Author contributions: D.P.: conception and design, collection and assembly of data, data analysis and interpretation, manuscript writing, and final approval of manuscript; K.L. and G.B.: collection and assembly of data and data analysis and interpretation; H.B.: collection and assembly of data; D.H.: conception and design, financial support, collection and assembly of data, data analysis and interpretation, manuscript writing, and final approval of manuscript.
Disclosure of potential conflicts of interest is found at the end of this article.
First published online in STEM CELLSEXPRESS August 16, 2012.
Umbilical cord blood (UCB) represents a readily available source of hematopoietic and endothelial precursors at early ontogeny. Understanding the proangiogenic functions of these somatic progenitor subtypes after transplantation is integral to the development of improved cell-based therapies to treat ischemic diseases. We used fluorescence-activated cell sorting to purify a rare (<0.5%) population of UCB cells with high aldehyde dehydrogenase (ALDHhi) activity, a conserved stem/progenitor cell function. ALDHhi cells were depleted of mature monocytes and T- and B-lymphocytes and were enriched for early myeloid (CD33) and stem cell-associated (CD34, CD133, and CD117) phenotypes. Although these cells were primarily hematopoietic in origin, UCB ALDHhi cells demonstrated a proangiogenic transcription profile and were highly enriched for both multipotent myeloid and endothelial colony-forming cells in vitro. Coculture of ALDHhi cells in hanging transwells promoted the survival of human umbilical vein endothelial cells (HUVEC) under growth factor-free and serum-free conditions. On growth factor depleted matrigel, ALDHhi cells significantly increased tube-like cord formation by HUVEC. After induction of acute unilateral hind limb ischemia by femoral artery ligation, transplantation of ALDHhi cells significantly enhanced the recovery of perfusion in ischemic limbs. Despite transient engraftment in the ischemic hind limb, early recruitment of ALDHhi cells into ischemic muscle tissue correlated with increased murine von Willebrand factor blood vessel and CD31+ capillary densities. Thus, UCB ALDHhi cells represent a readily available population of proangiogenic progenitors that promote vascular regeneration. This work provides preclinical justification for the development of therapeutic strategies to treat ischemic diseases using UCB-derived ALDHhi mixed progenitor cells. STEM Cells2012;30:2248–2260
Ischemic disease is characterized by the reduction of blood flow to the heart or peripheral tissues and encompasses life-threatening disorders such as ischemic heart disease and critical limb ischemia. Despite advances in pharmacological and surgical management, ischemic disease remains one of the leading causes of morbidity and mortality worldwide [1, 2]. Thus, novel therapies to promote the regeneration of damaged vasculature are under intense preclinical investigation [3–5].
Angiogenesis, the sprouting of new blood vessels from pre-existing vessels, is a central process for tissue repair, allowing the delivery of circulating proangiogenic cells that form a vascular regenerative microenvironment . Asahara et al. first described human bone marrow (BM)-derived cells that promoted the recovery of blood flow and colocalized with new vessels after transplantation in vivo . Later studies have shown that these rare cells, termed endothelial precursor cells (EPC), express the cell markers CD34, CD133, and VEGFR2 . Because these markers were also expressed on clonally distinct cell types of both hematopoietic and nonhematopoietic lineages, there has been considerable controversy over which cell subtypes best promote vascular regeneration.
Endothelial colony-forming cells (ECFC) are rare blood-derived cells that form late-outgrowth adherent colonies of proliferative cells when propagated under strict endothelial cell (EC) growth conditions and can be isolated at increasing frequency from human peripheral blood, BM, and umbilical cord blood (UCB). ECFCs are distinguished from proangiogenic hematopoietic cell types in human blood as they are exclusively CD45− and can integrate directly into perfused vessels in vivo. Recently, it has been demonstrated that many of the cell types previously referred to as EPC are not EC precursors per se but represent myeloid/macrophage lineage cells that promote vascular repair through proposed paracrine signaling to vessel-derived EC [8, 9, 10–12]. In mouse models, circulating CXCR4+/VEGFR1+ hematopoietic cells are recruited to sprouting vessels by secretion of stromal-derived factor 1 or vascular endothelial growth factor-A (VEGF-A) produced by damaged vasculature [4, 13]. Studies examining tumor metastasis have demonstrated that even small numbers of these hematopoietic cells play a major role in establishing a permissive niche for tumor vascularization [14–16].
Numerous studies have established that a decreased frequency or impaired function of circulating CD133+ or CD34+ cells is associated with increased cardiovascular risk in patients with ischemic heart disease and diabetes [2, 9, 17–20]. This, in addition to preclinical data showing the proangiogenic potential of BM-derived cells, has led to a number of clinical trials investigating treatment of cardiac [21–23] and limb ischemia [24, 25] by transplantation of heterogeneous human BM mononuclear cells (MNCs). Although these studies have demonstrated safety in the autologous setting, they have shown only modest improvements in clinical endpoints . Thus, purification of specific proangiogenic cellular subtypes from nonautologous sources and further characterization of specific subtype contribution during the coordination of vascular repair are warranted to improve current cell-based therapies for ischemic diseases.
We have recently shown that UCB represents an excellent source of regenerative progenitor cells . Early in ontogeny, UCB has been shown to contain a higher frequency of ECFC , and UCB-derived ECFCs have more robust vessel-forming capacity than ECFC from adult blood [9, 29]. Due to recent widespread initiatives to HLA-phenotype and cryopreserve this material, UCB now represents a readily available source of progenitor cells for cell therapy applications. In addition, allogeneic UCB progenitor cell transplantation may have distinct advantages over autologous BM in patients with obesity, chronic diabetes, or vascular disease-related pathologies due to compromised progenitor cell content and function [30–33].
Using high aldehyde dehydrogenase (ALDHhi) activity, a cytosolic enzyme involved in retinoic acid metabolism and cellular self-protection from oxidative damage [34–39], we have developed fluorescence-activated cell sorting (FACS) strategies to simultaneously enrich for progenitor cells from hematopoietic, endothelial, and mesenchymal lineages . We have demonstrated that transplanted ALDHhi progenitors reconstitute hematopoiesis [34, 35], exhibit widespread tissue distribution , and recruit to areas of ischemia after transplantation . Furthermore, BM-derived ALDHhi cells can augment revascularization and blood flow of ischemic limbs in transplanted mice with acute ischemic injury . Recently, we proposed a model of coordinated neovascularization after human progenitor cell transplantation wherein hematopoietic and nonhematopoietic progenitors act in concert to formulate a proangiogenic niche for vascular regeneration .
Here, we show that selection of human UCB cells with high ALDH activity simultaneously purifies a mixed population of primitive hematopoietic colony-forming cells (HCFCs) and ECFCs with potent proangiogenic functions. UCB ALDHhi progenitors demonstrated a proangiogenic transcription profile and supported EC survival in liquid culture and tube-like cord formation in growth factor-reduced matrigel. Transplanted ALDHhi cells improved recovery of blood flow after femoral artery ligation in non-obese diabetic/severe combined immunodeficient (NOD/SCID) mice. Thus, UCB progenitors purified by high ALDH-activity represent an allogeneic population of proangiogenic cells for the revascularization of ischemic tissues.
MATERIALS AND METHODS
Progenitor Cell Isolation from Human UCB
UCB was collected after informed consent by venipuncture at the London Health Sciences Birthing Centre. The Human Research Ethics Committee at the University of Western Ontario approved all studies. After hypaque-ficoll centrifugation, MNCs were depleted of erythrocytes by ammonium chloride lysis and assayed for ALDH activity using Aldefluor reagent (Stem Cell Technologies, Vancouver, Canada, stemcell.com), as previously described . FACS was performed to isolate cells with low side scatter and high (ALDHhi) versus low (ALDHlo) ALDH activity to >99% purity. Sorted cells were washed in phosphate-buffered saline (PBS) to allow efflux of Aldefluor substrate via reactivation of inhibited adenosine-5′-triphosphate-binding cassette-transporters.
Cell Surface Phenotype Analysis
UCB MNC, ALDHhi or ALDHlo cells were labeled with anti-human antibodies against pan-leukocyte marker CD45 and colabeled with CD4 and CD8 (T-cells), or CD19 and CD20 (B cells), or CD33 (myeloid) and CD14 (monocytes), or CD31 (PECAM-1) and CD144 (VE-cadherin, GeneTex, Irvine, CA, genetex.com), or CD34 and CD38, or CD117 (c-kit) and CD133 (Miltenyi Biotec, Cologne, Germany, miltenyibiotec.com), or CXCR4 and VEGFR2. All antibodies were from BD Biosciences unless otherwise indicated. Cell surface marker expression was measured using a FACSCalibur cytometer (BD Biosciences, San Jose, CA, www.bdbiosciences.com) and analyzed using FlowJo software (Tree Star, Ashland, OR, flowjo.com).
Colony-Forming Cell Assays
UCB MNC, ALDHhi or ALDHlo cells were cultured in methylcellulose (Methocult H4434, Stem Cell Technologies) for 14 days and hematopoietic colony formation was enumerated based on cell morphology identifying burst-forming units of erythrocytes, colony-forming units of granulocytes, macrophages, granulocytes and macrophages, or multilineage mixed colonies containing cells of all three types. ECFCs were enumerated after adherent culture for 14 days in complete endothelial growth media (EGM-2 + 2% fetal bovine serum + epidermal growth factor [EGF], VEGF-A, fibroblast growth factor [FGF], insulin-like growth factor-1 [IGF-1], Lonza, Basel, Switzerland, www.lonza.com), as previously described .
mRNA Isolation and Microarray Analyses
mRNA was from three sample-matched UCB ALDHhi and ALDHlo cell cohorts using mRNeasy Mini kits (Qiagen, Dusseldorf, Germany, qiagen.com). Microarrays were performed in triplicate using 18 human gene array chips (Affymetrix, Santa Clara, CA, affymetrix.com) at the London Regional Genomics Facility. Transcripts differentially expressed by UCB ALDHhi versus ALDHlo cells were analyzed using Partek Genomics Suite (Partek Inc., St Louis, MO, partek.com).
Coculture Assays with Human Umbilical Vein ECs
To assess paracrine functions of UCB cells on EC (human umbilical vein EC, HUVEC) survival and proliferation during coculture in vitro, 40 × 103 HUVEC were cultured in endothelial basal media (EBM2, Lonza) with or without angiogenic growth factors (EGF, VEGF-A, FGF, and IGF-1) and serum (2%) supplementation. Under each condition, 105 ALDHhi or 105 ALDHlo cells were seeded into porous 5 μm hanging transwells suspended above the HUVEC. The number of trypan blue excluding HUVEC present after 72 hours coculture was enumerated by blinded hemocytometer counts. For tube-like cord formation assays, 50 × 103 HUVEC, or HUVEC supplemented with 20 × 103 ALDHhi or ALDHlo cells were seeded on growth factor-reduced Matrigel (BD Biosciences) and cultured for 72 hours. The number of complete branch points in the cord structures formed by HUVEC was enumerated on an inverted microscope by blinded manual counts of four fields of view at 24, 48, and 72 hours. All coculture experiments were performed in triplicate.
Murine Femoral Artery Ligation Surgery and Transplantation of Human UCB Cells
Right femoral and saphenous artery and vein ligation with complete excision of the femoral artery and vein was performed on anesthetized NOD/SCID (Jackson Laboratory, Bar Harbor, ME, jax.org) or NOD/SCID/mucopolysaccharidosis type VII (MPSVII) mice as previously described . Within 24 hours of surgery, nonirradiated or sublethally irradiated (300 cGy) mice were transplanted by tail vein injection with PBS, 2 × 105 ALDHhi cells, 10 × 106 ALDHlo cells, or 20 × 106 MNC. Cell doses were chosen based on the relative frequency of each cell type within UCB MNC. NOD/SCID mice were used for functional studies to monitor the recovery of perfusion and to quantify blood vessel density as described below. NOD/SCID/MPSVII mice were used to detect human cell engraftment kinetics and localization in muscle sections from ischemic limbs using β-glucuronidase (GUSB) staining as described below.
Quantification of Hind Limb Perfusion Using Laser Doppler Perfusion Imaging
For laser Doppler perfusion imaging (LDPI; MoorLDI-2; Moor Instruments, Devon, U.K., moor.co.uk), anesthetized NOD/SCID mice were heated at 37°C for 5 minutes and blood flow was quantified within the hind quarters of live mice after surgery, and at 3, 7, 14, 21, and 28 days post-transplantation. Perfusion ratios (PRs) comparing blood flow in the ischemic versus nonischemic limb were quantified by averaged units of flux using Moorflow software (Moor Instruments).
Quantification of Blood Vessel Density
Adductor muscles from transplanted mice were embedded and frozen in optimum cutting temperature medium (Tissue Tek, Sakura Finetek, Tokyo, Japan, sakura-finetek.com), cryosectioned, fixed in 10% formalin (Sigma), and blocked with mouse-on-mouse reagent (Vector Labs, Burlingame, CA, vectorlabs.com). Murine blood vessel density in ischemic and nonischemic limbs was quantified at day 28 by counting mouse CD31+ capillaries and mouse von Willebrand factor (vWF+) blood vessels using rat anti-mouse CD31 (1:100; BD Biosciences) or rabbit anti-mouse vWF antibody (1:200; Millipore, Billerica, MA, millipore.com) and peroxidase-labeled anti-rat or anti-rabbit IgG secondary antibodies (Vector Labs) visualized using 3,3′-diaminobenzidine substrate (Vector), and counterstained with hematoxylin. Blood vessel density was counted in a blinded fashion from nine photomicrographic fields using light microscopy.
Quantification of Human Cell Engraftment by Flow Cytometry
BM and adductor muscles from the ischemic and nonischemic limbs were analyzed for human cell engraftment at 3, 7, and 28 days after transplantation by flow cytometric detection of human CD45 and HLA-A,B,C in combination with 7-AAD viability dye (BD Biosciences), as previously described . Muscle samples were digested with type II collagenase (Worthington Biochemical, Lakewood, NJ, worthington-biochem.com) prior to antibody labeling.
Detection of Human Cells by GUSB Activity
Transplanted NOD/SCID/MPSVII mice with unilateral hind limb ischemia were sacrificed at 3, 7, and 28 days after transplantation to allow histochemical detection of single human cell engraftment in ischemic muscle sections (GUSB-deficient) based on the detection of ubiquitous GUSB activity in human cells as previously described . Muscle sections were costained with vWF to assess the distribution of engrafted human cells with respect to murine vasculature and with human CD45 to assess the lineage of engrafted cells.
Analysis of significance was performed by one-way analysis of variance (ANOVA) for cell surface marker and progenitor frequency assays and by two-way ANOVA with Bonferroni post hoc tests for all other assays using Graphpad Prism software.
UCB ALDHhi Cells Were Enriched for Myeloid Progenitor Cell Surface Phenotypes
Human UCB MNCs were collected without prior lineage depletion  and purified based on cells with low side scatter (nongranular) and with low (ALDHlo, 41.36% ± 3.6%) versus high (ALDHhi, 0.4% ± 0.1%) ALDH activity compared to ALDH-inhibited (diethylaminobenzaldehyde, DEAB) controls (n = 10, Fig. 1A, 1B). Compared to unpurified UCB MNC (97.7% ± 0.8% CD45+) or ALDHlo cells (99.1% ± 0.5% CD45+), ALDHhi cells showed lower expression of the pan-leukocyte marker CD45 (93.3% ± 1.9% CD45+; *, p <.05; n = 4). Thus, sorted cell populations were gated into hematopoietic (CD45+) and nonhematopoietic (CD45−) subsets and further analyzed for primitive versus mature hematopoietic or EC surface marker expression.
CD45+ ALDHhi cells showed reduced expression lineage-specific markers for T- (CD4 and CD8, Fig. 1C) and B-lymphocytes (CD19 and CD20, Fig. 1D) compared to unsorted MNC or ALDHlo cells (**, p <.01; Table 1A). More than 90% of CD45+ ALDHhi cells expressed CD33, an early myeloid lineage marker, but coexpression of the mature monocyte marker CD14 was low (3.3% ± 1.0%, Fig. 1E; Table 1A). Similarly, ALDHhi cells showed near homogeneous expression of CD31 or PECAM-1 (96.7% ± 2.9%) but low coexpression of CD144/VE-cadherin (3.2% ± 1.0%, Fig. 1F; Table 1A). Notably, the CD45+ ALDHhi cells were highly enriched for the coexpression of the progenitor markers CD34 (89.6% ± 5.9%, Fig. 1G), CD117 (80.2% ± 4.9%), and CD133 (61.5% ± 2.3%, Fig. 1H) compared to unsorted UCB MNC or ALDHlo cells (*, p <.05, Table 1A). CD45+ ALDHhi cells were enriched for SCID repopulating progenitor phenotype (CD34+CD38−, Fig. 1G) , and ALDHhiCD34+ cells showed coexpression of the homing marker CXCR4 (14.2% ± 2.9%, Fig. 1I) but low expression of VEGFR2/kinase insert domain receptor (KDR) (0.5% ± 0.1%, Fig. 1J).
Table 1. Cell surface marker expression for (A) hematopoietic (gated CD45+) component and (B) nonhematopoietic (gated CD45−) component of human UCB MNC, ALDHhi cells, and ALDHlo cells
Human UCB MNC, or ALDHhi or ALDHlo cells were stained with anti-human antibodies to assess the cell surface marker phenotype on (A) hematopoietic and (B) nonhematopoietic cells. UCB ALDHhi hematopoietic cells were depleted of mature T- and B-lymphocytes and monocytes and were highly enriched for the early myeloid marker CD33 and platelet-derived endothelial cell adhesion molecule CD31 (PECAM-1). UCB ALDHhi hematopoietic and nonhematopoietic cells were enriched for progenitor cell markers CD34, CD117, and CD133. Data represents mean ± SEM (n = 4), different letters denote different expression patterns at *, p <.05; **, p <.01 compared to unpurified MNC.
CD45− cells, representing approximately 7% of the ALDHhi population, also showed enriched expression of CD34 (65.0% ± 14.8%), CD117 (64.9% ± 12.4%), and CD133 (22.5% ± 3.9%) progenitor markers (*, p <.05) but did not show increased expression of the mature EC-associated adhesion molecules (CD31 and CD144) compared to UCB MNC or ALDHlo cells (Table 1B). Collectively, these data indicate that the hematopoietic fraction of ALDHhi cells was depleted of mature lymphocytes and monocytes yet enriched for primitive myeloid progenitor phenotype, and the CD45− portion of the ALDHhi cells similarly retained a primitive nonhematopoietic cell phenotype.
UCB ALDHhi Cells Demonstrated Hematopoietic and Endothelial Colony Formation
UCB MNC, ALDHlo or ALDHhi cells were plated under lineage-specific growth conditions for 14 days to assess HCFC and ECFC frequency. After culture in methylcellulose, ALDHhi cells were highly enriched for myeloid lineage colony-forming cells (one HCFC in 3.6 cells) compared to ALDHlo cells (one HCFC in 6.6 × 103 cells; **, p <.01; n = 4). ALDHhi HCFC demonstrated multipotent lineage differentiation with increased production of erythroid, macrophage, granulocyte, and mixed morphologies (***, p <.001; Fig. 1K). To assess ECFC function in vitro, ALDH-purified UCB cells were cultured under stringent conditions established by Yoder et al. . Colony enumeration revealed increased ECFC content in ALDHhi cells (one ECFC in 4.6 × 104 cells) compared to ALDHlo cells (one ECFC in 2.2 × 105 cells; *, p <.05; Fig. 1L). ECFC progenies were highly proliferative, did not express hematopoietic markers (CD45 and CD14), and expressed EC markers (CD31, CD105, and CD144, data not shown). Unlike human BM , MNC or ALDH-sorted cohorts from UCB did not establish plastic adherent multipotent stromal cell colonies in vitro (n = 10). Thus, UCB ALDHhi cells represented a mixture of cells with enriched HCFC and ECFC function.
UCB ALDHhi Cells Demonstrated a Proangiogenic Transcription Profile
Affymetrix microarray identified 1,253 transcripts with significantly increased (*, p <.05) expression in three UCB ALDHhi cell samples compared to sample-matched ALDHlo cells. These transcripts were first filtered to include unique genes translating to plasma membrane bound (196 hits) or extracellular (108 hits) locations, and subsequently filtered for documented angiogenic biological functions (Supporting Information Tables S1, S2). As a control for the selection of cells with high ALDH-activity, ALDH1A1 mRNA expression was increased 42-fold in ALDHhi versus ALDHlo cells (p = 2.08 × 10−5). Furthermore, significantly increased CD34, PROM1, and KIT mRNA expression in ALDHhi cells (Supporting Information Table S1) correlated with increased CD34, CD133, and c-kit/CD117 cell surface protein expression previously observed by FACS (Fig. 1, Table 1). Transcripts associated with early myeloid progenitor differentiation (FLT3, CSF3R, and CD33), integrin-mediated cell adhesion (ITGA9, ITGAV, and ITGA2B), and vascular cell signaling (TIE1, TIE2, FGFR1, and endoglin) were also highly expressed by UCB ALDHhi cells (Supporting Information Table S1). Interestingly, transcripts for secreted cytokines associated with vascular functions (ANGPT1, FGF16, and VEGFA), immune cell migration (IL18, IL1β, IL8, CSF1, and CXCL2), and modulators of EGF and IGF signaling (EREG, AREG, HBEGF, and IGFBP7) were also increased in UCB ALDHhi cells (Supporting Information Table S2). Collectively, these data suggested that UCB ALDHhi cells possessed a proangiogenic transcription profile compared to ALDHlo cells.
UCB ALDHhi Cells Augment the Survival of HUVEC In Vitro
To assess the paracrine functions of ALDH-purified UCB cells on EC survival and proliferation in vitro, 40 × 103 HUVECs were seeded into transwell cocultures for 72 hours using a porous 5 μm hanging insert containing 105 UCB ALDHhi or ALDHlo cells (or basal media control) suspended above HUVEC. Under optimal conditions containing growth factors (GF+) and serum, noncontact coculture with ALDHlo or ALDHhi cells did not impact the fivefold expansion of HUVEC observed over 72 hours (Fig. 2A, 2E). Conversely, in GF−/serum-free conditions with or without ALDHlo cell supplementation, HUVEC invariably died within 72 hours (Fig. 2B, 2E). In contrast, coculture with ALDHhi cells under GF−/serum-free conditions preserved HUVEC viability for 72 hours (*, p <.05; Fig. 2B, 2E). Under GF+/serum-free conditions, ALDHhi cell coculture promoted a modest expansion (1.7-fold) of HUVEC (67.4 ± 7.1 × 103 cells), compared to more than threefold HUVEC loss with basal media (12.4 ± 2.6 × 103 cells) or ALDHlo cell supplementation (14.4 ± 2.8 × 103 cells; *, p <.05; Fig. 2C, 2E). Under GF−/serum+ conditions, ALDHlo or ALDHhi coculture did not significantly alter HUVEC numbers (Fig. 2D, 2E). Thus, noncontact coculture with human ALDHhi cells promoted HUVEC survival under GF−/serum-free conditions. These effects were consistent with the prosurvival stimulus in HUVEC provided by VEGF-A treatment , a central proangiogenic cytokine with increased mRNA-expression in UCB ALDHhi cells (Supporting Information Table S2).
UCB ALDHhi Cells Augment Tube-Like Cord Formation by HUVEC In Vitro
HUVECs (50 × 103) were cultured alone or cocultured with ALDHhi or ALDHlo cells (20 × 103) on growth factor-reduced Matrigel. HUVEC established partially completed cord networks within 24 hours (Fig. 3A). Although coculture of HUVEC with UCB ALDHlo cells did not significantly alter cord formation, supplementation with ALDHhi cells augmented HUVEC tube-like cord formation at 24 hours (Fig. 3B, 3C). Although UCB ALDHhi cells established ECFC in liquid culture (Fig. 1L, inset), UCB ALDHlo or ALDHhi cells seeded alone on growth factor-reduced Matrigel did not spontaneously align to form cord networks (Fig. 3D), and tube-like cord stability of HUVEC was not enhanced by coculture with UCB ALDHlo or ALDHhi at 72 hours (Fig. 3E–3H). Enumeration of completed branch points at 24 and 48 hours revealed that coculture with UCB ALDHhi cells augmented cord formation by HUVEC compared to media controls (n = 6; *, p <.05, Fig. 3H). Therefore, coculture with UCB ALDHhi cells promoted the cord-forming function of HUVEC in vitro.
Transplanted UCB ALDHhi Cells Enhanced the Recovery of Hind Limb Perfusion
To determine whether ALDH-purified human UCB cells could support blood vessel regeneration after intravenous transplantation in vivo, we induced acute unilateral hind limb ischemia in NOD/SCID mice by right femoral artery ligation with complete excision of a 0.5-cm section of the femoral artery and vein. Blood flow was assessed by LDPI and the PR in the ischemic versus control limb was decreased >10-fold (PR = 0.08 ± 0.02) postsurgery confirming consistent induction of unilateral hind limb ischemia (Fig. 4). Within 24 hours of surgery, mice were tail vein injected with PBS vehicle control, unsorted UCB MNC, ALDHhi cells, or ALDHlo cells. The representative LDPI shown in Figure 4 documents the recovery of limb perfusion after transplantation. PBS-injected control mice showed recovery of PR from 0.07 ± 0.01 after surgery to 0.36 ± 0.05 by day 7 postsurgery without subsequent improvement at later time points (Fig. 4A, 4E). This baseline recovery of perfusion was sufficient to prevent excess morbidity or limb loss in PBS transplanted mice. Mice transplanted with 20 × 106 MNC or 10 × 106 ALDHlo cells showed equivalent recovery of limb perfusion similarly compared to PBS controls (Fig. 4B, 4D, 4E). In contrast, mice transplanted with 50–100-fold fewer ALDHhi cells (2 × 105) showed increased perfusion by day 14 post-transplantation (PR = 0.58 ± 0.07) compared to all other treatments and augmented perfusion was maintained for 28 days post-transplantation (PR = 0.64 ± 0.06; **, p <.01; Fig. 4C, 4E). Thus, intravenous transplantation of human UCB ALDHhi cells improved the limb perfusion in mice with acute unilateral hind limb ischemia.
Transplanted UCB ALDHhi Cells Increased Capillary Density in the Ischemic Limb
To address the recovery of blood vessels in the ischemic hind limb, sections from the adductor muscle at the site of artery ligation were stained for both murine CD31+ capillary (Fig. 5A–5C) and vWF+ blood vessel (Fig. 5D–5F) density. Compared to the nonischemic limb at day 28, PBS-injected mice showed significant reduction (*, p <.05) of both CD31+ capillaries (Fig. 5G) and vWF+ blood vessels (Fig. 5F) within the ischemic hind limb. However, mice transplanted with UCB ALDHhi cells had significantly more CD31+ capillaries compared to mice transplanted with ALDHlo cells (*, p <.05) or PBS (**, p <.01; Fig. 5G). Similarly, mice transplanted with ALDHhi cells showed higher vWF+ blood vessel density than mice injected with ALDHlo cells or MNC (*, p <.05), or PBS-injected controls (***, p <.001, Fig. 5H). Despite only partial recovery of perfusion, CD31+ and vWF+ vessel density in the ischemic limbs of ALDHhi cell-treated mice were equivalent to the vessel density of the nonischemic control limb at day 28 (Fig. 5G, 5H). Thus, UCB ALDHhi cell transplantation increased both CD31+ capillary and vWF+ blood vessel density in the ischemic hind limb after femoral artery ligation.
Transplanted UCB ALDHhi Cells Were Recruited to the Ischemic Muscle at Low Frequency
In order to correlate the recovery of perfusion with human cell engraftment, murine BM and adductor muscle were collected at 3, 7, and 28 days after transplantation and analyzed for the presence of human CD45+ or HLA-A,B,C+ cells by flow cytometry. Despite consistent human cell hematopoietic chimerism in the BM of mice transplanted with ALDHhi cells (Fig. 6A, 6C) or MNC (Fig. 6C), human cells were not significantly detected in the adductor muscle of ischemic limbs by flow cytometry (Fig. 6B, 6D) at day 28. In addition, mice processed for engraftment at earlier time points showed no evidence of human cells in the ischemic limb at three (n = 3) or 7 days (n = 3) post-transplantation by FACS. Therefore, we repeated engraftment experiments using identical cell doses in the GUSB-deficient NOD/SCID/MPSVII recipient to permit the detection of transplanted human cells at single-cell resolution. At 3 and 7 days post-transplantation, the detection of GUSB-expressing cells confirmed the recruitment of human cells to the ischemic hind limbs of mice that received MNC (n = 4, Fig. 6E, 6F) or ALDHhi cells (n = 4, Fig. 6H, 6I). However, human cells were not detected in the ischemic limbs of mice that received ALDHlo cells (n = 4). Although transplanted cells were never detected in the nonischemic limb (data not shown), engrafted MNC and ALDHhi cells were observed in ischemic muscle sections surviving out to 28 days (Fig. 6G, 6J). Staining for GUSB-activity combined with murine vWF showed engraftment of single human cells between muscle fibers (Fig. 6K) and in association with the connective tissue surrounding blood vessels (Fig. 6L). Also, the majority of GUSB+ human cells surviving at day 28 costained for CD45 (Fig. 6M). Despite low frequency engraftment in the ischemic region, early recruitment of ALDHhi progenitor cells to the site of ischemic injury was sufficient to augment blood vessel density and improve limb perfusion.
Sublethal Irradiation Did Not Improve Perfusion in the Ischemic Hind Limb
The requirement for preparative irradiation is an important consideration in the design of clinically applicable cell therapies for ischemic diseases. In an attempt to improve human cell engraftment in the ischemic limb, transplantation experiments were repeated after sublethal irradiation (300 cGy) administered as a means to reduce xenorejection by residual innate immunity or natural killer-cell activity in the NOD/SCID model. Surprisingly, in mice receiving UCB ALDHhi cells, perfusion at early time points (days 3, 7, and 14) was similar for all transplanted groups (Supporting Information Fig. S1A–S1E). Although perfusion improved in the ALDHhi cell transplanted cohort at day 21, perfusion was reduced to baseline by day 28 (Supporting Information Fig. S1E), indicating that only a transient improvement in limb perfusion was observed. Compared to mice transplanted without preparative irradiation (Fig. 4), irradiated mice receiving an equal dose of UCB ALDHhi cells actually showed delayed recovery of perfusion at day 14 (PR300 cGy = 0.40 ± 0.05 vs. PR0 cGy = 0.58 ± 0.07). Nonetheless, at day 28, irradiated mice injected with ALDHhi cells still showed improved vWF+ blood vessel density (Supporting Information Fig. S2) similar to nonirradiated mice.
Irradiated mice also showed enhanced recruitment of ALDHhi cells to the BM at day 7 (1.3% ± 0.2%, Supporting Information Fig. S3A, S3C), resulting in extensive hematopoietic repopulation by day 28 (44.7% ± 2.0%, Supporting Information Fig. S3A, S3C). Unexpectedly, human cells in the ischemic limb were not detected by FACS or GUSB staining at day 7 (Supporting Information Fig. S3B, S3D, S3E) and were only rarely detected at day 28 (Supporting Information Fig. S3F). Thus, irradiation prior to ALDHhi cells transplantation did not augment engraftment or the recovery of perfusion in the ischemic limb.
In this study, we demonstrate that ALDH-expressing cells from human UCB, representing <0.5% of total UCB cells and primarily composed of primitive myeloid progenitor cells, demonstrated increased expression of mRNA transcripts for several angiogenic cytokines and stimulated the survival and tubule-forming functions of EC during coculture in vitro. Intravenous transplantation of UCB ALDHhi cells also augmented recovery of limb perfusion induced by femoral artery ligation and transection. Although the engraftment of human cells within ischemic muscle was infrequent, UCB ALDHhi cells were specifically recruited to the ischemic limb within 3 days of transplantation and stimulated the recovery of blood vessel and capillary density. Furthermore, ALDHhi cells function within the pancreas to promote islet revascularization after transplantation into streptozotocin-treated mice [45, 46]. Collectively, these data suggest that UCB ALDHhi progenitor cells represent a readily available population of proangiogenic cells for the development of cellular therapies to promote endogenous revascularization.
The UCB ALDHhi population was significantly enriched for human hematopoietic progenitor function. Approximately one in four ALDHhi cells demonstrated multipotent CFU capacity in methylcellulose cultures. UCB ALDHhi cells also showed increased expression of primitive cell surface markers (CD34, c-kit/CD117, and CD133), and >90% of UCB ALDHhi cells expressed the early myeloid marker CD33, indicating a substantial enrichment of early myeloid progenitors. At the same time, UCB ALDHhi cells were depleted of mature T- and B-lymphocytes as well as CD14-expressing monocytes. This is notable as CD14+ monocytes have been specifically implicated in the paracrine support of angiogenesis . However, transplantation of UCB ALDHhi cells, containing only 3% CD14+ monocytic cells, significantly improved perfusion and blood vessel density in the ischemic limb, whereas transplantation of mice with 100-fold higher doses of unsorted MNC, containing more than 17% CD14+ monocytes, failed to promote recovery from hind limb ischemia. Recent studies have demonstrated that murine common myeloid progenitors preferentially differentiate into proangiogenic monocytes that support neovessel formation in vivo . Although UCB ALDHhi cells may be capable of differentiating into monocytes in vivo, our data indicate that primitive UCB ALDHhi myeloid progenitors play a significant role in support of angiogenesis in vivo.
UCB ALDHhi cells also possessed enriched ECFC capacity in vitro. In contrast to the dominant myeloid progenitor composition, ECFCs were infrequent at approximately one ECFC in 50,000 ALDHhi cells. Although this represents a significant ECFC enrichment from unpurified UCB MNC , it is unlikely that the few ECFCs within 40 × 103 ALDHhi cells make a major contribution toward the maintenance of HUVEC survival and the induction of tubule-forming function that we observed under growth factor and serum-starved conditions in vitro. However, paracrine contributions of the nonhematopoietic component of the ALDHhi population should not be overlooked in vivo. Approximately 7% of the UCB ALDHhi population was CD45−, and these cells demonstrated high coexpression of primitive endothelial precursor markers (CD34 and CD133). Therefore, we propose that UCB ALDHhi cells represent a heterogeneous mixture of primitive hematopoietic and nonhematopoietic progenitors that may act in synergy to generate a supportive microenvironment to promote EC survival and revascularization function.
After intravenous transplantation, UCB ALDHhi cells induced stable recovery from acute ischemic injury within 2 weeks of transplantation into mice with unilateral hind limb ischemia. However, permanent engraftment of ALDHhi cells within the ischemic muscle was below the threshold of detection by flow cytometry. Notably, transplantation of 50–100-fold greater cell doses of unpurified UCB MNC or ALDHlo cells also showed little engraftment by FACS. Using NOD/SCID/MPSVII mice, a unique cell tracking model capable of detecting GUSB-expressing human cell at the single-cell level, we demonstrate that intravenous-injected UCB ALDHhi cells specifically recruited into the ischemic limb at 3 and 7 days after transplantation, and these cells survived in ischemic muscle tissue, albeit at low frequencies, for up to 28 days. Nonetheless, this low frequency engraftment of ALDHhi cells in the ischemic region at early time points was sufficient to augment the revascularization of ischemic muscle indicating a vascular stabilization was mediated by the ALDHhi cells.
Recent models have proposed a transient role for proangiogenic hematopoietic cells in the stabilization of injured vasculature and in the activation and recruitment of vessel resident endothelial precursors to the sprouting vessel branch . In support of this concept, we show that sublethal irradiation prior to transplantation of ALDHhi cells reduced recruitment of human cells to the ischemic region, increased engraftment of ALDHhi cells in the murine BM, and delayed the functional recovery of perfusion. Therefore, we propose that early recruitment of ALDHhi cells to the ischemic region is critical for vascular recovery, allowing ALDHhi cells to release proangiogenic stimuli that activate an endogenous program for collateral vessel formation resulting in improved limb perfusion. Thus, strategies to improve the efficiency ALDHhi progenitor cell delivery, or to prolong survival at or near the site of ischemic injury, are expected to further augment vascular regeneration via paracrine or contact dependent mechanisms.
Strategies that purify proangiogenic cell subtypes from alternate human sources are required to improve the efficacy of human cell therapy trials . Our studies outline several important caveats relevant to the development of cellular therapies to treat ischemic diseases. First, we have identified a mixed population of hematopoietic/myeloid and nonhematopoietic/endothelial ALDHhi progenitor cells from human UCB as a readily available and clinically applicable cell population with potent proangiogenic function. Second, the potential use of allogeneic UCB cells will open new avenues toward clinical cell-based therapies for ischemic disease, as mounting evidence indicates the potential for progenitor deletion and dysfunction of autologous BM cells in patients with severe diabetes and cardiovascular disease [2, 18–20]. Third, we have shown that neither high-level nor permanent engraftment of human cells were necessary to mediate augmentation of perfusion and improved vessel density in vivo. In an allogeneic context, such as intramuscular delivery of UCB ALDHhi cells in patients with ischemic heart disease or critical limb ischemia [49, 50], future clinical transplantation strategies using UCB may require only short-term immune suppression to initiate potent revascularization. Future studies will increase our understanding of the cellular constituents and paracrine pathways that modulate regenerative angiogenesis and will lead to novel therapies to improve tissue vascularization during ischemic diseases.
This work was supported by the Canadian Institutes of Health Research (MOP#86759) and by a grant-in-aid (NA 7295) and the MacDonald scholarship/new investigator salary award from the Heart and Stroke Foundation of Canada to D.A.H. D.M.P. received support from a Canadian Institutes of Health Research studentship and an Ontario Graduate Scholarship. We wish to acknowledge the contributions of Drs. Kristin Chadwick and Karen Morley at the London Regional Flow Cytometry Facility and David Carter at the London Regional Genomics Facility.
DISCLOSURE OF POTENTIAL CONFLICTS OF INTEREST
The authors indicate no potential conflicts of interest.