Turning bone marrow into brain

Most of the studies addressing the contribution of bone marrow-derived cells to the adult brain were performed in bone marrow chimeras, which were generated by the transplantation of genetically marked bone marrow cells into conditioned hosts (Fig. 2). Donor-derived cells were subsequently detected in the host CNS based on the expression of a transgene, retroviral tag, or the Y chromosome (when male bone marrow was transplanted into female recipients).

To identify bone marrow cells that have turned into brain cells, tissue sections were analyzed for cells showing colocalization of the bone marrow-label with tissue-specific markers and a distinctive morphology indicative of cell fate change. Unless the boundaries determined by embryologic trilaminar origin are not maintained in the adult, cells of mesodermal origin, such as microglia, brain macrophages, and endothelial cells, are more likely to be generated from the transplanted bone marrow than cells of ectodermal origin, including neurons, astrocytes, and oligodendrocytes. However, the signals regulating division and differentiation in the adult brain may be quite distinct from those in the immature brain, and the acquisition of specific cell phenotypes may therefore largely depend on local inductive signals.

Harvesting of bone marrow Harvesting of bone marrow Harvesting of bone marrow Transplantation into Transplantation Into Transduction with GFP

irradiated mouse irradiated mouse .—

Transplantation into irradiated mouse

Transplantation into irradiated mouse

Fig. 2. Schematic view of the generation of bone marrow (BM) chimeras. Bone marrow cells are harvested from donor animals and injected into myeloblated (e.g., lethally irradiated) recipient animals. Reconstitution of hematopoiesis with donor-derived peripheral blood cell progeny occurs within several weeks after transplantation. Donor-derived cells can be identified by the presence of the Y chromosome in female animals transplanted with male bone marrow cells. When donor bone marrow is derived from a transgenic animal or is genetically marked by viral trans-duction, donor-derived cells can be detected in the recipient animals based on the expression of the transgene/viral-encoded protein (e.g., P-galactosidase or GFP).

Fig. 2. Schematic view of the generation of bone marrow (BM) chimeras. Bone marrow cells are harvested from donor animals and injected into myeloblated (e.g., lethally irradiated) recipient animals. Reconstitution of hematopoiesis with donor-derived peripheral blood cell progeny occurs within several weeks after transplantation. Donor-derived cells can be identified by the presence of the Y chromosome in female animals transplanted with male bone marrow cells. When donor bone marrow is derived from a transgenic animal or is genetically marked by viral trans-duction, donor-derived cells can be detected in the recipient animals based on the expression of the transgene/viral-encoded protein (e.g., P-galactosidase or GFP).

3.1. Microglia and Perivascular Cells

Microglia are resident immunological effector cells that sense pathological events in the CNS (23). Their origin has been one of the most contentious issues in glial research over the past decades, but it is now generally accepted that microglia are ontogenetically related to cells of the mono-nuclear phagocyte lineage. Resting microglia show a ramified morphology with downregulated immunophenotype; other brain macrophages, such as perivascular cells and leptomeningeal macrophages, express major histocompatibility complex (MHC) class II antigen and high levels of CD45 (24,25). Microglial activation occurs in response to even subtle changes in the brain microenvironment, and activated microglia express macrophage-

related antigens, such as the complement receptor 3, MHC class I and class II antigens, and a number of cytokines and cell adhesion molecules (26).

In the developing brain, pial mesenchymal progenitor cells from the yolk sac seed in the brain parenchyma before its vascularization; later, circulating monocytes contribute to the population of ameboid microglia that colonize the embryonic brain (27). After extensive proliferation, these ameboid microglia ultimately transform into the ramified microglia that can be found in the postnatal brain (28,29). Some authors have challenged this view and suggested that microglial cells derive from neuroectodermal glioblasts (30,31). Moreover, studies using irradiated adult rats transplanted with MHC-mismatched bone marrow cells strongly suggested that only perivas-cular cells and leptomeningeal macrophages, but not parenchymal micro-glia, were bone marrow derived (32). Even after severe inflammatory conditions of the brain, resident microglia represented a very stable cell pool (33).

In contrast, mice transplanted with bone marrow cells expressing the green fluorescent protein (GFP) showed substantial microglial engraftment up to 12 mo after bone marrow transplantation (BMT) (7,34,35). In these chimeras, host perivascular cells were also completely substituted by donor-derived macrophages within 4 mo after BMT (36). Although the appearance of GFP-expressing cells in perivascular and leptomeningeal sites occurred throughout the brain, preferential microglial engraftment was observed in the olfactory bulb and later in the cerebellum.

From the data of Brazelton et al. (6), an estimated 5% of the microglia in the olfactory bulb were generated from bone marrow cells 8-12 wk after transplantation. By 15 wk post-BMT, up to a quarter of the cerebellar microglial population was found to be donor derived (34). De Groot et al. (37) reported that approx 10% of the white matter microglia arose from the transplanted bone marrow, and studies using transplantation of murine bone marrow cells transgenic for P-galactosidase (38) suggested that 20-30% of all brain macrophages originated from the donor marrow at 4-12 mo after transplantation. In mice homozygous for a mutation in the PU.1 gene and transplanted with wild-type bone marrow cells at birth without irradiation, all microglia and macrophages throughout the brain arose from the donor bone marrow (5).

Thus, it can be concluded that microglia in the postnatal murine brain may be generated in the bone marrow compartment. Because mature mono-cytes traffic between the blood and the brain despite an intact blood-brain barrier (39,40), these cells are likely to replace the perivascular and leptom-eningeal macrophage cell pools continuously and to differentiate into resi-

Fig. 3. Central nervous system engraftment of bone marrow-derived cells. Chimeric mice were generated by the transplantation of GFP-marked bone marrow cells into lethally irradiated wild-type mice (7,34). (A) Four weeks after transplantation, the middle cerebral artery was occluded for 60 min, and after 14 d of reperfusion, the ischemic hemisphere of a chimera was infiltrated by donor-derived GFP-expressing microglia/macrophages. (B) Twelve months after transplantation, a rare GFP-expressing Purkinje neuron is seen in the cerebellum of a bone marrow chimera.

Fig. 3. Central nervous system engraftment of bone marrow-derived cells. Chimeric mice were generated by the transplantation of GFP-marked bone marrow cells into lethally irradiated wild-type mice (7,34). (A) Four weeks after transplantation, the middle cerebral artery was occluded for 60 min, and after 14 d of reperfusion, the ischemic hemisphere of a chimera was infiltrated by donor-derived GFP-expressing microglia/macrophages. (B) Twelve months after transplantation, a rare GFP-expressing Purkinje neuron is seen in the cerebellum of a bone marrow chimera.

dent microglia. In vitro evidence suggests that microglial differentiation of peripheral blood monocytes may result from an interaction with astrocytes (41).

Interestingly, microglial engraftment is significantly enhanced following CNS damage. Thus, focal cerebral ischemia leads to a dramatic recruitment of bone marrow-derived cells into the ischemic brain, and almost one-third of these cells develop into microglia (Fig. 3) (34). Even remote lesion of the CNS by facial nerve axotomy in rodents induced microglial engraftment in proximity to the injured motoneurons (34,42). In a murine model of globoid cell leukodystrophy, bone marrow-derived microglia were distributed diffusely in both gray and white matter after 100 postnatal days (43). Donor-derived cells of the macrophage lineage also infiltrated the CNS in a regionally specific manner following BMT in mouse models of GM2 gangliosidosis (44) and Gaucher disease (45). Finally, peripheral macrophages were specifically recruited in cuprizone-induced CNS demyelination (46). It has thus been suggested that microglia may be used therapeutically to deliver genes of interest to the diseased brain (34).

3.2. Astrocytes

Astrocytes are the most numerous glial cells of the CNS. They seem to be generated from the dorsal regions of the neural tube during development (47) and throughout life are continuously replenished from multipotent neuroepithelial stem cells and glial-restricted precursors (48). The glial fibrillary acidic protein (GFAP) represents the major component of intermediate filaments in mature astroglia, but astrocytes may also express S100P (49). Extending processes to blood vessels, astrocytes participate in the formation of the blood-brain barrier, and they contribute to tissue homeostasis (50).

Astrocytes have been shown to control neuronal life directly by regulating synaptogenesis and neurogenesis (51,52). Moreover, in the adult brain, GFAP-expressing astrocytes have been suggested to represent neural stem cells capable of generating macroglia and neurons (53-55). It is therefore surprising that Eglitis and Mezey (3) found that 0.5-2% of the bone marrow-derived cells engrafting in the murine CNS after BMT expressed GFAP. Marrow stromal cells injected into the lateral ventricle of neonatal mice also differentiated into GFAP-immunoreactive astrocytes within 12 d (56).

On the other hand, none of the chimeric mice transplanted with bone marrow cells expressing GFP revealed any astroglial differentiation of the donor cells (6,10,34,35,57). Only when GFP-marked bone marrow cells were injected directly into the brain did they develop into GFAP-expressing astrocytes (35).

In vitro, bone marrow stromal cells were induced to differentiate into astrocytes in the presence of growth factors or differentiation factors such as Noggin (58,59). Multipotent adult progenitor cells (MAPCs) generated from murine MSCs differentiated into GFAP-expressing astrocytes in vitro and gave rise to astrocytes throughout the brain after injection into an early blas-tocyst (60). However, these cells failed to generate astroglia after transplantation into adult NOD/SCID mice. In the studies of Woodbury et al. (61) and Deng et al. (62), human and rat bone marrow stromal cells developed into neuronal phenotypes in vitro, but did not differentiate into GFAP- or S100P-expressing astroglia.

Because astrocyte proliferation is dramatically enhanced after CNS injury (63), the contribution of bone marrow-derived cells to the population of reactive astroglia was studied after cerebral ischemia. In chimeric rats subjected to middle cerebral artery occlusion (MCAO), Eglitis et al. (64) observed that the number of bone marrow-derived astrocytes was twice as high in the ischemic hemisphere compared to the contralateral side and almost 10 times as high as in control rats within 48 h. MSCs grafted into the ischemic brains of rodents (65,66) or injected into the peripheral circulation after transient MCAO in rats (67) were reported to differentiate into GFAP-expressing astrocytes in high numbers. In contrast, Hess et al. (57) failed to detect donor-derived astrocytes in chimeric mice subjected to MCAO after transplantation of GFP-marked bone marrow cells. Moreover, MSCs transplanted into the rat spinal cord after contusion did not express GFAP up to 5 wk after lesion (68). In contrast, MSCs administered intravenously to rats 1 d after traumatic brain injury gave rise to GFAP-expressing cells in the brain (69). Thus, differentiation of bone marrow cells into astroglia remains controversial, and the switch of cell fate may depend on specific local signals.

3.3. Oligodendrocytes

In contrast to astrocytes, oligodendrocytes seem to originate more from the embryonic ventral neural tube (70). Precursors of oligodendrocytes from the ventricular zone migrate to extraventricular sites during CNS development and continue to divide throughout life (71). Maturation of oligodendrocytes includes the expression of specific markers, such as O4 and galactocerebroside, and the extension of endfeet toward axons, followed by the process of myelination (72). In the adult, myelination is thought to occur by the recruitment of quiescent oligodendrocyte precursor cells (73). Recently, adult bone marrow cells enriched in c-Kit-positive hematopoietic progenitor cells were found to differentiate into O4-immunoreactive oligodendrocytes within 6 d after intracerebral transplantation into the neonatal mouse brain (74). Similarly, GFP-expressing bone marrow cells grafted directly into the adult murine brain differentiated into oligodendroglia-expressing carbonic anhydrase II within 12 wk after injection (35).

In contrast, murine MAPCs transplanted into adult NOD/SCID mice did not turn into galactocerebroside-expressing oligodendroglia, whereas MSCs could be induced to adopt an oligodendroglial fate in vitro (59,60). Similarly, GFP-marked bone marrow cells transplanted into myeloablated adult mice failed to generate oligodendrocytes for up to 6 mo after BMT (35).

Experimental evidence suggested that GFP-expressing murine MSCs can form functional myelin on injection into a focal demyelinated lesion in the rat spinal cord (75). Within 3 wk after injection, bone marrow-derived cells were found to express the myelin proteins myelin basic protein and P0. Moreover, rat MSCs predifferentiated in vitro into a Schwann cell-like phe-notype were found to myelinate the regenerating fibers of the axotomized sciatic nerve 3 wk after transplantation (76). As for astrocytes, the differentiation of bone marrow-derived cells into oligodendroglia therefore seems to depend largely on specific local cues.

3.4. Endothelial Cells

In the embryo, endothelial cells arise either from endothelial progenitors (angioblasts) or from stem cells, giving rise to both endothelial and hematopoietic cells (hemangioblasts) (77). Even after birth, circulating EPCs generated in the adult bone marrow can be assembled into endothelial channels after in situ differentiation (13). This vasculogenesis by EPCs has been reported in connection with wound healing and tumor vascularization and in response to ischemia (78).

Although the cerebral vascular system is primarily developed by angio-genesis (sprouting), experiments in chimeric mice have recently revealed that bone marrow-derived cells contribute substantially to neovascularization after focal cerebral ischemia. Thus, mice transplanted with bone marrow cells expressing P-galactosidase under the control of the endothelial Tie2 promoter showed bone marrow-derived endothelia in vessels at the border of the infarct, but not in intact parenchymal cerebral vessels (79). In mice transplanted with GFP-expressing bone marrow cells, 42% of the endothelial cells in the infarct were coexpressing GFP with endothelial markers, such as von Willebrand factor, CD31, and isolectin B4, at 3 d after transient MCAO. This number decreased to 26% by 14 d after MCAO (57). Finally, mobilization of EPCs by statin treatment enhanced endothelial regeneration after carotid artery lesion (80).

3.5. Neurons

Undoubtedly, the most intriguing example of adult stem cell plasticity is the conversion of bone marrow-derived cells into neuronal phenotypes. During development, reinforced by signals from the mesoderm, ectoderm from the dorsal side of the embryo forms neural tissue (81). Anterior-posterior neural patterning occurs soon after neural induction, and signals from the underlying mesoderm and the epidermis influence dorsal-ventral patterning.

Neurogenesis occurs in defined regions of the patterned neural plate, and the developmental fate of a neural crest cell depends critically on the signals it receives from the environment through which it migrates. Neurogenesis persists throughout the life of the organism, and small populations of hip-pocampal, cortical, and olfactory bulb neurons continue to be born in the adult dentate gyrus and the subventricular zone (53,82,83). Thus, the signals required for the neuronal differentiation of stem cells are maintained at least in some parts of the adult brain.

Interestingly, in mice transplanted with GFP-transgenic bone marrow cells, up to 0.3% of all neurons in the olfactory bulb were found to express

GFP within 4 mo after BMT (6). Bone marrow-derived cells in the brain were characterized as neurons based on their expression of neuronal antigens, such as neuronal nuclei (NeuN), neurofilament, and neural nuclei class III P-tubulin. Although most of these cells did not display the morphological charac-teristics of neurons, the presence of phosphorylated cyclic adenosine *5'-monophosphate response element-binding protein suggested that the donor-derived cells responded to cues in their environment in a manner consistent with the surrounding neurons.

Similarly, rare immature bone marrow-derived cells in the mouse spinal cord and in sensory ganglia expressed NeuN, neurofilament, and class III P-tubulin at 3 mo after BMT (84). When female mice homozygous for a mutation in the PU.1 gene were rescued by postnatal intraperitoneal injection of adult male bone marrow cells, 0.3-2.3% of all NeuN-immunoreac-tive cells throughout the brain were found to be Y chromosome-positive after 1-4 mo (5). There was no overall increase in the density of donor-derived cells in the neurogenic regions.

In contrast, a subsequent study failed to detect any donor-derived neurons in the brain within 4 mo after transplantation of bone marrow cells expressing P-galactosidase into adult wild-type mice (9).

In the experiments of Nakano et al. (35), none of the GFP-marked bone marrow cells transplanted by systemic infusion gave rise to cells expressing neuron-specific enolase in the brain after 6 mo. However, GFP-marked bone marrow cells injected directly into the brain were found to differentiate into neuronal phenotypes within 4 mo. Marrow stromal cells injected into the lateral ventricle of neonatal mice differentiated into neurofilament-immu-noreactive cells in the brain stem (56).

MAPCs generated from murine MSCs gave rise to NeuN-expressing cells throughout the brain after injection into an early blastocyst (60). However, these cells failed to generate neuronal phenotypes 1-6 mo after transplantation into adult NOD/SCID mice.

Perhaps the most consistent finding is the appearance of rare GFP-expressing Purkinje cells in the cerebellum several months after transplantation of GFP-marked bone marrow cells (Fig. 3) (7,10,85). Based on morphologic criteria and the expression of the y-aminobutyric acid (GABA)-synthesizing enzyme glutamic acid decarboxylase, these newly generated neurons were considered fully developed and functionally integrated into the cerebellar cytoarchitecture (7). Moreover, analysis of the brains of mice transplanted with a single GFP-marked HSC revealed only one GFP-positive nonhematopoietic cell, a Purkinje cell (10).

It is thus conceivable that less than 0.1% of the Purkinje cells in the adult may arise from hematopoietic stem cells by way of transdifferentiation. Nevertheless, it has to be taken into account that intermediate stages of development have not yet been described for the bone marrow-derived Purkinje cells, and issues of potential cell fusion have to be addressed in the light of in vitro findings on stem cell fusion (86). However, in the study of Corti et al. (84), all the cells coexpressing GFP with neuronal markers were mononucleate and diploid.

There is also ample in vitro evidence to suggest that clonal GFP-express-ing MSCs can be induced to differentiate into mature neurons expressing markers of dopamine synthesis, serotonin and GABA (60). Clones of rat and murine MSCs produced by limiting dilution were found to adopt a neuronal phenotype when exposed to combinations of specific growth factors or differentiation factors (59,61). Human MSCs could also be induced to differentiate into cells expressing neuronal antigens in vitro (58,62).

Several recent studies suggested that the injured CNS provides cues for the engraftment and neural differentiation of bone marrow-derived cells. Thus, MSCs grafted into the ischemic brains of rodents (65,66) or injected into the peripheral circulation after transient MCAO in rats (67,87) were found to differentiate into NeuN-, neurofilament-, class III P-tubulin-, or microtu-bule-associated protein 2 (MAP2)-expressing cells within 2 wk. In chimeric mice transplanted with GFP-expressing bone marrow cells, scattered donor-derived cells were immunoreactive for NeuN in the ischemic striatum 1-7 d after transient MCAO (57). Almost 1% of the murine MSCs grafted into the striatum of mice treated with 1-methyl-4-phenyl-1,2,3,6-tetrahydropyridine (MPTP) acquired a dopaminergic (tyrosine hydroxylase immunoreactive) phenotype within 1 mo after transplantation (88). Finally, MSCs injected into the rat spinal cord after contusion displayed NeuN immunoreactivity after 5 wk, but failed to express neurofilament or MAP2 (68).

In chimeric mice transplanted with bone marrow cells expressing P-galactosidase, no donor-derived neural cells were observed in the brain up to 5 mo after stab injury (9). In contrast, MSCs administered intravenously after traumatic brain injury in rats gave rise to NeuN-expressing cells in the damaged CNS (69).

Was this article helpful?

0 0

Post a comment