Bone marrow-derived SP cells can contribute to the respiratory tract of mice in vivo

Adult stem cells from bone marrow with the potential to give rise to cells of multiple lineages have been described. Cells derived from whole bone marrow (Theise et al., 2002), or from defined marrow populations (Gussoni et al., 1999; Jiang et al., 2002; Krause et al., 2001; Lagasse et al., 2000), have been reported in a variety of tissues including skeletal and cardiac muscle (Ferrari et al., 1998; Gussoni et al., 1999), endothelium (Jackson et al., 2001), brain (Mezey et al., 2000), lung epithelium (Kotton et al., 2001; Krause et al., 2001), gut (Okamoto et al., 2002), skin and liver (Korbling et al., 2002; Lagasse et al., 2000; Newsome et al., 2003). This apparent stem cell `plasticity' has great therapeutic potential, but requires rigorous validation before use in a clinical setting. Should a bone marrow-derived cell population have the ability to contribute to the respiratory tract then the possibility of autologous therapy for cystic fibrosis, for example, is an exciting prospect.

Of greatest interest in understanding the process are reports that describe the use of defined cell populations. Multipotent adult progenitor cells (MAPC), derived from adult bone marrow and then cultured for many cell doublings, have been reported to contribute to many organs in the embryo following blastocyst injection (Jiang et al., 2002). The lung was one of the tissues that appeared to show high-level chimaerism with donor and host cells. Other work using single cells that had previously homed to the bone marrow, reported up to 20% engraftment of the lung 11 months after donor cell injection (Krause et al., 2001). High levels of engraftment in the lung were also reported using whole bone marrow as a source of donor cells (Theise et al., 2002). Radiation pneumonitis apparently induced by the irradiation procedure encouraged repopulation of 2-14% of type II pneumocytes in the alveoli. Recently it was reported that 0.6% of lung epithelial cells in mice following bone marrow transplantation were donor-derived (Harris et al., 2004). Non-fused donor-derived cells that express epithelial markers were also detected, but at lower frequency, in the liver (0.05%) and the skin (0.1%).

High-level repopulation in the lung, after radiation-damage, is in keeping with what is observed in other organs where tissue damage is generally required to raise the frequency of donor-derived cells above an almost undetectable level. Indeed several publications using unchallenged subjects, have reported an inability to detect this phenomenon at all (Balsam et al., 2004; Castro et al., 2002; Wagers et al., 2002). Wagers et al. (Wagers et al., 2000) reported little evidence for developmental plasticity of adult haematopoietic cells using GFP marked single donor cells and this included the lung (Wagers et al., 2002). This lack of ability to detect donor cells is also demonstrable in mice where the donor cells are either injected directly into the damaged organ of interest, the heart in this case (Balsam et al., 2004; Murry et al., 2004), or in parabiotic mice where the donor and recipient blood circulations are joined (Wagers et al., 2002).

It is important to investigate the ability of defined donor cell populations to carry out this process. Introduction of the side population (SP) of bone marrow cells into irradiated mdx mice, appeared to rescue the depleted haematopoietic component and also resulted in the appearance of donor cells in skeletal muscle fibres (Gussoni et al., 1999). The SP population, isolated by virtue of its ability to exclude the DNA-binding dye Hoechst 33342 and analysed by flow cytometry, is present in bone marrow at a low level (0.02-0.08%) (Goodell et al., 1996). SP cells in the bone marrow express markers consistent with its blood stem cell phenotype, i.e. CD45⁺, ScaI⁺, lin^–. Single SP cells from bone marrow isolated by fluorescent activated cell sorting (FACS) have been shown to have the ability to rescue the haematopoietic defect in lethally irradiated mice, and give rise to differentiated skeletal muscle cells, functional hepatocytes and osteoblastic cells (Camargo et al., 2003; Camargo et al., 2004; Olmsted-Davis et al., 2003). It was suggested that, in transplanted animals, the donor-derived cells arise by fusion with host cells (Camargo et al., 2004). However, all these cell types exist normally as fused cells so this may be just the cells behaving as the tissue dictates. Recently, two studies have been reported showing no evidence for cell fusion in the lung or skin (Brittan et al., 2005; Harris et al., 2004).

Here we investigate the ability of bone marrow-derived SP cells to contribute to the upper respiratory tract, an organ that does not normally contain fused cells. We demonstrate the ability of our freshly isolated, genetically marked, SP cells to rescue lethally irradiated mice. Analysis of X-gal staining and ROSA26 allele-specific PCR indicated that these cells did not have the ability to contribute to the developing blastocyst, nor were they able to contribute to primary epithelial cultures grown at an air liquid interface (ALI) or denuded tracheal xenografts. Finally we examine the ability of male SP cells in engrafted female mice to contribute in vivo to damaged respiratory epithelium. We show clear evidence for the presence of donor cells by Y chromosome FISH and donor cell-specific PCR. In addition, we show that some of these donor cells express cytokeratin, a marker of epithelial cells.

Male ROSA26 mice of 6-10 weeks of age and backcrossed for five generations to CBA/Ca, were bred in house and used to isolate bone marrow. SCID female mice age 4-6 weeks, obtained from Charles River UK (England, UK) were used as recipients for subcutaneous grafts and housed in individually vented cages (IVC). C57Bl/6 males aged 10-12 weeks were purchased from Charles River UK (England, UK) and the tracheas used for subcutaneous grafting. CBA/Ca male mice aged 4-6 weeks obtained from Charles River UK (England, UK) were used as control donors of primary epithelial cells. CD1 females aged 3-4 weeks obtained from Charles River UK (England, UK) were used as blastocyst donors and those aged 8-12 weeks used as recipients of injected blastocysts.

Bone marrow was extracted from the femurs and tibias of mice, and a single cell suspension prepared by passing the bone marrow through a 21-guage needle after which the cells were pelleted by centrifugation and resuspended at 10⁶ cells ml^–1 in DMEM that contained 2% fetal calf serum (FCS) and 10 mM Hepes (Gibco-BRL, Paisley, UK).

SP cells were isolated from the bone marrow according to published methods (Goodell et al., 1996). Briefly, bone marrow was resuspended at 10⁶ cells ml^–1 in prewarmed DMEM as outlined above. Hoechst 33342 (Bis-Benzimide) (Sigma Aldrich, Dorset, UK) was added to the cells at a final concentration of 5 μg ml^–1 and the cells incubated at 37°C for 90 minutes. After this time, the cells were pelleted and resuspended in ice cold HBSS containing 2% FCS and 10 mM Hepes (Gibco-BRL, Paisley, UK) and maintained at 4°C for analysis by flow cytometry.

Analysis and sorting was performed using a FACSVantage, equipped with a Coherent INNOVA Enterprise II laser (Becton and Dickinson, Oxford, UK) and performed as originally described by Goodell et al. (Goodell et al., 1996). Briefly, the Hoechst dye was excited at 350 nm and its fluorescence emission was collected with a 424/44 band pass (BP) filter (Hoechst blue) and a 675/20 BP filter (Hoechst red). A 610 SP was used to separate the blue and red emissions. Propidium Iodide (PI) (Sigma Aldrich, Dorset, UK) was added to the cells prior to sorting at a final concentration of 2 μg ml^–1 to distinguish and exclude dead cells in the bone marrow population. Sort gates were set as defined by Goodell et al. (Goodell et al., 1996). After acquiring 1×10⁵ live cells, the SP gate could be clearly defined. The SP cells were then sorted into polypropylene tubes (BD Labware Europe, Combourg, France) containing 100% FCS and an aliquot re-analysed to check cell sort purity. The SP cells were also immunophenotyped using FITC conjugated Sca-1 and R-phycoerythrin conjugated Gr-1 (BD Biosciences, Pharmingen, Oxford, UK).

Engraftment levels of the bone marrow of recipients of ROSA26-derived cells was assessed using the conversion of the non-fluorescent substrate Fluorescein Di-(β-D-galactopyranoside) (FDG) (Sigma, Poole, UK) to a fluorescent product. Briefly, 40 μl of pelleted cells were warmed to 37°C in a water bath, to which was added 40 μl working solution FDG (2 mM in distilled H₂O) diluted from stock made up in DMSO at 100 mM. Following a further 2 minutes at 37°C, 2 ml of ice cold PBS were added, cells were then kept on ice for a further 30 minutes before assessment of fluorescence by flow cytometry. Data for up to 40,000 cells were acquired using a FACSCaliber bench top cytometer equipped with a 488 nm laser (Becton and Dickinson, Oxford, UK) without compensation. Dead cells and erythrocytes were excluded from acquisition based on their forward and side scatter characteristics. The fluorescent product of FDG was detected in fluorescence 1 channel (FL1). Data was analysed using CellQuest software (Becton and Dickinson, Oxford, UK).

These cells were cultured according to the method of Davidson et al. (Davidson et al., 2000) that had been previously established in this laboratory. Briefly, tracheae were excised from the appropriate mice, washed in phosphate buffered saline (PBS) and cells dissociated in calcium and magnesium free minimum essential media (MEM) containing penicillin (60 i.u. ml^–1; Gibco BRL), streptomycin (60 mg ml^–1; Gibco BRL), pronase (1.4 mg ml^–1; Boehringer Mannheim UK, Lewes, East Sussex, UK) and DNAse (0.1 mg ml^–1; Sigma-Aldrich; Poole, Dorset, UK) at 37°C for 60 minutes. Cells were pelleted and resuspended in Epithelial Cell Culture Media (ECCM) containing a 1:1 mix of DMEM (Gibco BRL) and HAM'S F12 (Gibco BRL) supplemented with 5% FCS (Gibco BRL), penicillin (100 i.u. ml^–1; Gibco BRL), streptomycin (100 μg ml^–1; Gibco BRL) and Insulin (120 i.u. l^–1; AAH Pharmaceuticals, Glasgow, UK). This cell suspension was then incubated at 37°C for 2 hours on a 100 mm tissue culture dish (Becton Dickinson UK, Oxford, UK) and unattached cells collected, centrifuged and resuspended in ECCM. The appropriate number of epithelial cells were seeded onto 24 well semi-permeable support membranes (Corning Costar, High Wycombe, UK) that had been pre-coated with 50 μg human placental collagen (Sigma-Aldrich Company, Poole UK). The cells were then incubated at 37°C, 6% CO₂ in a humidified incubator and on day 4 the media on the apical surface of the cultured cells was removed and the media on the outside of the insert replaced with USG Media that contained 1:1 DMEM and HAM'S F12 containing penicillin and streptomycin as above and supplemented with 2% Ultroser-G serum substitute (USG) (Gibco BRL). The media on the outside of the insert was subsequently replaced twice weekly whilst the cultures were maintained. SP cells were added to the insert at the same time as the primary epithelial cells.

Total DNA was extracted from cells, tissues or tissue sections using a rapid lysis method. Briefly, 50 μl of 25 mM NaOH 0.2 mM EDTA pH 12 was added to the sample and heated to 95°C for 20 minutes. Then 50 μl of 40 mM Tris HCl was added and the solution was vortexed. The DNA was stored at –20°C until use in the polymerase chain reaction (PCR), which was carried out with Taq polymerase (Roche Diagnostics, Lewes, UK) according to the manufacturer's instructions. The presence of LacZ DNA was studied by PCR using sequence specific primers with amplification DNA isolated from ROSA26 tissue as an internal control. Wild type (WT) DNA was also amplified using sequence specific primers that span the site of the LacZ insertion. DNA could also be amplified using sequence specific Y chromosome primers. Sequences of the primer pairs were as follows (all written 5′-3′ with forward primer first): LacZ (product size 276 bp) GACGTCTCGTTGCTGCATAA and GCTCCACAGTTTCGGGTTTTCG; WT (product size 354 bp) GGCTTAAAGGCTAACCTGATGTG and GGAGCGGGAGAAATGGATATG; Y chromosome (product size 310 bp) GATTCCATGAGGCACCAT and TATCTGCTTTCTCCACGACC. The primers were used at a final concentration of 1 μM each in the PCR reaction, which were carried out under standard conditions. The thermal cycling protocol for LacZ and WT PCR comprised an initial denaturation step at 94°C for 4 minutes followed by 43 cycles of 94°C for 1 minute, 65°C for 1 minute and 72°C for 1 minute. The final cycle consisted of a re-annealing at 72°C for 10 minutes. A similar protocol was used for Y chromosome amplification except that the annealing temperature was 60°C for 45 seconds. PCR products were visualised on ethidium bromide stained 2% agarose gels using a Herolab Easy RH-3 system (Scotlab, Coatbridge, UK).

Fertilised blastocysts obtained from superovulated CD1 females aged 3-4 weeks were injected with SP cells that had been derived from ROSA26 bone marrow and then transferred into recipient CD1 females aged 8-12 weeks. Embryos were then harvested at appropriate developmental stages.

Slides were dewaxed by washing 3 times in xylene for 5 minutes before rehydration through an ethanol series (100%, 90%, 70%, 50%, 30%) and then placed into water. The slides were then microwaved for 20 minutes in 0.1 M citrate buffer ph 6.0 after which they were denatured for 3 minutes in 70% formamide 2×SSC, plunged into ice cold 70% ethanol and dehydrated through an alcohol series and air-dried.

Y chromosome FISH was carried out using a biotinylated whole chromosome Y paint (Cambio, Cambridge, UK) according to the manufacturers instructions. Briefly, the probe was hybridised to the section overnight at 37°C. Slides were then washed 4 times for 3 minutes in 2×SSC at 45°C then incubated for 30 minutes at 37°C with avidin Texas Red (Vector Laboratories, Peterborough UK). This was followed by washing 3 times for 2 minutes in 4×SSC, 0.1% Tween20 at 37°C after which time biotinylated anti-avidin (Vector Laboratories, Peterborough UK) was added and the slides incubated further for 30 minutes. The slides were then washed as above after which time avidin Texas Red was again added to the slides for a further 30 minutes incubation at 37°C. After washing the slides as above, they were mounted in Vectashield (Vector Laboratories, Peterborough UK) containing 1 μg ml^–1 4,6-diaminidino-2-phenylindole (DAPI) counterstain.

CBA/Ca female mice (8-12 weeks) were lethally irradiated with 1050 rads delivered from a GammaCell 40E (MDS Nordian, Fleuvus, Belgium) with a Cesium 137 source at a dose-rate of 114 Rads per minute. Following irradiation they received a single tail-vein infusion of 1×10⁴ bone marrow-derived SP cells. The animals were housed under specific pathogen free (SPF) conditions in IVCs for 3 months after which time the tracheas were damaged by instillation with 10 μl 2% polidocanol (Sigma, Poole, UK) and the animals harvested 7 days later. Experimental protocols involving animals were carried out in accordance with permits and guidelines issued by the MRC Ethical Review Committee and the United Kingdom Home Office.

Tracheas were removed from the appropriate animals and fixed in 4% paraformaldehyde (PFA), processed to paraffin wax blocks, cut to 7 μM sections by microtome. Sections were then deparaffinised in xylene and hydrated through an alcohol series before mounting for viewing, carrying out immunostaining or for FISH analysis. Some tissues or cell cultures were also stained for β-galactosidase enzyme activity using the standard X-gal substrate staining procedure after which time the tissues were processed to wax blocks as outlined above. Sections were stained with hematoxylin and eosin (H and E) according to a standard procedure.

Cytokeratin staining was performed after FISH. Slides were blocked with 1% IgG (Sigma, Poole, UK) diluted in 5% donkey serum (DS) (Sigma, Poole, UK) for 20 minutes at RT after which time the slides were washed in PBS after which time the slides were blocked again in 5% DS, 2% BSA in PBST (0.1% Tween in PBS) for 30 minutes at RT. The slides were washed again and incubated with pan cytokeratin antibody (Sigma, Poole, UK) (1 in 20) overnight at 4°C. The slides were washed and then incubated with FITC labelled anti-F(ab) fragment (1:50) (Jackson ImmunoResearch Laboratories, USA) secondary antibody for 2 hours at RT then washed and mounted with Vectashield containing 1 μg ml^–1 DAPI.

Tracheas were isolated from C57Bl/6 male mice aged 10-12 weeks. Cells were removed from the tracheas by the process of freezing and thawing the tissue repeatedly at –70°C. Having tied the lower end of the trachea, the appropriate cells or media alone were added into the trachea in ECCM and the upper end of the trachea tied securing the cells and media within. SCID female recipient mice aged 4-6 weeks were used as hosts for the grafts. Mice were anaesthetised with an intraperitoneal (ip) injection of Hypnorm (0.4 ml kg^–1) and midazolam (5 mg kg^–1) and shaved on both sides of the body. The skin was cleaned with 70% ethanol and a small incision was made. The skin was retracted and a donor trachea was inserted into the deep subcutaneous connective tissue. Autoclips (International Market Supplies, Cheshire, UK) were used to close the skin wound. Animals were then brought round from anaesthetic using an ip injection of Vetergesic (0.4 ml kg^–1). The tracheas were removed 9, 21 and 42 days post implantation into the recipients and fixed and processed for histology as outlined above.

Slides were visualised using a Zeiss Axioplan 2 microscope (Carl Zeiss UK, Welwyn Garden City, UK) equipped with Ludl filter wheel (Ludl Electronic Products, Hawthorne, NY) and Chroma 83000 triple bandpass filter set (Chroma Technology Corp., Rockingham, VT). Brightfield and phase microscopy were also performed on this microscope. Grayscale images were collected with a Coolsnap HQ cooled CCD camera (Roper Scientific, Tucson, AZ). In-house scripts written for IPLab (Scanalytics Corp., Fairfax VA) were employed for image capture and image processing. Slides were also visualised using a Zeiss LSM 510 Meta confocal microscope (Carl Zeiss UK, Welwyn Garden City, UK).

SP cells were isolated from the bone marrow of ROSA26 mice. These mice express the bacterial β-galactosidase gene under the control of a ubiquitous promoter (Friedrich and Soriano, 1991). Fig. 1A,B show the characteristic SP profile of these cells following Hoechst incubation. The cell population was found to be 69.16% positive for Sca-I and 4.20% positive for Gr-I (Fig. 1C), which compares well with phenotypic data previously described for bone marrow-derived SP cells (Camargo et al., 2003; Goodell et al., 1997). Cytospins of isolated SP cells showed them to have a high nuclear to cytoplasmic ratio (Fig. 1D). After injection of 1×10⁴ male bone marrow-derived ROSA26 SP cells into immunocompatible CBA/Ca lethally irradiated female mice, 5/6 animals survived. The surviving animals were sacrificed 3 months after injection and 7 days post polidocanol instillation, and the degree of marrow contribution was assessed. All showed high-levels of bone marrow reconstitution by the ROSA26 donor cells, as compared with the levels seen in positive and negative controls (ROSA26 and CBA/Ca respectively) (Fig. 2).

It has been reported that adult mouse mesenchymal and brain cells can contribute to all three germ layers in chimaeras following blastocyst injection (Clarke et al., 2000; Jiang et al., 2002). We injected a total of 2500 cells from the SP population, and saw no evidence of contribution after injection of 10-40 cells into 144 blastocysts and analysis of 65 live embryos. Donor cell presence was assayed by X-gal staining of embryos and confirmed by LacZ PCR specific for the ROSA26 allele (Fig. 3). Sensitivity was at 1 cell in 1000 (Fig. 3A) and PCR Gels were blotted and probed using a radio-labelled internal oligo probe to further increase the sensitivity (data not shown). In none of the 65 embryos resulting from SP injection did we observe a ROSA26 positive PCR signal (Fig. 3C,D) even when DNA was extracted from what appeared to be blue X-gal positively stained tissues. Camargo et al. (Camargo et al., 2003) report that 25% of their SP cell preparations are capable of rescuing a lethally irradiated mouse following intra venous injection of just a single cell. Thus, given the success of our SP population at rescuing lethally irradiated animals we would have expected to see evidence of chimaeric generation if these freshly isolated cells are truly pluripotent. We conclude that SP cells isolated direct from the bone marrow are not pluripotent.

The tracheal epithelium of mice is rich in Clara cells and, therefore, are a good model for bronchioles in human, which is the initiating site of lung disease in cystic fibrosis. We have previously described the isolation of terminally differentiated epithelial cells from the trachea of mice, and subsequent culture on collagen coated semi permeable supports at an air interface (Davidson et al., 2000). These air liquid interface (ALI) cultures form confluent polarized epithelia with high transepithelial resistances. Immunohistochemistry, electrophysiology, and light and electron microscopy demonstrated that the cells are epithelial in nature. SP cells from ROSA26 mice were seeded onto the membranes with primary epithelial cells isolated from the tracheas of C57Bl/6 mice. Eight days after seeding the cultures, ciliated cells were obvious by eye and resistances over 12 kΩ cm^–2 were recorded. Cultures set up using epithelial cells isolated from ROSA26 mice and C57Bl/6 mice alone were used as the positive and negative staining controls. Cultures were analysed after 28 days and stained for β-galactosidase activity using the X-gal substrate. ROSA26 cells showed characteristic blue stained cells (Fig. 4A) whereas C57Bl/6 epithelial cells showed no staining (Fig. 4B). There was no evidence of X-gal positive cells in the mature epithelium of cultures containing ROSA26 SP cells (Fig. 4C).

In an effort to improve the chance of the SP cells contributing to the epithelial cultures, we attempted to observe their contribution to respiratory epithelia grown on denuded tracheal graft maintained subcutaneously on the flanks of SCID mice. We hypothesised that perhaps extracellular matrix factors that were missing from the ALI cultures, would be present in this system. The normal epithelial layer of the trachea (Fig. 5A) is removed upon freeze-thawing of the tissue (Fig. 5B). However on seeding 1×10⁵ tracheal cells isolated from the tracheas of C57Bl/6 or ROSA26 mice into freeze/thaw denuded trachea, we were able to produce an epithelial layer of reasonable integrity (Fig. 5C). Media alone gave no cell layer (data not shown). When 2×10⁴ SP cells from ROSA26 bone marrow were seeded with 1×10⁵ unmarked, differentiated cells into denuded tracheas, an epithelial layer was formed (Fig. 5D), but no blue cells were present after X-gal staining (Fig. 5E). PCR specific for the ROSA26 allele was carried out on DNA extracted from the histology slides of the sub-cutaneous grafts and was found to be negative (data not shown) even though the PCR was found to be sensitive to 1 cell in 1000 (Fig. 3A).

In order to assess the ability of SP donor cells to contribute to the respiratory tract in vivo, we examined the airway tissue of the lethally irradiated female mice, rescued by injection of male SP Rosa26 bone marrow-derived cells.

Damage of tissues appears to improve the ability of marrow-derived cells to contribute to non-haematopoietic organs. We used 2% polidocanol to damage the trachea of transplanted mice and showed that this was effective at completely stripping the epithelial layer of the trachea (Fig. 6A,B). After 7 days the epithelial lining had been replaced (Borthwick et al., 2001). The mice were sacrificed 3 months after intra-venous injection of 1×10⁴ bone marrow-derived ROSA26 SP cells and 7 days following 2% polidocanol instillation. Although blue cells were readily detectable in ROSA26 tracheas by X-gal staining (Fig. 7A,B), they were not detectable in the tracheas of the transplanted mice (Fig. 7C,D). However, PCR analysis of DNA extracted from tracheal tissue sections of these mice, showed a strong LacZ product (Fig. 7E) indicating the presence of donor cells within this tissue. This finding was confirmed by Y chromosome PCR showing that male ROSA26 donor cells were present in the trachea of the animals (Fig. 7F). Y chromosome FISH analysis of tracheal sections showed the presence of donor cells on the tracheal epithelium of 4 out of the 5 surviving animals (Fig. 8A-D). The average number of Y chromosome positive donor cells detected on the airway epithelia was 0.83% (range 0.45-1.15%) (Table 1). However, this is likely to be an underestimate of the true proportion of donor-derived cells, since only 87% of nuclei in the control male section appeared Y chromosome positive by FISH analysis (609/700). This is because the whole nucleus is not present in the section, and the plane of focus used to capture an the image could exclude Y chromosome signals.

Sections were stained for Y chromosome FISH and then subjected to immunohistochemistry for the epithelial pan cytokeratin marker (Fig. 9). It is important to use the same slide for both analyses as serial sections can be misleading and error prone. Y chromosome positive (donor) cells within the epithelial layer were found to be cytokeratin negative (Fig. 9C-F) or positive (Fig. 9G-J). Slides subjected to this dual analysis and analysed by widefield microscopy were also visualised by confocal microscopy to confirm the presence of both Y chromosome and pan cytokeratin positive signals in the same cells (Fig. 9K). By confocal analysis, 55% of Y chromosome positive (donor) cells were found to be positive for the pan cytokeratin epithelial marker.

We have examined the multipotency of freshly isolated bone-marrow-derived SP cells and their ability to contribute to the respiratory tract.

Although our SP cells efficiently rescue lethally irradiated mice, they did not contribute to embryos after blastocyst injection. We injected between 10 and 40 SP cells into 144 blastocysts. This resulted in 65 embryos being analysed, but no donor-derived cells were detected. Camargo et al. (Camargo et al., 2003) demonstrated that 10% of their SP population was most enriched for haematopoietic stem cells (HSC) and 25% of single cells could rescue lethally irradiated animals. Consequently, the fewest cells we injected into the surviving blastocysts, was 975 cells and so we would expect these to contain at least 24 of the highly potent HSC. Clearly, had the SP population had the same properties as MAPC we might have hoped to see chimaeras by virtue of X-gal positive embryos. We therefore conclude that freshly isolated SP do not have the same multipotent properties as MAPC and are not pluripotent.

Terminally differentiated tracheal epithelia can be cultured in vitro at an ALI, where they form a polarised epithelia with electrophysiology and structure consistent with mature upper airway epithelium. SP cells from muscle (but not bone marrow) can undergo myogenic specification after co-culture with myoblasts and muscle-derived stem cells have the potential to give rise to myogenic cells via a myocyte-mediated inductive interaction (Asakura et al., 2002). Recently Coraux et al. (Coraux et al., 2005) have demonstrated that when cultured at an ALI, ES cells can give rise to a fully differentiated airway epithelium with ultrastructural features and secretory functions characteristic of this tissue. We cultured the adult Rosa26-marked marrow-derived SP cells in a very similar way and found that in co-culture with unmarked terminally differentiated tracheal cells, ALI cultures with beating cilia and high resistances were established. The SP cells however, did not contribute to these cultures.

In order to be confident that the correct extracellular matrix factors were present we also seeded denuded trachea with Rosa 26 marrow-derived SP cells and carrier primary tracheal cells and looked again for both X-gal staining or the Rosa 26 targeted allele PCR signal. The SP cell seeded xenografts were negative by both criteria. Clearly the necessary signals/factors are not present in these in vitro and ex vivo systems to allow marrow-derived SP cells to contribute to the formation of the epithelia.

Several reports, suggest that the lung in particular can support bone marrow-derived cell engraftment in both mice and human, at a level close to one that may be therapeutic (Grove et al., 2002; Theise et al., 2002). The number of bone marrow-derived cells following haematopoietic engraftment in liver and muscle is very low (Wagers and Weissman, 2004). Colonisation of the fah–/– mouse liver with donor cells appears to result from fusion and clonal expansion as a result of selection. However the number of clonal nodules of donor cells is estimated to be only 200-400 per liver, reflecting a donor cell hepatocyte population in normal mice without clonal expansion of only 1 in 300,000 (Camargo et al., 2004). This low level of tissue engraftment is also seen in muscle, although it does increase after injury (LaBarge and Blau, 2002). Harris et al., report lung epithelial cells being derived from mice following irradiation and bone marrow transplantation (Harris et al., 2004). They report 0.6% of alveolar lung cells being donor-derived with no evidence of cell fusion. We examined the tracheal epithelium in our female, lethally irradiated mice rescued by i.v. injection of male SP cells and subjected to Polidocinol damage, after 3 months. Initially we used the β-galactosidase colourimetric assay as all cells in the airway of Rosa 26 mice are heavily stained after X-gal treatment. This assay is frequently used to detect door cells in engrafted mice and we have used this assay previously in morula aggregation chimaeras to study cell lineages in the trachea (Borthwick et al., 1999). To our surprise we could see no blue cells in the epithelial lining the trachea of the SP cell rescued mice even following profound epithelial damage and repair. However, PCR analysis indicated a strong signal for both for Y chromosome and the ROSA26 allele in sections that were X-gal negative. Y chromosome FISH of tracheal tissue, revealed 0.8% of cells were indeed Y chromosome positive in 4 out of 5 mice. We showed that some of these donor-derived cells were cytokeratin positive.

Staining bone marrow cells with the β-galactosidase substrate FDG, revealed the level of chimaerism in these mice was comparable to a ROSA26 homozygote, and demonstrated the functionality of the β-galactosidase enzyme in the marrow.

Expression of a ubiquitously expressed marker gene is a common strategy to assay for donor-derived epithelial cells. Indeed the majority of studies in this field use expression of a marker gene to identify donor-derived cells. However, Alverez-Dolado et al. (Alverez-Dolado et al., 2003) reported donor-derived cells that expressed tissue specific markers that were negative for expression of a β-actin-driven donor cell carrying transgene. They identify donor-derived cells by virtue of cre/lox recombination and this also identifies cells that have undergone fusion. They suggest that cell fusion may result in donor gene inactivation or elimination over time (Alvarez-Dolado et al., 2003). Support for this theory also comes from a group that demonstrate wild type donor cells present in cardiomyocytes or skeletal muscle after transplantation into a sarcoglycan deficient mouse (Lapidos et al., 2004). The donor-derived muscle cells do not express the differentiated muscle gene sarcoglycan. Therefore it is possible that our 0.8% Y chromosome positive cells may be fused cells.

The fact that we do not see evidence of marrow-derived SP cell contribution to the ALI cultures or tracheal xenografts, but do see cells in the airway following transplantation implies that the SP cells must require cues from the damaged tissue to migrate from the blood to the airway epithelia. These experiments imply that SP cells themselves are not mature enough to differentiate to airway epithelial cells.

Our results are also consistent with the idea that mature myeloid cells are responsible for the transdifferentiation process. The elegant work by Camargo et al. (Camargo et al., 2003; Camargo et al., 2004) suggests that the myofibres and hepatocytes that express donor-derived markers, are derived from donor myeloid cell fusion with recipient tissue-specific cells. The SP cells in vitro or ex vivo will not have had the cues to differentiate towards a myeloid lineage. It will be interesting to observe whether the introduction of transgene marked mature myeloid cells into these systems will promote transgene positive epithelial cells. If the cells we observe as donor-derived that express cytokeratin (but are not positive for X-gal staining), arise by fusion, this will be novel as the muscle, bone and liver are normally fusogenic tissues whereas the cells of the trachea (predominantly Clara cells, basal cells and ciliated columnar cells) are not.

The process of fusion however cannot be the only mechanism responsible for apparent `transdifferentation' as it has been reported that NOD/SCID mice injected with human cord blood possess hepatocytes that have nuclei that are only positive for human DNA, and not mouse DNA (Newsome et al., 2003). In addition, Harris et al. (Harris et al., 2004), report a lack of evidence of fusion in the lung, and although there are some caveats to their observations, they are supported by studies in the epidermis (Brittan et al., 2005). Others (Zhang et al., 2004), report that both cell fusion and transdifferentiation may account for the transformation of CD34⁺ cells into cardiomyocytes in vivo.

Progenitor population(s) in trachea must exist as the airway is fairly quiescent, but damage (such as that precipitated by polidocanol) results in epithelial replacement within 7 days. We have previously identified a population of label retaining cells in the trachea of mice following epithelial damage of the airway in vivo. These are located in a putative stem cell niche at the neck of submucosal glands in the proximal region of the trachea, and at the intercartilaginous junctions further down (Borthwick et al., 2001). In the majority of experiments recording bone marrow-derived cell contribution to other organs, the population of local stem cell niches is not observed. In the small intestine, for example, crypts do not become clonal as would happen if a stem cell in the crypt was derived from the donor (Krause et al., 2001). Repopulation of a resident progenitor population has been reported in the muscle using a non-purified marrow cell population (LaBarge and Blau, 2002). In this experiment the satellite cells of the muscle appeared to be the target of reconstitution, but this is not observed in experiments using single SP HSC as donor cells (Camargo et al., 2003) where only mature muscle fibres were recognised as being of donor origin. Again this implies that the nature of the donor-derived engraftment occurs via circulating cells within a tissue. The number of stem cells will always be lower than any other cell type in a tissue and so if the process is purely by myeloid cells fusion with an endogenous cell of the tissue, then by chance stem cells will rarely be accessed and perhaps that is what is reflected in the patterns most frequently observed. We see no evidence of clonal expansion of the Y chromosome containing cells in the airway following detergent-induced epithelial stripping and repair. The tissue engraftment was analysed one week after the complete epithelial lining was stripped away by polidocanol. Thus the donor-derived cells identified must have repopulated the airway during the 7 day regeneration period.

In conclusion, we see no evidence for adult bone marrow-derived SP cells contributing to the developing embryos following blastocyst injection. In addition they do not contribute to the tracheal epithelium in our in vitro or ex vivo models of epithelial cell culture. However following irradiation and SP cell injection we do see evidence of Y chromosome containing donor cells in 0.8% of the tracheal epithelium. Cells carrying the donor cell-derived Y chromosome can be detected that are also positive for cytokeratin indicating their epithelial nature. It is encouraging for respiratory tract therapy that cells carrying an epithelial marker can be detected in the trachea and may provide a potential therapeutic avenue for diseases such as cystic fibrosis.

We are indebted to Cathy Simpson and Shonna Maccall for carrying out the FACS analysis. We would like to thank Paul Perry and Mike Miller for their assistance with microscopy and Sandy Bruce for helpin preparing the manuscript. We are grateful to the MRC and CF Trust UK for funding, and to N. D. Hastie for his continued support and enthusiasm for this project.