Multilineage stem cells in the adult

Introduction

A conservative theory would describe totipotent stem cells as the foundation of the embryo that yield increasingly committed progenitor cells at the origin of the diverse fetal organs. Accordingly, postnatal and adult tissues would be renewed and regenerated exclusively by tissue-specific stem cells. Such specialized stem cells have been, in some instances, precisely characterized (the paradigm being the hematopoietic stem cell), but it has long also been known that less committed multilineage stem cells persist in the adult. The mesenchymal stem cell (MSC) is the archetype of post-natal/adult cells endowed with multiple developmental potentials.¹ MSC can, indeed, give rise, experimentally, to bone, cartilage, fat and smooth and skeletal muscle, but the natural participation of MSC in the turnover of these tissues has been unknown. This is because MSC have been solely identified indirectly once they had proliferated in vitro. To what extent the outgrowth of MSC in cultures of multiple tissues reflects the native existence of multipotent mesodermal stem cells has been debated; yet, the ignorance of MSC true identity has not compromised the rise of this cell as one of the preferred candidates for many stem cell therapy treatments.^2,3

Recently though, candidate cells at the origin of mesenchymal stem cells have been described, identified and purified. These putative ancestors of MSC are components of blood vessel walls and, as such, disseminated throughout the organism. We briefly review hereafter the steps that have led us to propose a perivascular origin for mesenchymal stem cells and propose some speculations regarding the possible medical utilization of prospectively purified MSC originators as well as their natural function in the organism.

Mesenchymal Stem Cells Share Traits with Vascular Pericytes

Similarities have been described, in the past decade, to exist between MSC and pericytes in terms of phenotype and gene expression,^4–9 suggesting that MSC indeed represent a progeny of the perivascular cell compartment.¹⁰ This was relatively unexpected inasmuch as pericytes (a.k.a. mural cells or Rouget cells), which encircle endothelial cells in capillaries and microvessels, play a well documented role in angiogenesis and blood pressure regulation.^11–13 Of note though, a few pioneering studies had already contributed to identifying mesodermal developmental potentials in pericytes, assumed to be notably expressed in pathologic conditions such as vascular calcification.^14,15 Moreover, MSC could be derived from isolated blood vessels, reinforcing the possibility that these stem cells are of vascular affiliation.¹⁶

Prospective Identification of Vascular Pericytes as Mesodermal Stem Cell Ancestors

We have marked pericytes in multiple human tissues (skeletal muscle, myocardium, pancreas, adipose tissue, skin, placenta, umbilical cord, bone marrow, kidney, lung, brain, dental pulp) on surface expression of NG2, CD146 and PDGFR-β and absence of known hematopoietic, endothelial and myogenic cell markers.¹⁷ Human pericytes also express alkaline phosphatase in skeletal muscle¹⁸ and the human dermis-1 (HD-1^bri) antigen in the skin.¹⁹ Human pericytes purified to homogeneity from diverse organs by flow cytometry, as CD146⁺ CD34⁻ CD45⁻ CD56⁻ cells, were myogenic in culture and in vivo when injected into cardiotoxin-injured or genetically dystrophic SCID mouse muscles. Seeded in culture, purified pericytes were indistinguishable from conventionally derived mesenchymal stem cells in terms of adherence, morphology, mitotic activity, surface antigen expression (CD44, CD73, CD90, CD105) and, most importantly, developmental potential. Indeed, cultured pericytes from diverse human tissue origins differentiate, at the clonal level, into bone, cartilage and fat cells when cultured under respective relevant inductive conditions and are vigorously myogenic and osteogenic in vivo in immunodeficient mice^17,20 (also Zhang et al., submitted for publication). Further supporting the hypothesis of a perivascular origin of mesenchymal stem cells, we and others also observed that pericytes in their tissue of origin natively express the canonical MSC markers CD44, CD90, CD73 and CD105.^5,17

Vascular Pericytes For Cell Therapies and Tissue Engineering

The above described experiments have suggested that the elusive mesenchymal stem cell originates, at least in part, in vascular pericytes. Can, therefore, pericytes be contemplated as alternative therapeutic cells, and what should be the advantage of using such purified cells in place of crude MSC for regenerative medicine? From a practical point of view, some of us have shown that viable, developmentally potent pericytes can be recovered from cryopreserved human tissue and thus can be widely accessible as therapeutic cells.²¹ More fundamentally, perivascular cells exposed to the adverse hypoxic conditions encountered in the context of wound healing exhibit enhanced proliferation and migration,^22,23 and we have demonstrated that purified pericytes exposed to chemoattractive peptides migrate significantly more efficiently when cultured under hypoxia, as compared to normoxic conditions, in a chemotactism chamber.²⁴ We also confirmed in vivo that human pericytes injected into the skeletal muscles of genetically dystrophic mice migrate at a remote distance from the site of injection and therefore mediate more efficient myogenesis.²⁰

Purified human pericytes have been used experimentally in other, different models of tissue engineering or regeneration. Cylindrical biodegradable PEUU [poly(ester-urethane)urea] scaffolds seeded with skeletal muscle derived pericytes and implanted into the rat aorta developed into functional vascular grafts displaying normally organized intimal, medial and adventitial layers.²⁵ In a different experimental setting, human muscle pericytes injected into the infarcted SCID mouse myocardium restored cardiac anatomy, stimulated host neoangiogenesis and reduced local inflammation. As a result, cardiac function was significantly improved (Chen CW et al., submitted for publication). In a model of oxygen-induced arrested alveolar growth in newborn rats, human umbilical cord-derived pericytes injected intravenously restored the anatomy of the distal lung and normal respiratory function resumed (Pierro et al., submitted for publication). Of important note, in the two latter settings, no significant cell chimerism was detected in the regenerated myocardium and lung, indicating that besides their documented multilineage differentiation, potential pericytes can also contribute indirectly to tissue regeneration, possibly via promotion of angiogenesis, reduction of inflammation and stimulation of endogenous progenitor cells. Accordingly, multiple relevant growth factors and cytokines have been detected in the supernatant of cultured human pericytes,²⁶ further suggesting, indirectly, that mesenchymal stem cells can act as trophic mediators (reviewed in ref. 27).

Pericytes Are Not The Only Perivascular Cells Exhibiting Pre-MSC Characteristics

The above described identification of pericytes as mesenchymal stem cell originators raised the question as to whether MSC are all derived from pericytes.²⁸ Along a systematic search for alternative mesenchymal stem-like cells, we have dissected by flow cytometry the stromal cell compartments of human white adipose tissue, lung and skeletal muscle and identified a population of CD146⁻ CD34^hi CD31⁻ CD45⁻ cells, which, although phenotypically totally distinct from pericytes, exhibit the same developmental potential and ability to give rise to MSC in long term culture. Anatomically, these cells reside in the tunica adventitia of arteries and veins, as assessed in all organs analyzed. We have, therefore, demonstrated the coexistence of two separate perivascular presumptive mesenchymal stem cells: the pericytes lining microvessels and adventitial cells around larger vessels (Corselli et al., submitted for publication). Besides these two classes of perivascular cells, no MSC activity could be detected in any other cell compartment.

Do Perivascular Multilineage Mesodermal Stem Cells Play a Natural Role in Organogenesis and Tissue Repair and Regeneration?

Inasmuch as perivascular stem cells have been functionally identified, as summarized above, following dissociation, purification and introduction into artificial experimental models, the possible natural role played by these cells in tissue development and repair remains mostly speculative and yet, represents one of the most fascinating questions raised by this research. A few reports have described a physiological role of pericytes in tissue regeneration. Davidoff et al. used a model of Leydig cell chemical ablation in the rat to demonstrate that these testosterone-producing cells are regenerated by pericytes.²⁹ Tracking PPARγ-expressing cells in cre-lox transgenic mice revealed that vascular mural cells, i.e., pericytes, are at the origin of white adipocytes in the adult.³⁰ Finally, cell tracking experiments have also documented the pericyte origin of myofibroblasts responsible for kidney fibrosis.^31,32 Of note, our preliminary results also suggest that specialized pericytes constitute a stock of renin-producing cells in the juxta-glomerular region of the human kidney (Stefanska et al., unpublished observations).

Concluding Remarks

For more than thirty years, investigators have routinely derived mesenchymal stem cells indirectly, the only selection criterium being the ability of these cells to adhere and proliferate in long term cultures of bone marrow and other organs. Only in recent years was the native identity of these cells progressively uncovered, first by phenotypic correlations and eventually through prospective purification. It is now becoming increasingly accepted that the ancestor of the cultured MSC is involved in the architecture of blood vessels. This recognition is acknowledged by the recent publication of several review articles supporting this theory.^33–36 Physical association of the initial mesenchymal stem cell with blood vessel walls would explain why MSC have been extracted in cultures of numerous different organs and tissues. A fundamental question that remains unanswered, however, is whether vessel-associated MSC potential is expressed in normal development and in the maintenance of adult tissue homeostasis. Clearly, this potential is present but repressed in many organs: osteo-, chondro-, adipo- and myogenic pericytes have been extracted from sites where these cell lineages never develop, such as the placenta and pancreas.^17,20 Nonetheless, a few reports have convincingly documented the involvement of mural cells in the regeneration of specific cell types. Futher insight into this question will rely on cell lineage-tracing experiments in transgenic mice, several of which are under current development. Such models would also help further decipher the ontogeny of other vessel-associated multilineage mesodermal progenitors, such as myoendothelial cells³⁷ and mesoangioblasts.³⁸

Mesenchymal stem cells, regardless of their identity and anatomic origin, have long been contemplated as privileged therapeutic cells, be it for their multipotency or immunomodulatory ability (reviewed in ref. 39). Presumably, advantages of transplanting purified perivascular cells into patients over conventionally derived MSC include: (1) no need for in vitro expansion culture, thus limiting risks of infection, genetic instability and malignant transformation as well as avoiding exposure to animal products; (2) precise characterization in terms of native tissue localization, phenotype and developmental potential (conversely, MSC are derived from primary, heterogeneous cell cultures) and, consequently, (3) high purity; in this respect, there is evidence that endothelial cells present in conventional MSC cultures negatively regulate the differentiation potential of the latter.⁴⁰ Finally, from a regulatory agency perspective, it is most likely that the use of highly purified and characterized cells will facilitate demonstration of consistency in product identity, purity and potency.