We investigated whether fetal cells can enter the maternal brain during pregnancy. Female wild-type C57BL/6 mice were crossed with transgenic Green Mice ubiquitously expressing enhanced green fluorescent protein (EGFP). Green Mouse fetal cells were found in the maternal brain. Quantitative real-time polymerase chain reaction (PCR) of genomic DNA for the EGFP gene showed that more fetal cells were present in the maternal brain 4 weeks postpartum than on the day of parturition. After an excitotoxic lesion to the brain, more fetal cells were detected in the injured region. The presence of fetal cells in the maternal brain was also confirmed by quantitative real-time PCR for the sex-determining region of the Y chromosome. Four weeks postpartum, EGFP-positive Green Mouse fetal cells in the maternal brain were found to adopt locations, morphologies, and expression of immunocytochemical markers indicative of perivascular macrophage-, neuron-, astrocyte-, and oligodendrocyte-like cell types. Expression of morphological and immunocytochemical characteristics of neuron- and astrocyte-like cell types was confirmed on identification of fetal cells in maternal brain by Y chromosome fluorescence in situ hybridization. Although further studies are required to determine whether such engraftment of the maternal brain has any physiological or pathophysiological functional significance, fetomaternal microchimerism provides a novel model for the experimental investigation of the properties of fetal progenitor or stem cells in the brain without prior in vitro manipulation. Characterization of the properties of these cells that allow them to cross both the placental and blood–brain barriers and to target injured brain may improve selection procedures for isolation of progenitor or stem cells for brain repair by intravenous infusion.
Fetal cells can enter maternal blood circulation during pregnancy , persist in maternal circulation after pregnancy , and engraft many maternal tissues, including bone marrow (BM), spleen, and liver in both mice [3, 4] and humans [5–7]. In humans, such cells have been found to persist in maternal blood as long as 27 years postpartum . However, it was not known whether fetal cells capable of crossing the placental barrier to enter maternal blood during pregnancy could also cross the blood–brain barrier to enter the maternal brain. Under certain conditions, umbilical cord blood cells have been reported to be capable of expressing some proteins characteristic of neural cell types [8–10] and, when injected intravenously into rats with traumatic brain injury or stroke, of entering the brain and expressing immunocytochemical markers for neural cell types [11, 12]. This evidence suggests the hypothesis that fetal cells may enter the maternal brain during pregnancy and differentiate into neural cells. In this study we report fetal microchimerism in the maternal mouse brain together with morphological and immunocytochemical evidence that these fetal cells can express characteristics of perivascular macrophage-like and neural-like cell types in the maternal brain.
Materials and Methods
Eight- to 12-week-old male homozygous C57BL/6CrSlcTgN(act-EGFP) OsbC14-Y01-FM131 mice (Green Mice) were crossed with wild-type young adult female C57BL/6 mice (8–10 weeks old). All of the offspring were hemizygous Green Mice. Control wild-type young adult female C57BL/6 mice remained virgins. Green Mouse pups (7 days old) provided positive controls for hemizygous Green Mouse cells. Ex-breeder wild-type female C57BL/6 mice were purchased from the Laboratory Animals Center, National University of Singapore. These mice had been held as breeding stock since the age of 6–8 weeks and were retired at the age of 6–7 months after weaning their last litter. They were used at the age of 8–9 months old, at least 2–3 months after delivering their last litter. All experiments were conducted under the institutional guidelines of the Animal Ethics Committee of the Singapore General Hospital and the Institutional Animal Care and Use Committee, Office of Life Sciences, National University of Singapore, in accordance with the Guide for the Care and Use of Laboratory Animals, National Institutes of Health, and following the International Guiding Principles for Animal Research .
Excitotoxic Lesions of the Brain
Diffuse lesions of the cortex and hippocampus of wild-type female C57BL/6 mice (7–8 weeks old) were produced by stereotaxic microinjection of 1.2 μl of 1 mg/ml N-methyl-D-aspartic acid (NMDA) in 0.9% saline into the lateral ventricles bilaterally. Two weeks after lesioning, the mice were crossed with male Green Mice (8–12 weeks old).
Detection of Green Mouse Cells in Maternal Blood by Fluorescence-Activated Cell Sorting
Maternal blood samples (500 μl) were taken from the lateral tail vein of non-lesioned mothers (n = 10) at 7 days after delivery. The blood specimens were suspended in lysis buffer (0.83 g NH4Cl, 0.10 g KHCO3, 3.7 mg NaEDTA in 100 ml ddH2O) for a few seconds. Fractions containing nucleated cells were obtained by centrifugation in FicollPaque Plus (Amersham Biosciences, Piscataway, NJ, http://www.amersham.com) degradient solution and fixed with 2% paraformaldehyde in phosphate buffered saline. The cells were then screened by fluorescence-activated cell sorting (FACS) in a flow cytometer (FACScan; Becton, Dickinson and Company, Franklin Lakes, NJ, http://www.bd.com) using an Argon ion laser for excitation at 488 nm and detecting emission of enhanced green fluorescent protein (EGFP)–positive cells from 500–545 nm. FACS signals were corrected to the blood of wild-type virgin female C57BL/6 mice and Green Mouse pups as negative and positive controls, respectively. Data are expressed as mean ± standard error of mean (SEM).
Visualization of Fetal Green Mouse Cells in Maternal Blood and Brain
Samples of the nucleated cell fractions were also visualized directly by phase-contrast and epifluorescence microscopy. Four weeks after delivery, the mothers were euthanized by an overdose of anesthetic and perfused with 0.9% saline followed by 4% paraformaldehyde in phosphate buffer (pH 7.4). The brains were serially sectioned (20 μm) on a cryostat and viewed under epifluorescence microscopy.
Quantitative Real-Time Polymerase Chain Reaction
For quantitative real-time polymerase chain reaction (PCR) for the EGFP gene, brains of young adult mice whose pups were fathered by Green Mice were collected on the day of parturition (n = 4) and 4 weeks postpartum both in non-lesioned mice (n = 4) and in mice with excitotoxic lesions of the brain (n = 4). Additionally, brain samples were collected from young adult maternal mice at 4 weeks postpartum (n = 4) and ex-breeder stock female mice 2–3 months after delivering their last litter (n = 9) for quantitative real-time PCR for the sex-determining region of the Y chromosome. Maternal mice were euthanized with an overdose of sodium pentobarbital (>200 mg kg−1 intraperitoneal) and transcardially perfused with cold 0.9% saline to wash out the blood. The brain was removed and cut into four blocks. Block 3 corresponded to the region at which the NMDA was injected to induce an excitotoxic lesion. Genomic DNA was extracted with the DNeasy Tissue Kit (Qiagen, Hilden, Germany, http://www1.qiagen.com) following the protocol provided by the manufacturer. The DNA yield was quantified spectrophotometrically by absorbance at 260 nm (Shimadzu UV-Visible Spectrophotometer), and the DNA was stored at −20°C.
For the real-time quantitative PCR for the EGFP gene, the primers and FAM-labeled probes were designed and synthesized by Applied Biosystems (Foster City, CA, http://www.appliedbiosystems.com). The primer and probe sequences and reaction conditions are indicated in Table 1. The plasmid pEGFP-tub (Clontech, Palo Alto, CA, http://www.clontech.com) was prepared and purified with a spin miniprep kit (Qiagen) and serially diluted to generate a standard curve for each real-time PCR run. For the real-time quantitative Y chromosome–specific PCR, we used a TaqMan Gene Expression Assay primer-probe mixture designed for selective amplification of the sex-determining region of the mouse Y chromosome (Applied Biosystems). The standard curve was created by dilution of total genomic DNA from a male mouse brain. All samples were run on an ABI PRISM 7000 Sequence Detection System (Applied Biosystems).
Table Table 1.. Primer and probe sequences and reaction conditions for real-time polymerase chain reaction
5′-TGC TGC TGC CCG ACA A-
5′-TGT GAT CGC GCT TCT CGT T -
5′-*FAMCCA CTA CCT GAG CAC CC-
Each of 40 cycles
2 min 50°C
10 min 95°C
15 sec 95°C
1 min 60°C
Each hemizygous fetal Green Mouse cell contains only one copy of the EGFP gene in genomic DNA . Likewise, each fetal male C57BL/6 laboratory mouse (Mus musculus) cell contains only one copy of the sex-determining region of the mouse Y chromosome . To express the number of cells positive for these fetal markers relative to the total number of cells, the number of cells yielding the genomic DNA analyzed was estimated based on the approximations that the haploid mouse genome is 2.7 × 109 base pair (bp) and that the average molecular weight of a bp is 610 g (i.e., approximately 5.47 pg of genomic DNA for each diploid nucleus). Data are expressed as mean ± SEM.
Fluorescence In Situ Hybridization for the Y Chromosome
Four weeks after delivery, non-lesioned young adult mothers (n = 6) and ex-breeder stock females 2–3 months after delivering their last litter (n=6) were perfused with 0.9% saline followed by 4% paraformaldehyde in phosphate buffer (pH 7.4). Serial coronal sections (10 μm) were cut on a cryostat, recovered on glass slides, and washed briefly in phosphate buffered saline (PBS) before fluorescence in situ hybridization (FISH) with a Cy3-labeled mouse Y chromosome–specific probe (Cambio, Dry Drayton, U.K., http://www.cambio.co.uk) according to an adaptation of the manufacturer's recommended protocol. Briefly, sections were incubated with methanol and acetic acid (3:1 vol/vol) for 45 seconds, air-dried, and incubated with sodium thiocyanate solution (1 M) for 5 minutes at 75°C. The sections then received two 5-minute washes in deionized, distilled water and were postfixed in 4% paraformaldehyde in phosphate buffer (pH 7.4) for 10 minutes and air-dried. Next, the sections were denatured by three 5-minute washes with 60% formamide in ×2 standard saline citrate (SSC) at 75°C, quenched in ethanol, dehydrated through graded alcohols, and air-dried. The Cy3-labeled mouse Y chromosome probe (Cambio) was prewarmed to 37°C, 10–15 μl was applied to the center of the sections, and the sections were individually covered with glass coverslips and sealed with rubber cement. The sections were further denatured in the presence of the probe by cycling to 90°C for 5 minutes on, 2 minutes off, 4 minutes on, 1 minute off, and finally 3 minutes on. The slides were then hybridized overnight at 37°C in a humidified chamber. The coverslips were removed, and the sections underwent three 5-minute washes in 50% formamide in ×2 SSC, three 5-minute washes in ×2 SSC, and two 10-minute washes in 0.05% Tween 20 in ×4 SSC at 42°C. After FISH, the sections were immunostained as described below and mounted with an antifade mounting medium contain 4,6-diamidino-2-phenylindole (Vectashield, Vector Laboratories, Burlingame, CA, http://www.vectorlabs.com).
Four weeks after delivery, mice (n = 10) were perfused with 4% paraformaldehyde in phosphate buffer (pH 7.4). Serial coronal sections (20 μm) were cut on a cryostat. Fluorescent EGFP-positive fetal cells were identified in parenchyma of the cortex, hippocampal formation, or subiculum and adjacent to blood vessels in these regions. Selected sections were immunostained. Primary antibodies used were rat anti-mouse CD11b Alexa Fluor 647-conjugated (1:100; BD Pharmingen, San Diego, http://www.bdbiosciences.com/pharmingen), mouse monoclonal anti-CD45 (1: 100; Santa Cruz Biotechnology, Santa Cruz, CA, http://www.scbt.com), rat anti-mouse F4/80 (1:100; Serotec Ltd., Oxford, U.K., http://www.serotec.com), rabbit anti-GFAP (1:200; DakoCytomation, Glostrup, Denmark, http://www.dakocytomation.com), mouse monoclonal anti-NeuN (1:200; Chemicon, Temecula, CA, http://www.chemicon.com), and polyclonal rabbit anti–von Willebrand factor (anti-vWF, 1:100; DakoCytomation). The secondary antibodies used were FluoroLink Cy2- and Cy3-labeled goat anti-mouse immunoglobulin G (IgG) (Amersham Biosciences), Alexa Fluor 488 and Alex Fluor 568 goat anti-rat IgG (Molecular Probes, Eugene, OR, http://probes.invitrogen.com), Alexa Fluor 488 and Alex Fluor 568 donkey anti-rabbit IgG (Molecular Probes), Alexa Fluor 633 goat anti-rabbit IgG (Molecular Probes), and Alexa Fluor 568 donkey anti-goat IgG (Molecular Probes). All of the secondary antibodies were used at 1:200 dilutions. Briefly, sections were permeablized with 100% cold acetone, washed three times with PBS containing 0.1% Triton X-100 (PBS-Triton), and blocked with 5% bovine serum albumin and 0.1% goat serum or 5% donkey serum, as appropriate for the secondary antibodies used, for 1 hour. Slides were incubated with the primary antibody overnight at room temperature. The slides were then washed three times for 5 minutes in PBS-Triton and incubated with the appropriate secondary antibody for 1 hour at room temperature and then washed three times for 5 minutes in PBS-Triton.
For multiple labeling, the blocking and primary and secondary antibody incubations were repeated sequentially. Some of the sections were further immunostained for EGFP with a mouse monoclonal antibody to GFP (1:100, 1.5 hours; Chemicon) conjugated to the photostable fluorophore Alexa Fluor 488 (Zenon; Molecular Probes).
Sections were examined by sequential scanning and optical sectioning with a confocal microscope (FV500 [Olympus, Tokyo, http://www.olympus-global.com] or LSM510 [Carl Zeiss, Jena, Germany, http://www.zeiss.com]). For illustration, extended-focus XY confocal images were generated from the 3D image stacks, and orthogonal slices through the Z-stacks along the X-, Y-, and Z-axes are shown for selected cells.
Fetal Green Mouse Cells in Maternal Blood and Brain
EGFP-positive Green Mouse fetal cells were detected by FACS in mononuclear cell fractions of maternal blood taken 7 days after delivery from young adult mothers (n = 10) whose pups were fathered by Green Mice (Fig. 1A). When the FACS signals were corrected to the blood of young adult wild-type virgin females as negative controls (Fig. 1B) and the blood of 7-day-old Green Mouse pups as positive controls (Fig. 1C), 0.08 ± 0.02% of nucleated cells in the blood were found to be Green Mouse fetal cells. Samples of the nucleated cell fractions were also visualized directly by phase-contrast and epifluorescence microscopy. EGFP-positive Green Mouse cells were found in the blood of mothers of hemizygous Green Mouse pups (Fig. 1E).
Four weeks after delivery, the mothers were euthanized and perfused with 4% paraformaldehyde in phosphate buffer (pH 7.4). The brains were serially sectioned. Small numbers of EGFP-positive green mouse fetal cells were found in the maternal brains under epiflourescence microscopy (Fig. 1H).
Quantification of Fetal Mouse Cells in Maternal Brain
Quantitative real-time PCR of genomic DNA revealed that, on the day of parturition, 7.1 ± 1.8 cells per 1,000 maternal cells were EGFP-positive Green Mouse fetal cells in the intact brain of young adult mothers (n=4) whose pups were fathered by Green Mice. At 4 weeks postpartum (n = 4), there were 17.5 ± 3.7 fetal cells per 1,000 maternal cells (Fig. 2D), which was significantly greater than on the day of parturition (t-test; p < .05). The number of Green Mouse fetal cells was not significantly greater in the lesioned brain overall (23.3 ± 4.3 fetal cells per 1,000 maternal cells; Fig. 2D). However, within the block of tissue containing the site of the lesion, Block 3, the number of Green Mouse fetal cells was significantly greater in the lesioned maternal brain (29.9 ± 8.5 fetal cells per 1,000 maternal cells) than in the intact maternal brain (5.3 ± 2.6 fetal cells per 1,000 maternal cells; Fig. 2E; t-test, p < .05).
In a separate experiment, real-time PCR of genomic DNA from the brains of young adult wild-type C57BL/6 mothers (n = 4) for the Y chromosome–specific sex-determining region of the mouse Y chromosome produced a similar estimate for the number of fetal cells in the intact maternal brain at 4 weeks postpartum (5.5 ± 1.6 male fetal cells/1,000 maternal cells, which equates to approximately 11 fetal cells/1,000 maternal cells). Quantitative real-time PCR of genomic DNA from brains of C57BL/6 ex-breeder stock female mice at least 2–3 months after delivering their last litter (n = 9) for the sex-determining region of the mouse Y chromosome revealed male cells in the brains of four out of nine female mice. The male cells were found almost exclusively in Block 1, corresponding largely to the olfactory bulb. In those ex-breeder stock females in which male cells were found, the mean number of male cells in Block 1 was 95.8 ± 69.8 male cells per 1,000 maternal cells.
Location of Fetal Cells in the Maternal Brain
Double-immunostaining with anti-GFP antibodies to identify Green Mouse fetal cells and anti-vWF antibodies to identify endothelial cells revealed perivascular fetal Green Mouse cells juxtaposed to blood vessels in the brains of non-lesioned young adult mothers 4 weeks postpartum (Figs. 3, 4). Rarely, these cells juxtaposed to blood vessels appeared to be binucleated (Figs. 3G–3K). Other Green Mouse fetal cells were observed within the brain parenchyma with no obvious association to blood vessels (Figs. 5, 6) and occasionally aligned with the maternal brain cell layers (Figs. 5A–5H). Likewise, fetal cells were identified in the brain parenchyma with no obvious association to blood vessels by FISH for the Y chromosome in the brains of ex-breeder stock female mice at least 2–3 months after delivering their last litter.
Morphology and Immunocytochemistry of Fetal Cells in the Maternal Brain
Four weeks after delivery, Green Mouse fetal cells in the mothers' brains were capable of expressing morphological and immunocytochemical characteristics of diverse cell types. Some of the Green Mouse fetal cells observed within the brain juxtaposed to blood vessels were found to immunostain with an antibody to the macrophage marker F4/80 and would occasionally appear to wrap processes around adjacent endothelial cells (Figs. 4F, 4H). Occasionally, perivascular maternal cells immunolabeling for F4/80 were also observed with similar morphologies (Figs. 4I– 4L). However, there was no evidence of expression of CD11b by fetal cells in the maternal brain. No instances of engulfment of host neural cells by fetal cells labeling with the anti-F4/80 antibody were observed.
Double-labeling with antibodies to GFP and to the neuronal markers MAP2ab and NeuN provided immunocytochemical evidence for expression of characteristics of neuronal cells by Green Mouse fetal cells in the brains of young adult mothers 4 weeks postpartum (Figs. 5A–5H). These anti-GFP–labeled cells also displayed morphological features consistent with neuronal cells and were occasionally found organotypically aligned with the cells of the host CA1 pyramidal layer (Figs. 5A–5H). There was no evidence of F4/80 or CD45 immunoreactivity in Green Mouse fetal cells immunolabeling for these markers of neuronal cell type. Male fetal cells identified by FISH for the Y chromosome in the brains of ex-breeder stock female mice at least 2–3 months after delivery of their last litter also double-labeled with the anti-MAP2ab antibody (Figs. 5I–5L).
Although fetal cells were occasionally closely juxtaposed to NG2-positive host cells, no evidence was found for double-labeling of Green Mouse fetal cells by an antibody to NG2, a marker for immature oligodendrocytes (Figs. 6A–6C). However, other Green Mouse fetal cells labeled with antibodies to MAG and GFAP, markers for oligodendrocytes and astrocytes, respectively (Figs. 6D–6I). There was no evidence of F4/80 or CD45 immunoreactivity in Green Mouse fetal cells immunolabeling for these markers of neural cell type. Male fetal cells identified by FISH for the Y chromosome in the brains of ex-breeder stock female mice at least 2–3 months after delivery of their last litter also double-labeled with the anti-GFAP antibody (Figs. 6J, 6K).
Fetal Green Mouse cells were found to enter wild-type maternal blood and brain, and some of these cells expressed immunocytochemical markers for neural cell types. The presence of fetal Green Mouse cells in maternal blood confirms that, as has been previously demonstrated for crossings of other mouse strains [3, 4], fetal Green Mouse cells can enter the maternal circulation in C57BL/6 mice. Although we also confirmed our findings by using Y chromosome–specific markers to identify male fetal cells, the use of Green Mice ubiquitously expressing EGFP facilitated detection of the fetal cells in maternal blood and tissues. However, EGFP should never be visualized in frozen sections without prior fixation because leaching of soluble EGFP can lead to false negatives. Neither ethanol nor acetone fixatives are sufficient to stabilize EGFP in tissue and, as previously described by others [16, 17], we found it important to fix the EGFP with a formaldehyde-based fixative before sectioning to avoid leaching of soluble EGFP. Subsequent immunofluorescence with an anti-GFP antibody was also beneficial.
We found that at least some of the fetal cells that spontaneously enter maternal circulation during pregnancy are capable of entering the maternal brain. Although fetal Green Mouse cells were rare in the brain, the number recovered is not insubstantial. However, although the blood was removed by perfusion, the number of cells identified in brain tissue by the real-time PCR analysis does not reflect only neural cells but also includes cells engrafted into other niches, such as the perivascular environment. Studies of ex-breeder stock females showed that in the intact brain, fetal cells are present at least 2–3 months after pregnancy in some but not all individuals. In humans, individual differences in engraftment of the blood and skin of mothers by fetal cells have also been reported . These individual differences may be due to differences in traffic across the placental barrier between pregnancies or to the mothers' immune systems and the degree of histocompatibility between the fetal cells and the mothers. The larger numbers of cells observed in some of the individual mice that retained fetal cells suggests that fetal cells can accumulate over multiple pregnancies or can proliferate in the mothers. That greater numbers of fetal cells are found in the maternal brain at 4 weeks postpartum than on the day of parturition could suggest that these cells are progenitor or stem cells capable of proliferation. However, this does not necessarily mean that proliferation occurs within the brain because it has been reported that, at least in humans, fetal cells engraft the BM . In the intact brains of non-lesioned young adult mice and the ex-breeder stock females, the fetal cells were preferentially found in the region of the olfactory bulb. The subventricular zone has been reported to support survival and limited proliferation, migration, and immunocytochemical differentiation of umbilical cord blood cells and BM cells [19, 20]. Perhaps the subventricular zone and the rostral migratory stream to the olfactory bulb offer a niche facilitating fetal cell incorporation into the maternal brain, fetal cell proliferation, or fetal cell survival.
When the maternal brain is injured, these cells preferentially enter the region of the injury. Previous studies had shown that umbilical cord blood cells can cross the blood–brain barrier to enter the injured brain, where they can express some immunocytochemical markers of neural cell types [11, 12]. Injury to the brain could increase the entry of cells from maternal blood circulation by compromising the blood–brain barrier or by releasing signaling molecules that cause fetal cells to be recruited to the brain. In both intact and lesioned brains, fetal cells were found both closely juxtaposed to blood vessels and within the brain parenchyma with no obvious association to blood vessels. Fetal cells expressing morphological and immunocytochemical features characteristic of neuronal cells were also found organotypically aligned with host neurons in the pyramidal cell layer of CA1 of the hippocampus (Figs. 5A–5H). Together, these data suggest that the fetal cells may have migrated within the host brain and developed in response to cues from the host.
Within the brain, fetal cells were capable of expressing morphological and immunocytochemical characteristics of various cell types. Some fetal cells remained in close association with blood vessels, but we found no evidence for endothelial cells of fetal origin in the maternal brain. It might have been predicted that fetal engraftment of the endothelium of maternal vessels would be observed because it has been reported that after focal ischemia the adult mouse brain can recruit endothelial progenitors cells from BM for neovascularization . However, at least in the rat brain, the rate of turnover of brain endothelial cells is very slow , and it may be that unless injury triggers neovascularization, circulating fetal endothelial cells will not be recruited. Besides, at least in humans, it appears that endothelial cells in the blood of pregnant women are of maternal rather than fetal origin [23, 24].
Fetal cells were also observed within the brain in close juxtaposition to the endothelial cells of the host blood vessel wall. The location and morphology of these fetal cells resemble those of BM-derived perivascular macrophages [25–28]. Consistent with differentiation into macrophages, some of these fetal cells were immunolabeled for the macrophage marker F4/80. However, there was no evidence at 4 weeks postpartum in the intact maternal brain that such fetal cells expressed CD11b. This may suggest incomplete differentiation or maturation of perivascular macrophages of fetal origin. In the intact mouse brain, such BM-derived perivascular cells have been reported to phagocytose host endothelial elements . It may be that the apparent wrapping of processes around neighboring endothelial cells observed here (Figs. 4F, 4H) is evidence of similar phagocytosis by fetal-derived perivascular macrophages. After brain injury, BM-derived macrophages and microglia are also reported to infiltrate the brain parenchyma [26–28]. It has been reported that such BM-derived parenchymal microglia can engulf host neural cells . Observations of such engulfment events would show that the donor cells are functional, but we saw little evidence of infiltration of the intact maternal brain parenchyma by fetal-derived macrophages or microglia and did not observe any instances of engulfment of host neural cells by fetal cells.
We observed a rare example of a seemingly binucleated EGFP-positive cell adopting a putatively perivascular macrophage-like location juxtaposed to a blood vessel. This may represent evidence of cell fusion either between two fetal cells or between a host cell and a fetal cell. Alternatively, it may represent a fetal cell caught in the act of division or the development of a polynucleated cell type. Binucleated and multinucleated microglia and macrophages frequently occur in association with central nervous system injury, especially chronic inflammatory injury [30, 31]. It may be that the presence of fetal cells in the maternal brain has led to an inflammatory response. If this is a fusion event, we cannot determine whether this cell fused because it became perivascular macrophage–like or whether it became perivascular macrophage–like because it fused with a host perivascular macrophage.
Fetal cells in the maternal brain were also capable of developing gross morphological similarities to neural cell types and expressing immunocytochemically labeled protein markers normally associated with neural cell types (Figs. 5, 6). As some fetal cells were observed with macrophage-like characteristics, it is important that it has been demonstrated that serial confocal sectioning allows engulfment of differentiated host cells by EGFP-positive donor cells to be readily distinguished from expression of neural markers by EGFP donor cells . The presence of EGFP within the soma and arborizations of cells colabeling for neural markers is inconsistent with macrophage engulfment (e.g., Figs. 5D, 5H). Further studies might determine whether these cells become electrophysiologically functional neurons. We also found no evidence of EGFP-positive fetal cells with neural-like morphologies or immunocytochemically labeling for neuronal markers colabeling for F4/80 or CD45. The lack of expression of these markers suggests that these cells were not macrophages. The lack of expression of CD45 by fetal cells in the maternal brain may also suggest that the fetal cells that infiltrate the maternal brain and adopt characteristics typical of neural cell types are not of hematopoietic origin. However, we cannot exclude the possibility that these cells expressed CD45 before entry into the maternal brain and then stopped expressing CD45 when in the brain.
No Green Mouse fetal cells were found expressing the immature oligodendrocyte marker NG2, although occasionally EGFP-positive cells were found in close juxtaposition to NG2-positive cells (Fig. 5C). This could suggest that, when they differentiate along an oligodendrocytic path, the Green Mouse fetal cells do not form an active oligodendrocyte progenitor population in the maternal brain but rather differentiate into nonproliferating mature oligodendrocytes.
The Green Mouse fetal cells observed in the maternal brain usually occurred in clusters. Frequently, these clusters contained a number of EGFP-positive cells with different morphological and immunocytochemical profiles (e.g., Fig. 6). This clustering may indicate that a single fetal progenitor or stem cell entered the brain and then proliferated to produce daughter cells following various differentiation pathways in subsequent generations. Alternatively, clustering may indicate that multiple fetal cells enter the brain at particular locations, perhaps attracted by release of signaling molecules or because of local variations in the permeability of the blood–brain barrier to infiltration by fetal cells, and that these multiple fetal cells then followed different differentiation pathways or fused with different types of maternal cells. Further studies are required to investigate the mechanisms of this clustering and why these cells express markers for different cell types.
These data suggest that pregnancy is a minimally invasive model allowing entry of fetal cells into the maternal brain. This model will facilitate comparisons of the fate of fetal cells and endogenous adult progenitor or stem cells in the adult brain. It has been speculated that in vitro isolation and culture can alter the properties of progenitor or stem cells. The pregnancy model will also allow investigation of the fate of fetal cells in the adult brain without in vitro manipulation. Further characterization of fetal cells capable of crossing the blood–brain barrier may improve selection procedures for isolation of progenitor or stem cells from other sources, for example, directly from umbilical cord blood, for brain repair by intravenous infusion.
Recently, there has been much speculation that fetal microchimerism may have implications in maternal health [32–35]. It is possible that fetal microchimeric cells may participate in the maternal response to injury . It is known that hormonal changes during pregnancy can influence neurogenesis , and it may be that pregnancy makes certain niches in the brain a more receptive environment for fetal cells. Further studies are required to determine whether there are any functional or pathological implications of the engraftment of the maternal brain with fetal cells during pregnancy.
This research was supported by the National University of Singapore Young Investigator Award to G.S.D. and by grants from the National Medical Research Council of Singapore, Singapore Health Services Pte Ltd., and the Department of Clinical Research, Singapore General Hospital, to Z.C.X. We thank Prof. Catherine J. Pallen for helpful comments on the draft manuscript. We thank the Department of Experimental Surgery, Singapore General Hospital, for provision of animal housing facilities.
The authors indicate no potential conflicts of interest.