Disclosure of potential conflicts of interest is found at the end of this article.
Author Contributions: A.E.B. wrote the manuscript.
First published online in STEM CELLSEXPRESS February 4, 2011; available online without subscription through the open access option.
Flow cytometry has been used to detect cells that express high levels of the aldehyde dehydrogenase activity in normal tissues. Such ALDH bright (ALDHbr) cell populations have been sorted from human cord blood, bone marrow, mobilized peripheral blood, skeletal muscle, and breast tissue and from the rodent brain, pancreas, and prostate. A variety of hematopoietic, endothelial, and mutiltipotential mesenchymal progenitors are enriched in the human bone marrow, cord, and peripheral blood ALDHbr populations. Multipotential neural progenitors are enriched in rodent brain tissue, and tissue-specific progenitors in the other tissue types. In xenograft models, uncultured human bone marrow and cord ALDHbr cells home to damaged tissue and protect mice against acute ischemic injury by promoting angiogenesis. Uncultured cord ALDHbr cells also deploy to nonhematopoietic tissues and protect animals in CCl4 intoxication and chronic multiorgan failure models. Mouse ALDHbr cells and cells derived from them in culture protect animals in a chronic neurodegenerative disease model. Purifying ALDHbr cells appears to increase their ability to repair tissues in these animal models. Clinical studies suggest that the number of ALDHbr cells present in hematopoietic grafts or circulating in the blood of cardiovascular disease patients is related to clinical outcomes or disease severity. ALDHbr cells have been used to supplement unrelated cord blood transplant and to treat patients with ischemic heart failure and critical limb ischemia. ALDH activity can play several physiological roles in stem and progenitor cells that may potentiate their utility in cell therapy. STEM Cells 2011;29:570–575
Several allogeneic and autologous cell populations are in clinical development for use as agents to repair tissue damage. This concise review focuses on one of these cell populations that express high activity of the enzyme aldehyde dehydrogenase (ALDH). These cell populations known as ALDH bright (ALDHbr) cells are isolated from adult tissues by flow sorting. Jones et al.  developed the first method to sort living cells based on ALDH activity and demonstrated that human and mouse bone marrow hematopoietic progenitor cells (HPCs) expressed high levels of the enzyme. Storms et al.  synthesized a new substrate for the flow assay and characterized human cord blood (CB) HPC populations with high ALDH activity and low light scatter properties. Since the commercialization of the latter method, now widely known by its trade name Aldefluor (Aldagen, Inc, Durham, NC, www.aldagen.com.), it has been used to enrich populations of stem and progenitor cells from a variety of adult tissues, from primary cancers, and from cultured cells. This review briefly outlines the basis for selection of ALDHbr cells by cell sorting, explores roles that ALDH may play in stem cell physiology and tissue repair, and then details tissue repair activities that have been demonstrated in ALDHbr cells isolated from various tissues. Efforts to translate basic studies into clinical use of ALDHbr cells in regenerative medicine are summarized. The length limitation of a review precludes detailed discussion of other aspects of roles of ALDH in cells or the literature on expression of ALDH by tumor cells. However, several recent reviews of these topics are available [3–6].
HOW ARE ALDHBR CELLS DEFINED, AND WHAT IS THE SIGNIFICANCE OF HIGH ALDH ACTIVITY IN CELLS?
ALDH is a generic designation for a closely related superfamily of 19 human enzymes that can catalyze the pyridine nucleotide linked oxidization of aldehydes into carboxylic acids (Fig. 1). The individual members of the superfamily show substrate preferences for different aldehydes and some also have other functions unrelated to aldehyde oxidation . A variety of compounds that inhibit ALDH activity is available; these compounds may inhibit the activities of several different ALDH gene products . One inhibitor diethylaminobenzaldehyde (DEAB) is routinely used as an inhibitor of the Aldefluor reaction . Figure 1 explains how the Aldefluor reaction is carried out and how ALDHbr cells are defined for cell sorting.
The full range of ALDH isozymes that oxidize Aldefluor under reaction conditions conventionally used to prepare samples for sorting of ALDHbr cells is unknown. However, immunochemical and DEAB inhibition studies suggested that ALDH1A1 is a major ALDH gene product over expressed in human bone marrow stem cells  and breast cancer stem cells and that ALDH1A1 activity is correlated with high Aldefluor activity . Recently, RNA knockdown and antibody staining methods have implicated ALDH1A3, a molecule that is closely related to ALDH1A1, as the major contributor to Aldefluor oxidation in some ALDHbr breast cancer stem cells .
ALDH1A1 and/or ALDH3A1 may play an important role in defining the ALDHbr population is important, because these are two of four ALDH isoforms (ALDH1A1, ALDH1A2, ALDH1A3, and ALDH8A1) that oxidize retinaldehyde to retinoic acids . One hypothesis suggests that the ALDH activity detected by Aldefluor influences stem cell and tissue repair cell activity through retinoic acids [11, 12]. Retinoic acids formed by the action of these four ALDH isoforms could bind to transcription factors that regulate many developmental programs [13, 14]. In principle, ALDH1a1 and ALDH1A3 could indirectly regulate cell activity and proliferation by controlling intracellular retinoid concentrations and could modulate the activity of other cells by releasing retinoids and/or paracrine factors under retinoid control. To date, ALDH1A1 is the only over expressed ALDH isoforms with retinoid oxidizing activity detected in ALDHbr populations sorted from normal tissues [12, 15]. However, targeted deletion of ALDH1A1 activity does not diminish Aldefluor oxidation by mouse bone marrow cells , suggesting that other ALDH isozymes can oxidize Aldefluor oxidization in these animals. Similarly, knocking down transcript levels for each of the 15 known zebra fish ALDH isoforms fail to reduce retinoid responsive expansion of ALDHbr hematopoietic progenitors . The redundancy of ALDH isozymes and complexity of retinoid regulation have made it difficult to determine the full range of ALDH molecules that participate in retinoid mediated activities in cells.
ALDH isozymes that do not have oxidize retinoids, including mitochondrial ALDH isozymes, are also expressed in ALDHbr cells [10, 12, 15]. In all cases, these ALDH isozymes are coexpressed with ALDH1A1 or ALDH1A3 and whether these other isozymes contribute to the oxidation of the Aldefluor substrate is unknown. Several authors have pointed out that activities of ALDH unrelated to retinoid metabolism can be important in stem cell biology [3–6]. ALDH1A subfamily proteins and other ALDH isozymes oxidize aldehydes besides retinoids, including lipid oxidation products formed when cells are subjected to oxidative stress . Several stem cell types, including some cancer stem cells, reside in and have metabolic pathways attuned to hypoxic environments , and the increase in ALDH activity may reflect the demands of surviving in such niches. Evidence consistent with this idea is presented below. Thus, ALDH isozymes over expressed in ALDHbr cell populations can regulate a variety of functions that are important in stem cell biology and tissue repair, but if and how these molecules contribute to tissue repair in the specific experimental and clinical contexts described in the following sections are largely unknown.
ALDHBR CELLS FROM SOURCES OF HEMATOPOIETIC GRAFT IN TRANSPLANTATION AND TISSUE REPAIR
ALDHbr populations derived from human CB [2, 19–24], bone marrow [1, 25–27], and cytokine mobilized peripheral blood [28–30] are highly enriched in early and lineage committed HPC and heterogeneous in expression of canonical HPC surface markers such as CD133 and CD34. All myeloid lineages, B-cells, and natural killer cells arise from ALDHbr CB cells in mouse xenograft models. ALDHbr CB populations include all of the long-term and most of the short-term cells that reconstitute hematopoiesis in these models. Recently, Liu et al.  demonstrated that T-cells also arise from highly purified cord ALDHbrCD34+ in nonobese, diabetic, severe combined immunodeficiency (NOD-SCID) γ−/− mice and that the pace and extent of short-term and long-term engraftment of the bone marrow and release of mature blood cells to the periphery depends on the dose of ALDHbrCD34+ cells administered.
Methods to detect and enrich ALDHbr HPCs are moving into the clinic. Measuring circulating ALDHbr cells has been used to monitor effectiveness of HPC mobilization and to detect differences in HPC subsets mobilized by different agents [29, 30]. Also, the number of ALDHbr cells in hematopoietic grafts is being explored as a metric of graft potency. In retrospective studies, the dose of ALDHbr cells administered to patients is inversely correlated with hematopoietic engraftment time [28, 29, 31]. Preliminary reports suggest that ALDHbr content of thawed cords, particularly if carried out prospectively on segments attached to cryopreserved CB units during unit selection, could be useful as a measure of the unit potency [32, 33]; formal prospective trials are anticipated. Finally, CB ALDHbr cell products have been used to supplement conventional allogeneic CB transplant in children in hopes of ultimately accelerating engraftment. A preliminary report suggests that the procedure is feasible and safe .
Two general points regarding the use of ALDHbr cells in regenerative medicine applications emerge from the studies of HPC activity in human ALDHbr cells. First, high ALDHbr activity discriminates different types of cells bearing CD34 and CD133 surface markers. For example, most of the CD34+ and CD133+ cells in CB do not express high ALDH activity, but as already noted, these ALDHlow populations contain short-term but not long-term reconstituting cells [19–24]. Similarly, ALDHbr cells from human bone marrow contain early multilineage progenitors that make megakaryocytes, but ALDHlow cells do not . ALDHbr and ALDHlow population isolated from other tissues also exhibit important functional differences even though they express similar surface antigens, as discussed below.
Second, ALDHbr cells from hematopoietic graft sources are enriched in stem cells besides HPC. Approximately 10%–15% of the ALDHbr cells in cord [2, 18–24] and mobilized peripheral blood are CD34− [28–30]. These cells, CD34− and CD133− ALDHbr cells, have not been found to have any HPC activity in vitro or in vivo [19, 20, 28]. In bone marrow, approximately 50% of the ALDHbr cells are CD34− [25–27], and these cells are likely to have nonhematopoietic activity. ALDHbr cells from human bone marrow are highly enriched in progenitors that give rise to multipotent mesenchymal cells [25, 26]. Endothelial progenitor cells are also highly enriched in human ALDHbr populations from bone marrow [25, 26] and normal peripheral blood [35, 36]. In the context of tissue repair, human ALDHbr cells, especially cells derived from bone marrow, can provide a mixture of potential progenitor and, as discussed below, paracrine activities from a variety of stem cell types.
ALDHBR CELL FROM HUMAN CB AND BONE MARROW IN REPAIR OF ISCHEMIC INJURY
Two studies in mouse xenograft models demonstrate that freshly isolated, uncultured human ALDHbr cells home rapidly to ischemic tissue damage. Nanoparticle labeled CB ALDHbr cells administered intravenously 1 day after left anterior ascending coronary artery occlusion, preferentially homed to the infarcted anterior surface of the heart and could be detected there for 3 days, but labeled ALDHlow cells were detected only in the spleen . In a hind limb ischemia model , labeled bone marrow ALDHbr cells were administered intravenously 1 day after permanent ligation of the blood supply to the right hind limb, and animals were imaged after transplant. Labeled ALDHbr cells were specifically detected at the ischemic lesion and persisted there for up to 7 days.
In both these models, ALDHbr cells that homed to ischemic sites mediated local formation of new blood vessels. In the coronary model, engraftment with ALDHbr cells, but not ALDHlow cells, led to an increase in the number of large diameter mouse blood vessels specifically in the infarcted area of the heart. No improvement in cardiac function was noted. In the hind limb model, perfusion of the ligated limb increased 7 days after infusion of the ALDHbr cells and increased steadily to >80% of control level over the next 2 weeks. Restoration of perfusion was associated with a twofold increase in capillary density at the ligation site. As in the coronary model, few human cells were detected in the damaged leg 3 weeks after transplantation. Transplantation with ALDHlow cells or with unsorted bone marrow mononuclear cells did not restore tissue perfusion or increase in capillary density. These results show that ALDHbr cell populations contain the angiogenic cell compartment of human bone marrow and CB and that these cells can stimulate formation of new blood vessels at sites of ischemic injury.
Because human cells did not persist in large numbers at injury sites in these ischemic injury models, angiogenesis was attributed to elaboration of angiogenic factors by the transplanted ALDHbr cells. Smith et al.  have reported that human bone marrow ALDHbr cells express angiogenic cytokines, that expression of these cytokines is upregulated by hypoxia, and that medium conditioned by these cells can protect endothelial cells from ischemic damage. Thus, ALDHbr cell populations have a variety of progenitor cell activities and can respond to ischemic tissue damage through paracrine mechanisms.
In the hind limb ischemia study, unsorted bone marrow cells did not increase limb perfusion or capillary density even though the dose administered contained two- to fourfold more ALDHbr cells than the angiogenic ALDHbr preparations. This suggests that ALDHlow cells in human bone marrow can inhibit the homing and/or angiogenic activity of ALDHbr cells and that sorting potentiates angiogenic activity. This provides a rationale for sorting ALDHbr cells from bone marrow for potential therapeutic uses.
ALDHbr cell technology is moving into the clinic in the context of cardiovascular medicine and ischemic diseases. Povsic et al. [35, 36] reported that the number of ALDHbr cells circulating in the blood of cardiovascular patients was inversely correlated with the extent of coronary artery disease. This group suggested that circulating ALDHbr cell levels could be used to assess risk of cardiac disease and to monitor the efficacy of therapeutic interventions. The number of ALDHbr cells in the bone marrow of noncardiac patients undergoing hip replacement remains reasonably constant, but peripheral blood ALDHbr cells varied with age . Thus, unless disease or medical treatment depletes the bone marrow ALDHbr compartment , ALDHbr populations from elderly patients should be accessible for use in autologous cell therapy. Clinical trials testing the feasibility and safety of using uncultured autologous bone marrow-derived ALDHbr cell populations in treating critical limb ischemia  and ischemic heart failure  have been completed; preliminary reports suggest that ALDHbr cells may improve perfusion of ischemic leg or heart tissue.
OTHER TISSUE REPAIR MODELS INVOLVING HUMAN CB ALDHBR CELLS
Cells derived from uncultured CB ALDHbr cells during a 2 to 3-month period after transplantation can home to several nonhematopoietic tissues in response to chronic injury caused by genetic lysosomal storage disease  and by CCl4-induced liver damage  in mouse xenograft models. Such dispersion of human cells to tissues was not observed in animals engrafted with ALDHlow cells. In the intoxication model, prior engraftment with ALDHbr cells reduced mortality to 58% when compared with untreated animals and animals receiving ALDHlow cells. As human cells did not persist in large numbers in the liver, the authors suggested that the cells derived from the ALDHbr cells protected engrafted mice by secreting paracrine factors.
ALDHBR CELLS FROM NEURAL TISSUE
ALDHbr multipotent neuroprogenitors cells are present in the developing rat embryonic neural tube , fetal mouse brain , and both subventricular and subcortical zones of adult mouse brain . These cells form neurospheres that can be propagated and also driven to form neurons, astrocytes, and glia in culture. When adult brain neural progenitor cells (NPC) were sorted into subpopulations based on prominin-1, Lewis X, and high ALDH activity, only ALDHbr cells demonstrated the ability to form neurospheres and retained multipotency in vitro . Thus, ALDH activity identifies functional neuroprogenitor subpopulations within broader populations defined by surface antigens.
ALDHbr cells were cultured under conditions that promoted formation of motor neurons and transplanted i.t. into mouse models of spinal motor atrophy [48, 49]. Both transplanted neurons and uncultured ALDHbr cells migrated to ventral horns and established functional neuromuscular junctions. Transplantation significantly ameliorated all aspects of disease progression and extended life, but did not rescue the animals. Although migration of transplanted cells was observed and motor neurons arose from the transplanted cells, the protective effects of cell transplantation were attributed to production of angiogenic and neuroprotective cytokines by the cells. Cells derived from the ALDHbr cells strongly expressed such cytokines. Uncultured, freshly isolated ALDHbr cells also improved the disease progression and secreted several of protective cytokines, although not as much as the cultured cells. Corti et al.  suggested that isolating ALDHbr cells and culturing them in the absence of the inhibitory effects of spinal cord glial cells promotes the development of cells that can produce neuroprotective cytokines. This recalls the effect of enriching ALDHbr cells on efficacy of human bone marrow in the hind limb ischemia model discussed above.
ALDHBR CELLS FROM OTHER TISSUES
Skeletal Muscle Progenitors
ALDHbr populations derived from biopsies  or primary explants  of human skeletal muscle are enriched in myoblast progenitors. ALDHbr cells from biopsies contained CD34+ and CD34− cells, and none of the cells initially expressed the myoblast marker CD56. ALDHbrCD34+ cells had some characteristics of mesenchymal stem cells (MSCs) and gave rise in culture to a heterogeneous population of CD56− cells. The ALDHbrCD34− fraction developed in vitro as a highly enriched population of CD56+ myoblasts that formed myotubes and showed strong myogenic potential on i.m. transplantation . When explant cultures were driven to differentiate, only ALDHbr cells gave rise to myoblasts. Differentiated ALDHbr myoblasts survived treatment with H2O2 and DEAB, transplantation, and engrafts skeletal muscle better than ALDHlow populations . This result is consistent with the idea that high ALDH activity allows ALDHbr populations to survive oxidative stress encountered in tissue repair contexts. Finally, when immunoselected CD56+ cells from the differentiated cultures were analyzed, no differences were found in the myogenic potential of ALDHbr and ALDHlo cells in vitro . This result is consistent with observations on MSCs [52, 53] and endothelial cells  showing that ALDH expression is sometimes uncoupled from functional activity when differentiated cells are propagated in culture.
ALDHbr cells from human breast reduction samples include cells that formed mammospheres at 20-fold higher frequency than unsorted cells and had 10-fold higher clonogenicity. The cells were multipotential, giving rise to uncommitted, myoepithelial, luminal epithelial and mixed colonies, and ducts, when transplanted into mammary fat pads. The ALDHdim population gave rise only to luminal epithelial cells . In contrast to these results with human cells, ALDHbr cells isolated from bovine mammary epithelium gave rise to cells with luminal phenotype in culture, while ALDHlow cells were myoepithelial progenitors .
Rovira et al.  used Aldefluor to sort the ALDHbr population of central acinar/terminal duct cells from peripheral acinar duct units of adult mice. These cells expressed early embryonic pancreas markers and formed spherical pancreatospheres. These spheres accumulated secretory products and retained some ALDHbr cells, although ALDH activity was extinguished in most cells. The derived endocrine cells showed glucose regulated insulin secretion. Spheres could be serially propagated. On transplantation into mouse embryos, ALDHbr cells, but not dim cells, contributed to both exocrine and endocrine lineages in the developing pancreas. In a repair model of pancreatitis, ALDHbr cells expanded in structures derived from acinar cells.
Mouse prostate cells contain an ALDHbr population that gives rise to functional basal and luminal duct tissue and neuroendocrine cells when transplanted under the renal capsule with feeder cells. Unlike ALDHlow cells, the ALDHbr cells express basal epithelial and characteristic prostate progenitor cell markers and proliferated robustly .
Corneal Limbic Cells
ALDHbr and ALDHlow cell populations derived from cadaveric human limbic tissue differ in clonogenicity and expression of putative limbic progenitor cell markers. However, in this case, limbic progenitor cell activity was enriched in the ALDHlow cells . Several ALDH isoforms are expressed highly in the cornea where ALDH activity protects the cornea from oxidative damage . Thus, high ALDH is a characteristic of differentiated corneal cells.
ALDHbr cell populations enriched in stem cell activity have been sorted from several normal tissues using the Aldefluor method. Identifying the specific ALDH isozymes that oxidize Aldefluor in each ALDHbr population will help clarify the physiological functions of ALDH in these cells. ALDHbr cells reconstitute hematopoiesis and effect tissue repair in animal models. Clinical monitoring of graft ALDHbr cell content and circulating ALDHbr cell levels in patients suggests that ALDHbr cells play important roles in hematopoietic transplantation and in ischemic cardiovascular disease. ALDHbr cells have been manufactured from unrelated donor CB and from autologous bone marrow, and have been administered to patients in early clinical trials. Additional uses for ALDHbr cells in cell therapy may emerge from the preclinical models reviewed above.
Several issues concerning the potential clinical use of regenerative therapies based on ALDHbr cells remain to be addressed. Most significantly, one ALDHbr cell product is manufactured for one patient, a characteristic shared with other therapies based on banked unrelated CB, unmanipulated autologous bone marrow, autologous apheresis products, and autologous induced pluripotential cell products, among others. In principle, manufacture and delivery of these lot-of-one products pose more complex logistic problems than off-the-shelf allogeneic products made by processes that yield doses for many patients form a single production run. Manufacturing a clinical ALDHbr cell product, which takes about half-a-day with current sorting technologies, is at an intermediate level of complexity when compared with unmanipulated banked CB and bone marrow products, which require minimal processing before delivery, and expanded CB and bone marrow products or induced pluripotential cells, which require cultivation. Gene therapy approaches add to complexity. Scaling ALDHbr cell manufacture, like scaling of cell therapies requiring cell culture or manipulation, will be challenging.
Whether cell sorting from an adult tissue sample can yield a therapeutically effective dose of ALDHbr cells remains unknown. In preclinical models and early clinical trials, enriched ALDHbr cells have been administered in numbers one or two orders of magnitude lower than most other cell therapies. This reflects the rarity of these cells, the size of the cord or the bone marrow samples from which the ALDHbr cell products have been manufactured, and efficiency of manufacturing. As noted above, preclinical data suggests that enriching ALDHbr cells increases their potency in tissue repair when compared with unsorted tissue cell preparations. How the potency of unexpanded ALDHbr cells that have not been exposed to cytokines in culture will compare with other therapies based on expanded cells is unknown. Another technical issue that will influence the efficacy of ALDHbr and all other cell products is the route of delivery. Interactions among dose, potency, and delivery will vary for different clinical indications. These questions will only be answered as ALDHbr cell products and other candidate cell therapies move from initial safety trials into more advanced clinical trials focused on efficacy in specific clinical indications.
Disclosure of Potential Conflicts of Interest
The author holds equity in and is a consultant to Aldagen, Inc., a company that is developing research, diagnostic, and therapeutic products based on detecting and isolating ALDHbr cells.