoblasts, produce PAI-1 in response to myeloablation, which negatively regulates the hematopoietic regeneration in the BM microenvironment. Total body irradiation in mice dramatically increased the local expression levels of fibrinolytic factors, including tissue-type plasminogen activator (tPA), plasmin, and PAI-1. Genetic disruption of the PAI-1 gene, or pharmacological inhibition of PAI-1 activity, significantly improved the myeloablation-related mortality and promoted rapid hematopoietic recovery after HSCT through the induction of hematopoiesis-promoting factors. The ability of a PAI-1 inhibitor to enhance hematopoietic regeneration was abolished when tPA-deficient mice were used as recipients, thus indicating that PAI-1 represses tPA-dependent hematopoietic regeneration. The PAI-1 inhibitor not only accelerated the expansion of the donor HSCs during the early-stage of regeneration, but also supported long-term hematopoiesis. Our results indicate that the inhibition of PAI-1 activity could be a therapeutic approach to facilitate the rapid recovery and sustained hematopoiesis after HSCT. Stem Cells2014;32:946–958
Hematopoietic stem cell transplantation (HSCT) is used as a therapy for patients who suffer from hematological malignancies. In general, such patients are myeloablated by chemotherapy and/or radiotherapy to eradicate the deranged host hematopoietic system, followed by transplantation of healthy donor-derived hematopoietic cells . However, due to a low engraftment efficiency and delayed bone marrow (BM) reconstitution, these patients occasionally suffer from severe immunodeficiency, thus leading to an increased susceptibility to serious infectious diseases, and therefore, to a high-risk of transplant-related death . The establishment of an efficient strategy to improve the recovery and sustain hematopoiesis is a goal in the treatment of patients undergoing HSCT.
The fibrinolytic pathway breaks down fibrin clots in the blood, and plasmin plays a central role in this process . The proenzyme plasminogen (Plg) is produced from the liver and circulates in the blood stream. Under certain circumstances, such as wound healing, Plg is proteolytically converted into the active enzyme plasmin by tissue-type plasminogen activator (tPA), which is released from endothelial cells . The blood also contains negative regulators of the fibrinolysis pathway, including plasminogen activator inhibitor-1 (PAI-1). The production of plasmin from plasminogen is thus regulated by a balance between activator molecules (e.g., tPA) and their inhibitors (e.g., PAI-1) [4, 5]. Therefore, inhibiting the PAI-1 activity is expected to accelerate the activation of the tPA-mediated fibrinolytic pathway.
Recently, Hattori and colleagues demonstrated that the fibrinolytic pathway regulates hematopoietic regeneration [6, 7]. They showed that the deletion of the Plg gene impaired the entry of quiescent HSCs into the cell cycle and delayed hematopoietic regeneration. In contrast, the activation of Plg by the exogenous administration of recombinant tPA promoted HSC proliferation and differentiation through the potentiation of matrix metalloproteinase (MMP)-mediated release of c-kit ligand (c-kitL) from BM stromal cells. The fibrinolytic pathway thus plays a role in hematopoiesis.
We have recently developed a low molecular weight synthetic inhibitor of PAI-1, TM5275 (5-chloro-2[((2-[4-(diphenylmethyl) piperazine-1yl]-2-oxoethoxy)acetyl) amino]benzoate) [8, 9]. TM5275 binds selectively to the A β-sheet (s4A) position of the PAI-1 molecule, preventing the formation of the PAI-1/tPA complex, thereby preserving active tPA. Previous studies have demonstrated that TM5275, which was given orally, provided antithrombotic benefits without prolonging the bleeding time in rodent and monkey thrombosis models . Given the potential importance of the fibrinolytic pathway in HSCT, in this study, we addressed whether the suppression of PAI-1 activity could affect the hematopoietic regeneration after transplantation in PAI-1 knockout (KO) mice and using the PAI-1 inhibitor, TM5275. Our study demonstrates that PAI-1 is a negative regulator of hematopoietic regeneration, and that suppression of the PAI-1 activity leads to both a rapid recovery and long-term maintenance of donor-derived hematopoiesis. Therefore, the inhibition of PAI-1 activity could be a therapeutic approach to facilitate rapid recovery and sustained hematopoiesis.
Materials and Methods
Eight- to twelve-week-old C57Bl/6J mice were purchased from CLEA Japan (Tokyo, Japan, www.clea-japan.com). PAI-1-deficient mice (B6.12952-Serpine1tm1Mg/J) [11, 12] were purchased from Jackson Laboratory (Bar Harbor, ME, www.jax.org). tPA-deficient mice (B6.129S2-Plattm1Mlg/J)  were kindly provided by Dr. Koichi Hattori, University of Tokyo, Japan. All mice were housed in cages at the animal facility of Tokai University School of Medicine. All the protocols for animal experiments were approved by the Animal Care Committee of Tokai University, and animals were treated in accordance with the institutional guidelines.
To distinguish donor (Ly5.1)- and recipient (Ly5.2)-derived BM cells, we used the Ly5.1/Ly5.2 congenic system and analyzed the reconstitution of hematopoietic cells by a fluorescence-activated cell sorting (FACS) analysis by gating Ly5.1+ donor-derived cells. Before transplantation, the recipient mice were lethally irradiated (9 Gy) in an x-ray irradiator (MBR-1520R-3, Hitachi Medico, Tokyo, Japan, www.hitachi-power-solution.com). Ly5.1+ BM mononucleic cells (MNCs, 2.5 × 106) were transplanted intravenously into the retro-orbital plexus of Ly5.2+ congenic mice.
For the secondary transplantation, 1 × 106 donor-derived Ly5.1+ BM MNCs from the primary recipients were retransplanted into Ly5.2+ secondary recipients which had been irradiated with 9 Gy. To protect secondary recipients from radiation-related lethality, 5 × 105 Ly5.2+ competitor cells were transplanted along with the Ly5.1+ donor cells. To compare the proportion of long-term HSCs, Ly5.1+ BM cells from the primary recipients were serially diluted and administered along with 5 × 105 Ly5.2+ competitor cells into secondary recipients that had been irradiated with 9 Gy. At 12 weeks after secondary transplantation, the BM MNCs were collected and stained with APC-conjugated anti-Ly5.1, FITC-conjugated anti-B220, PE-conjugated anti-Gr-1 and anti-Mac-1 antibodies, and were analyzed by FACS LSRFortessa (BD Bioscience, San Jose, CA, www.bdbioscience.com). The proportion of donor cells was calculated from a total of 200,000 events. Successfully engrafted mice were defined as recipients that contained more than 1.0% Ly5.1+ donor-derived cells with both lymphoid (B220+) and myeloid (Gr-1+/Mac-1+) differentiation markers. In the radioprotection assay, we used 12 Gy irradiation, which results in 20% survival when 1 × 106 cells were transplanted (Supporting Information Fig. S1).
Administration of the PAI-1 Inhibitor or tPA
TM5275 is a specific inhibitor of PAI-1 molecules that had a half-maximal inhibition (IC50) value of 6.95 µM in a tPA-dependent hydrolysis assay [8, 10]. It does not interfere with other serpin/serine protease systems, such as the alpha1-antitrypsin/trypsin and alpha2-antiplasmin/plasmin systems.
After BM transplantation, TM5275 (100 mg/kg) or vehicle (saline) was administered daily to mice via oral gavage using a feeding needle for 5 consecutive days. Recombinant tPA (10 mg/kg, Eisai, Tokyo, Japan, www.eisai.com) was administered to the mice daily by intraperitoneal injection for 5 consecutive days. The dosage is equivalent to that used in the clinical setting.
Isofluorane-anesthetized mice were perfused with 4% paraformaldehyde in phosphate buffered saline (PBS) through the left ventricle. The femur and tibia were removed, decalcified, embedded in OCT compound, and frozen in liquid nitrogen. Alternatively, the decalcified bones were embedded in paraffin. The deparaffinized sections were stained for tPA by incubation with a rabbit anti-mouse tPA polyclonal antibody (Santa Cruz Biotechnology, California, CA, www.scbt.com), for Plg/plasmin with a rabbit anti-human Plg/plasmin polyclonal antibody (Santa Cruz Biotechnology) or for PAI-1 with a rabbit anti-mouse PAI-1 polyclonal antibody (Abcam, Cambridge, MA, www.abcam.com), followed by visualization with a catalyzed signal amplification II system (Dako, California, CA, www.dako.com). The slides were then developed with diaminobenzidine and counterstained with methyl green. For the double immunohistochemical staining analysis, the bone sections were stained with PAI-1 or tPA antibodies, followed by costaining with either rat anti-mouse PECAM-1 (CD31) monoclonal antibodies (BD BioSciences), goat anti-mouse osteocalcin polyclonal antibodies (Santa Cruz Biotechnology), or goat anti-mouse alpha one chain of type I collagen (Col(I) α1) polyclonal antibodies (Santa Cruz Biotechnology). Serial sections of bone were stained with rabbit anti-mouse c-kit polyclonal antibodies (Santa Cruz Biotechnology) or rabbit anti-mouse proliferation cell nuclear antigen (PCNA) polyclonal antibodies (Abcam). Fluorescent immunohistochemistry was also performed with secondary antibodies as follows: Alexa Fluor 488 goat anti-rabbit IgG, Alexa Fluor 594 donkey anti-goat IgG or Alexa Fluor 594 goat anti-rabbit IgG secondary antibody (Life Technologies Corporation, Grand Island, NY, www.lifetechnologies.com), followed by counterstaining with 4′,6-diamidino-2-phenylindole. The endosteal region of the BM was defined as that within 12 cells from the endosteum [14, 15]. Images were captured using a HS All-in-one Fluorescence Microscope Biorevo 9000 (Keyence Corporation, Osaka, Japan, www.keyence.com) and analyzed by the BZ II analyzer software program (Keyence Corporation).
Evaluation of the tPA, Plasmin, PAI-1, MMP-9, and C-KitL Levels in Blood Plasma and BM Fluid
Plasma was prepared from peripheral blood (PB) with EDTA as an anticoagulant, and then the specimens were centrifuged at 11,600g for 10 minutes to completely remove platelets. BM liquid was collected as described previously . Briefly, four long leg bones were perused with 1 ml of PBS containing 2 mM ethylenediaminetetraacetic acid (EDTA) and 0.5% bovine serum albumin (BSA). BM cells were removed by centrifugation at 350g, and the resulting supernatant was designated as BM liquid. The volume of the BM cavities of the bones was assumed to be 25 µl in this study, and the cytokines concentration were multiplied by this dilution factor (1,000/25 = 40×) and expressed per unit BM volume as described previously . The concentrations of plasmin (Innovative Research, Novi, MI, www.innov-research.com), tPA (Innovative Research), active PAI-1 (Innovative Research), total MMP-9 (R&D System, MN, www.rndsystems.com), and stem cell factor (SCF) (c-kitL, R&D System) in the plasma and in BM fluid were determined by enzyme-linked immunoabsorbent assay (ELISA) kits according to the manufacturers' instructions.
PB Cell Counts
PB was collected and a complete blood count was determined using a Sysmex Hematology Analyzer (Sysmex Co., Kobe, Japan, www.sysmex.com).
Analysis of HSC and Cell Engraftment
On day 2 and 1, 3, and 15 weeks after the infusion of MNCs, the mice were euthanized, and the BM MNCs were collected from the femurs and tibiae. The BM MNCs were counted, and aliquots of cells were stained with various antibodies as noted below. Donor-derived hematopoietic cells were labeled with a PE-conjugated anti-Ly5.1 antibody (CD45.1, BD Biosciences) and biotin-conjugated antibody cocktail for lineage markers (CD5, CD11b, CD45R, Gr-1, 7-4, and Ter119; Miltenyi Biotec, Bergisch Gladbach, Germany, www.miltenyibiotec.com), followed by perinidin chlorophyll protein-cyanine 5.5 (PerCP-Cy5.5)-conjugated streptavidin (BD Biosciences). The labeled cells were divided into two aliquots, each of which was then mixed with either antibody cocktail A (APC-conjugated anti-mouse c-kit [CD117] antibody [eBioscience, San Diego, CA, www.ebioscience.com], PE-Cy7-conjugated anti-mouse Sca-1 [Ly6A/E] antibody [eBioscience], and FITC-conjugated anti-mouse CD34 antibody [eBioscience]) or with antibody cocktail B (APC-conjugated anti-mouse CD48 antibody [eBioscience] and PE-Cy7-conjugated anti-mouse CD150 antibody [eBioscience]). A flow cytometric analysis was performed on the FACS LSRFortessa (BD Bioscience) instruments using the FACSDiva software program (BD Bioscience). Dead cells were gated out by staining with propidium iodide. The proportion of each lineage was calculated from 1,000,000 events.
Cell Proliferation and Cell Cycle Analysis
At 1 week post-transplantation, Ly5.1+, lineage-negative, Sca-1-positive, c-kit-positive (LSK) cells were isolated and stained with anti-Ki67-FITC antibody according to the manufacturer's instructions (BD Bioscience).
The data were analyzed using unpaired two-tailed Student's t tests or the Log-rank test for the survival analysis using the PRISM software program (GraphPad software, LA Jolla, CA, www.graphpad.com). For comparisons of more than three groups, a one-way ANOVA followed by Bonferroni post-tests was performed. A value of p < .05 was considered to be significant.
Irradiation Activates the Fibrinolytic Pathway and Increases the Expression of PAI-1 in the BM Microenvironment
To examine the effects of irradiation on the fibrinolytic system in the BM microenvironment, the levels of fibrinolytic factors, such as tPA, plasmin/Plg, and its inhibitor, PAI-1, in the BM fluid and in plasma were measured by ELISA. Two days after 9 Gy irradiation, the levels of tPA and plasmin in the BM fluid, as well as in the plasma, were significantly elevated (Fig. 1A). The level of active PAI-1, a negative regulator of the fibrinolytic system, was also elevated in the plasma and BM fluid of the irradiated mice (Fig. 1A). Of note, the increases of these fibrinolytic factors and PAI-1 were much more prominent in the BM fluid than in the blood, suggesting that the irradiation dramatically activates the fibrinolytic pathway and its inhibitor, PAI-1, in the hematopoietic microenvironment.
Increases in the expression levels of tPA, plasmin (or Plg), and PAI-1 in the BM microenvironment were also confirmed by means of immunohistochemical approaches (Fig. 1B, 1C). The irradiation severely destroyed the BM structure, which may have increased the chance of non-specific staining. Therefore, the specificity of the antibodies used in these immunostaining was verified in tPA KO mice or PAI-1 KO mice. No substantial material in the tissue was stained with the respective antibodies under the same assay conditions in these mice (Supporting Information Fig. S2).
Next, the cells responsible for producing the fibrinolytic factors and PAI-1 in the irradiated BM were identified. tPA was detected in vasculature-lining CD31+ endothelial cells and was significantly elevated after irradiation (Fig. 1B, 1C and Supporting Information Fig. S3A). In agreement with previous reports [4, 5], PAI-1 was detected in megakaryocyte-like cells in untreated mice, but after irradiation, the PAI-1 producing megakaryocyte-like cells disappeared, and the PAI-1 levels in the endosteal region were increased. The tissue distributions of plasmin and Plg colocalized with that of PAI-1 after irradiation (Fig. 1B). Our subsequent double-staining studies revealed that the majority of PAI-1-expressing cells in the irradiated BM were both osteocalcin-positive and Col(I) α1-positive osteoblasts (Fig. 1C and Supporting Information Fig. S3B, S3C).
The activation of fibrinolytic factors and the fibrinolysis inhibitor, PAI-1, in nonhematopoietic cells in the hematopoietic niche shortly after irradiation was also conformed in in vitro studies. We observed that primary BM stromal cells, as well as the BM stromal cell line, HESS5, produced fibrinolytic factors and PAI-1 after irradiation (Fig. 1D and Supporting Information Fig. S4).
PAI-1 Is a Negative Regulator of Early-Phase Hematopoietic Regeneration
To investigate role of PAI-1 in the hematopoietic regeneration, we transplanted 2.5 × 106 Ly5.1+ BM MNCs into Ly5.2+ congenic mice, which had been myeloablated by 9 Gy irradiation. PAI-1 KO mice (Ly5.2+) were used as HSCT recipients to monitor the effects of the PAI-1 inhibition on hematopoietic regeneration. The PAI-1 KO recipient mice did not exhibit the induction of active PAI-1, regardless of the transfer of wild-type (WT; PAI-1+/+) hematopoietic cells, suggesting that donor-derived hematopoietic cells may not be involved in the regulation of the fibrinolytic pathway during hematopoietic regeneration (Fig. 2A).
A previous study by another group  reported that the activation of the fibrinolytic pathway by recombinant tPA promoted hematopoietic cell proliferation through the MMP-mediated release of c-kitL. We therefore evaluated the expression levels of active tPA and other hematopoietic regulatory factors (i.e., MMP-9 and c-kitL) in the PAI-1 KO mice. The results demonstrated a marked increase in the expression of these factors compared to the WT mice at each time point (Fig. 2B–2E).
We subsequently examined the efficiency of hematopoietic regeneration by a flow cytometric analysis (Supporting Information Fig. S5). Over 90% (93.96% ± 1.56%, n = 36) of the hematopoietic cells in the recipient BM were Ly5.1+ donor-derived cells. The absolute number of BM MNCs (Fig. 2F) and the proportion of Ly5.1+ donor-derived Lin− SLAM (CD150+CD48−) HSCs (Fig. 2G, 2H) were higher in the PAI-1 KO mice than in the WT mice during both the steady state and post-transplant periods. We also examined another HSC marker, CD34−LSK (Lin−Sca-1+c-kit+), and showed that the proportion of Ly5.1+ CD34−LSK cells was higher in the PAI-1 KO mice than in the WT mice (Fig. 2I, 2J), supporting our hypothesis that the induction of PAI-1 in the hematopoietic microenvironment inhibits hematopoietic regeneration. Collectively, our results demonstrated that radiation-induced myeloablation augments the expression levels of not only hematopoietic regeneration-enhancing factors, tPA and Plg/plasmin, but also simultaneously enhances the expression of their negative regulator, PAI-1.
The Expression of Fibrinolytic Factors Is Augmented by PAI-1 Inhibition During Hematopoietic Recovery
We tested our hypothesis that the pharmacological inhibition of PAI-1 can augment the endogenous tPA-mediated fibrinolytic pathway activity more efficiently than exogenous tPA administration, leading to more efficient hematopoietic reconstitution after BM transplantation. A PAI-1 inhibitor or recombinant tPA was thus administered to the WT irradiated mice, and the early phase of hematopoietic recovery and the changes in fibrinolytic factors and the fibrinolysis inhibitor in the plasma were monitored at several time points. The administration of a PAI-1 inhibitor resulted in almost complete suppression of the elevation of active PAI-1 after BM transplantation (Fig. 3A). The PAI-1 inhibitor significantly increased the plasma levels of active tPA and plasmin (Fig. 3B, 3C). Based on the fact that the half-life of recombinant tPA is only a few minutes in rodents , as well as the results in Figure 1, it is likely that the increased tPA present at 2 and 7 days post-transplantation reflects the local production of tPA in the BM generated by irradiation. Surprisingly, the PAI-1 inhibitor augmented the induction of fibrinolytic factors more strongly than did the direct administration of recombinant tPA. This may be explained, at least in part, by the fact high PAI-1 activity was maintained during recombinant tPA administration (Fig. 3A), which may limit its benefit on fibrinolytic factors.
We confirmed that administering a PAI-1 inhibitor to mice induced an increase in the plasma levels of hematopoiesis-promoting factors, such as MMP-9 and c-kitL (3- and 1.7-fold compared to the vehicle treatment, respectively) as shown in Figure 3D and 3E. These results demonstrate that the inhibition of PAI-1 effectively induces factors promoting hematopoiesis.
The Activation of the Fibrinolytic Pathway by PAI-1 Inhibition Enhances the Hematopoietic Reconstitution
The protection against BM damage after irradiation and/or chemotherapy is a primary factor that determines the survival rate of animals , which hinges on how rapidly the transplanted (i.e., following radioablation) or remaining (i.e., following chemotherapy) hematopoietic cells can reconstitute the crippled hematopoietic system. The survival rates of recipient mice after either a lethal dose of radiation (12 Gy) or 5-fluorouracil administration were investigated in the presence of either recombinant tPA or a PAI-1 inhibitor. The administration of tPA improved the survival rate after myeloablative treatment, but PAI-1 inhibitor treatment offered significantly more effective protection (Fig. 4A, 4B).
The benefits on the hematopoietic recovery were also confirmed in 9 Gy-irradiated mice, which had been transplanted with BM MNCs. The numbers of both white blood cells and platelets markedly increased after BM transplantation: the PAI-1 inhibitor treatment led to more successful hematopoietic recovery than did the treatment with recombinant tPA (Fig. 4C, 4D). These results demonstrate that the inhibition of PAI-1 promotes hematopoietic recovery and protects against myeloablation-induced mortality.
Suppressing the PAI-1 Activity Induces tPA-Mediated HSC Proliferation in the BM After HSCT
To further elucidate the mechanism(s) by which the PAI-1 inhibition improves the hematopoietic recovery, 2.5 × 106 Ly5.1+ BM cells were transplanted into the 9 Gy-irradiated Ly5.2+ congenic mice, and the effects of tPA or a PAI-1 inhibitor on the proliferation of HSCs were assessed. The absolute number of BM MNCs in the recipients indeed increased in the mice given a PAI-1 inhibitor (Fig. 5A). At 3 weeks after transplantation, both the proportion and the absolute number of phenotypically identified Ly5.1+ donor-derived HSC compartments were significantly higher in the PAI-1 inhibitor-treated mice than in the control or the tPA-treated mice (Fig. 5B–5E). The proportion of mature myeloid and lymphoid Ly5.1+ donor-derived cells in the recipient BM was equivalent in the vehicle-treated and PAI-1 inhibitor-treated recipients, suggesting that the PAI-1 inhibitor does not influence the differentiation of HSCs (Supporting Information Fig. S6). The ability of a PAI-1 inhibitor to induce hematopoietic regeneration, that is, upregulation of MMP and c-kitL production and enhancement of donor cell engraftment was completely negated when the tPA KO mouse was used as the recipient (Fig. 5F–5J), further supporting our hypothesis that the hematopoietic regeneration in our model was derived from the PAI-1 inhibition and subsequent tPA-mediated proliferation of HSCs during early hematopoietic reconstitution.
To confirm that the PAI-1 inhibitor induces HSC proliferation, the 9 Gy-irradiated Ly5.2+ congenic mice were transplanted with 2.5 × 106 Ly5.1+ BM cells and given a PAI-1 inhibitor for 5 consecutive days. The expression of Ki67 in the Ly5.1+ donor-derived hematopoietic stem and progenitor cells was examined 1 week after HSCT. The Ki67+ donor-derived LSK cells were detected at the highest level in the mice given the PAI-1 inhibitor (Fig. 6A), which correlated with the HSC proportion in the recipient. The inhibition of PAI-1 thus stimulates hematopoietic stem/progenitor cells (HSPCs) to enter into the cell cycle. The PCNA+c-kit+ proliferating HSPCs were preferentially located in the endosteal region of the BM (arbitrarily defined as within 12 cells of the endosteum) [14, 15] at 1 week after transplantation (69.4% ± 4.1%; n = 1,811 PCNA+c-kit+ cells in the vehicle-treated group and 76.7% ± 3.4%; n = 2,983 PCNA+c-kit+ cells in the PAI-1 inhibitor-treated group, p < .01), indicating that the proliferation of HSPCs in the PAI-1 inhibitor-treated recipients as regulated by their interaction with the niche (Fig. 6C).
In the early-stage of regeneration, the HSPCs underwent apoptosis at a higher rate in the vehicle-treated recipient, whereas the PAI-1 inhibitor prevented the apoptotic cell death of HSPCs (Supporting Information Fig. S7). Altogether, these findings indicate that the PAI-1 inhibitor enhances the proliferation of HSPCs and protects them from stress-induced apoptosis, leading to improve the hematopoietic regeneration.
PAI-1 Inhibition Potentiates the Self-Renewal Capacity of HSCs
The rapid proliferation of HSCs occasionally results in their exhaustion (i.e., loss of competence as HSCs), eventually leading to the failure of long-lasting hematopoiesis in the BM [19-21]. To examine whether the elevation of tPA activity, either by tPA- or PAI-1 inhibitor administration, induced transplanted HSCs to undergo exhaustion, the recipient BM was analyzed at 15 weeks in the Ly5.2+ mice that had been transplanted with Ly5.1+ donor BM cells and given a PAI-1 inhibitor. The results indicated that when the PAI-1 inhibitor was administered to mice, it increased not only the number of BM MNCs (Fig. 7A), but also the proportion, as well as the absolute number of phenotypic Ly5.1+ HSCs compared to the vehicle- or recombinant tPA-treated group (Fig. 7B-7E), suggesting that the inhibition of PAI-1 activity by the PAI-1 inhibitor efficiently prolonged the survival of donor-derived HSCs in the BM.
The self-renewal capacity of HSCs was also examined in a secondary transplant performed 15 weeks after the primary transplant. Twelve weeks after the secondary transplant, the chimerism of the primary donor-derived Ly5.1+ hematopoietic cells was fourfold higher in the PAI-1 inhibitor-treated group than in the vehicle group (Fig. 7F). The inhibition of PAI-1 therefore enhances not only the rapid hematopoietic recovery in the early phase, but also the long-term self-renewal capacity of HSCs.
To compare the repopulating activity of long-term HSCs, a small number of primary Ly5.1+ donor-derived hematopoietic cells after limiting dilution was transplanted into Ly5.2+ mice (6 × 104 or 2 × 104 cells per mouse). Recipient Ly5.2+ mice were thought to be successfully engrafted if Ly5.1+ cells with the potential for multilineage differentiation were detected in excess of 1% of the total BM cells in the recipient (representative FACS profiles are shown in Supporting Information Fig. S8). The results demonstrated that the chimerism of Ly5.1+ cells was higher in the PAI-1 inhibitor-treated group (1.26% ± 0.1% with 2 × 104 cells and 3.45% ± 1.09% with 6 × 104 cells) than in the vehicle-treated group (0.38% ± 0.26% with 2 × 104 cells and 1.04% ± 0.19% with 6 × 104 cells) (Fig. 7G). In addition, successful engraftment was achieved more often in the PAI-1 inhibitor-treated group (five of the five mice with 2 × 104 and 6 × 104 cells) than in the vehicle-treated group (one of the five mice with 2 × 104 cells and two of the five mice with 6 × 104 cells). Collectively, our results of the phenotypic (i.e., cell surface analysis) and repopulating (i.e., transplantation) assays revealed that inhibiting the PAI-1 activity during the early phase of reconstitution efficiently expands the long-term repopulating HSCs.
This study was undertaken to elucidate the role of PAI-1, a negative regulator of the fibrinolytic pathway, in hematopoietic regeneration after irradiation-induced myeloablation. The results demonstrated the following: first, the expression levels of fibrinolytic factors, such as tPA and plasmin, are markedly augmented in endothelial and nonhematopoietic stromal cells, respectively, in the BM of mice after sublethal irradiation. Second, the PAI-1 expression is simultaneously upregulated in nonhematopoietic stromal cells, especially osteoblasts. Third, PAI-1 negatively regulates hematopoietic regeneration: the genetic deletion of PAI-1, as well as the administration of a PAI-1 inhibitor, activates the tPA-mediated fibrinolytic pathway, eventually accelerating hematopoietic regeneration. Finally, the inhibition of PAI-1 activity at an early phase of transplantation facilitates the recovery and maintenance of hematopoiesis (Supporting Information Fig. S9).
Previous studies have shown that the levels of fibrinolytic factors and their inhibitors are elevated in several tissues following irradiation [22-24]. However, the dynamics of fibrinolytic system factors, particularly PAI-1, in hematopoietic regeneration of the BM have not been elucidated. This study revealed, for the first time, that irradiation has a substantial impact on the fibrinolytic system in the BM. Upon irradiation, the levels of fibrinolytic factors and PAI-1 in the BM fluid increased dramatically (∼30-fold), in sharp contrast to the moderate increase in these factors in the plasma (1.8–3-fold). Of note, their levels in the BM fluid were 1 or 2 orders of magnitude higher than those in the plasma (e.g., 300 and 8 ng/ml of PAI-1 in the BM fluid and plasma, respectively), demonstrating that BM cells have a potent capacity to produce fibrinolytic factors and PAI-1 upon irradiation. In good agreement with these in vivo results, the results of the in vitro cultured cell experiments demonstrated that nonhematopoietic BM stromal cells readily produce tPA, plasmin, and PAI-1 after irradiation.
On closer inspection by immunohistochemistry, we identified nonhematopoietic cells, especially stromal cells, as the primary source of fibrinolytic factors and PAI-1. Our data in the irradiated PAI-1 KO mice (recipient), which have undetectable levels of PAI-1 in the plasma after transplantation of normal hematopoietic cells (PAI-1+/+), suggested that the marked increase in PAI-1 after irradiation does not originate from donor hematopoietic cells. Nonhematopoietic stromal cells in the BM thus play a significant role in repairing the myeloablated hematopoietic microenvironment. This is not surprising, because the irradiation of mice with a lethal dose of radiation eradicates the hematopoietic cells and their progenitor cells, but not nonhematopoietic cells [25, 26].
We also elucidated the mechanism by which PAI-1 regulates the hematopoietic regeneration in the BM. A previous study by Hattori and colleagues demonstrated that activation of the fibrinolytic pathway by administration of recombinant tPA results in the conversion of the transmembrane form to the soluble form of c-kitL , a hematopoiesis-promoting factor. They suggested that activated plasmin subsequently transforms MMP-2/9 into active forms, which in turn release c-kitL from stromal cells. It is therefore likely that PAI-1 activity prevents the hematopoietic regeneration in the BM by inhibiting the fibrinolytic pathway and c-kitL release. Our results showing that the suppression of PAI-1 activity either by a pharmacological approach or by the genetic deletion of the PAI-1 gene can elevate the tPA/plasmin activity, promote c-kitL production, and lead to the rapid recovery of hematopoiesis after myeloablation. Although other products of fibrinolytic degradation may also promote engraftment, these results support the critical role of c-kitL production in the improvement of engraftment by PAI-1 inhibition.
Emerging evidence suggests that PAI-1 expression generally limits tissue repair by negatively regulating the fibrinolytic environment and by inhibiting cell migration . For example, the skin wound healing process is accelerated in PAI-1-deficient mice . PAI-1 KO mice are also protected against liver fibrosis  and radiation enteropathy . In terms of tissue regeneration, Galipeau and colleagues have demonstrated that mesenchymal stem cells (MSCs) derived from PAI-1 KO mice exhibited higher regenerative potential than those from WT mice . Furthermore, chemical manipulation of the PAI-1 activity improves the engraftment of MSCs, defining PAI-1 as a negative regulator of transplanted stem cell survival in vivo . This study clarified the active involvement of PAI-1 in the hematopoietic regeneration after irradiation.
Proper treatment of the initial stage of hematopoietic recovery and the prevention of premature HSC exhaustion could therefore significantly improve the clinical outcome of transplantation [1, 2]. In this regard, our study demonstrated that, despite a short period of administration, the suppression of the PAI-1 activity by a low molecular weight compound could induce both rapid hematopoietic regeneration through increased cycling of HSCs and expansion of the long-term HSCs. This opens a new avenue for improving HSCT. It should be emphasized that the PAI-1 inhibitor does not induce HSC exhaustion or malignancy in spite of its potent ability to increase the cycling of HSCs. The results of this study clearly demonstrated that c-kit+ HSPCs in the group treated with the PAI-1 inhibitor preferentially localized to the BM niche, just like in the vehicle-treated group, suggesting that the interaction between the hematopoietic progenitor cells and niche is maintained even after the treatment with a PAI-1 inhibitor. This may be a plausible explanation for why the PAI-1 inhibitor does not induce HSC exhaustion.
Both the PAI-1 inhibitor and tPA theoretically activate the fibrinolytic pathway and the subsequent hematopoietic regeneration, but their effects in vivo in animals appear to be different. This is partly explained by the differences in the routes of administration between tPA and the PAI-1 inhibitor, as well as the doses, mechanisms of action, and/or half-lives of these agents. Recombinant tPA is administered intravenously (a large amount of tPA is given directly into the circulation), and immediately activates the fibrinolytic pathway, but its half-life is only a few minutes . In contrast, the PAI-1 inhibitor was given orally, and was absorbed in the gut, entered into the circulation gradually, inhibited the PAI-1 moiety, and subsequently upregulated the tPA activity leading to its effects on the fibrinolytic pathway. The half-life of the PAI-1 inhibitor is much longer (6.5 hour) than that of tPA.
It is also important to note that tPA administration itself increased the PAI-1 level, suggesting a potential negative feedback effect in this pathway and limits to the therapeutic benefits of tPA for hematopoietic regeneration. In addition, the repopulating capacity of HSCs in tPA-treated mice showed a slight decrease, suggesting that tPA treatment may induce HSC exhaustion. It should also be mentioned that PAI-1 regulates not only tPA, but also other proteins (i.e., vitronectin, urokinase-type plasminogen activator (uPA), and low density lipoprotein receptor (LDLR)) [5, 31, 32] and thereby has an impact on broader biological systems.
In conclusion, our study provides the first direct evidence that PAI-1 is a negative regulator of hematopoietic regeneration, and that the inhibition of PAI-1 activity, either genetically or by a low molecular weight compound, significantly improves donor-derived hematopoiesis after transplantation. Our findings give new insights into the treatment of HSCT and for clinical transplantation medicine.
We appreciate the help of Dr. Koichi Hattori (Institute of Medical Science, University of Tokyo, Japan) for kindly providing the tPA KO mice. We thank Dr. Nobuo Watanabe (Tokai University School of Medicine, Japan) for helpful discussion and wrote the manuscript. We also thank the members of the Research Center for Regenerative Medicine of Tokai University, especially Tomomi Takanashi, Kozue Hiyama, and Tomoko Uno, for the technical support. We thank the members of the Animal Care Center of Tokai University for their meticulous care of the experimental animals. This work was supported by Japanese Grants-in-Aid for Scientific Research from the Ministry of Education, Culture, Sports, Science and Technology (MEXT), from the Ministry of Health, Labor and Welfare (MHLW), from the National Institute of Biomedical Innovation (NIBIO), from the Japan Science and Technology Agency (JST), and from the Tokai University School of Medicine Research Aid. Special thanks are due to the MARA Education Foundation, Malaysia, for supporting the scholarship awarded to A.A.I.
A.A.I.: collection and assembly of data, data analysis and interpretation, and manuscript writing; T.Y.: conception and design, data analysis and interpretation, manuscript writing, and financial support; M.O.: data analysis and interpretation; T.D.: provision of study material; C.v.Y.d.S.: manuscript writing; T.M.: provision of study material, data analysis and interpretation, and manuscript writing; K.A.: conception and design, data analysis and interpretation, financial support, and final approval of manuscript.
Disclosure of Potential Conflicts of Interest
The authors indicate no potential conflicts of interest.