Cells Capable of Bone Production Engraft from Whole Bone Marrow Transplants in Nonablated Mice

6–8-wk-old BALB/c H-2D mice were purchased from Charles River Labs. and were housed in a conventionally clean facility for at least 1 wk. All mice received mouse chow and acidified water ad libitum. The BALB/c H-2D strain was chosen because of its negligible immunoreactivity to the H-Y (Y chromosome– associated histocompatibility) antigen (12, 13).

Marrow was collected by flushing femurs and tibiae with cold PBS supplemented with 5% heat-inactivated FCS (Hyclone Labs.). The cells were filtered through a 40-μm filter (Becton Dickinson), washed, and resuspended for injection in 0.4 ml of PBS per recipient.

Transplants involved a single or three daily intravenous infusions into nonablated female recipients of the same age. After 6 wk or 6 mo, the presence of donor cells in the marrow was quantitated by Southern blot analysis and the spatial distribution of engrafted cells in the bone tissue was analyzed on whole femoral marrow sections using FISH.

Marrow DNA extracts were prepared, digested, and separated by gel electrophoresis, and analyzed by Southern blot as previously described (14). The presence of Y chromosome–specific DNA sequences was assessed using a pY2-cDNA probe (donated by I. Lemischka, Princeton University, Princeton, NJ) (15). Sample loading variability was assessed and adjusted by reprobing membranes with IL-3 cDNA (donated by J. Ihle, DNAX, Palo Alto, CA). Probes were labeled with ³²P. The percentage of male DNA was quantified by PhosphorImager (Molecular Dynamics) analysis.

In situ detection of individual cells using FISH on sections of paraffin-embedded whole murine femurs was done as previously described (16). In brief, mice were perfused with 4% paraformaldehyde at physiological pressure into the descending aorta. 5-μm sections were permeabilized using proteinase K enzymatic digestion. Sections were hybridized with a digoxigenin-labeled Y chromosome–specific painting probe (17) and stained using antidigoxigenin rhodamine-Fab fragments. Levels of autofluorescence were determined by counter-staining in 0.2 μmol DAPI (4,6-diamidino-2-phenylindole) (see Fig. 1 d). Sections were mounted in Vectashield (Vector Labs.).

To accurately assess the extent of donor chimerism and the histological location of engrafted cells within the bone tissue, cells of donor origin were detected by FISH, using a Y chromosome–specific painting probe, in whole femoral sections.

6 wk after a marrow transplant, cells of donor origin were detected as osteocytes encapsulated in lacunae (Figs. 1 and 3 A). As osteocytes are terminally differentiated bone cells within the bone matrix itself, these cells could not have been transplanted during the original infusion but must be the progeny of engrafted, functionally active donor cells. In addition, donor cells were evident both as flattened lining cells on the periosteal bone surface and as cells within the endosteal growth region (Fig. 3, D and G). It is also important to note that because whole marrow was transplanted, a significant proportion of marrow cells in the recipients were also of donor origin (refer to the legends of Figs. 1 and 3 for exact proportions).

A technical limitation of FISH is the removal of the actual bone matrix during the denaturation process. As a consequence, cells at the endosteal marrow surface were not scored, avoiding any possible confusion between positive bone marrow cells and positive bone cells once the bone matrix was removed. To confirm that the cells analyzed were both of donor and bone origin, serial, longitudinal femoral sections were cut and every second section was stained by FISH and every alternate section was stained with hematoxylin and eosin (H and E). Individual cells could then be identified in the bone lacunae, periosteum, or endosteal growth region on the H and E sections (Fig. 3, C, F, and I). Due to the thickness of each section (5 μm), a significant number of cells were present in both the FISH and serial H and E sections. For ease of identification, these cells are pictorially represented in each figure (Fig. 3, B, E, and H).

Bone cells of donor origin were easily detected both 6 wk and 6 mo after transplant, and no difference in the incidence of these cells was observed at either of these time points. In all sections analyzed after a transplant of 120 –180 × 10⁶ marrow cells (resulting in 19.2 ± 2.9 to 31.3 ± 2.5% marrow chimerism), at least one positive cell was detected. In greater than two-thirds of the sections analyzed (n = between 2 and 9 sections for 3 and 2 individual mice at 6 wk and 6 mo, respectively), three or more positive cells were detected. The maximum number of positive cells observed in the bone matrix and the periosteal region of any individual, centrally cut longitudinal section was 8, of which a significant number were in the matrix itself. With a section being 5 μm thick, and the central region of a femur from a 12-wk-old mouse ∼2-mm-thick, there is a conservative estimate of up to 3,200 of these positive cells per femur. A section contained 2,870 ± 114 (n = 8) osteocytes, and therefore there was a maximum of 3 donor cells per 1,000 osteocytes.

Even up to the age of 6 mo, a mouse grows significantly. Between the ages of 6 and 12 wk, the period of the 6-wk transplant, this growth resulted in a significant 16.5% increase in body weight (18.3 ± 0.2 g and 21.9 ± 0.4 g, SEM, respectively; P < 0.001 using a Wilcoxon rank-sum test, n = 15 and 13 respectively). During this period of time, bone growth, or specifically bone remodelling by resorption and formation, resulted in a 30.3% increase in femoral mass (30.4 and 43.6 mg, respectively), as well as a significant increase in size (Fig. 4). Consequentially, by 12 wk of age there is essentially complete turnover of the bone present at 6 wk of age, allowing osteocytes of donor origin to be found anywhere within the bone matrix. In addition, mice transplanted at 6 wk of age and analyzed at 6 mo of age would have many complete turnovers of bone matrix during the transplant period. This suggests that a cell of donor origin detected as an osteocyte in the bone matrix 6 wk after transplant is extremely unlikely to be present 6 mo after transplant. Cells present in the bone matrix 6 mo after transplant are most likely to have been derived from cells that were not actively producing bone 6 wk after transplant; instead, the donor osteocytes detected 6 mo after transplant probably arose from an originally transplanted osteoprogenitor.

Our data demonstrate that whole marrow contains cells of the bone lineage that can engraft post-intravenous infusion, form bone, and give rise to osteocytes and bone lining cells. These are easily detectable in the femur 6 wk and 6 mo after transplant into nonablated mice. This occurs long-term in vivo, without any in vitro culturing manipulation. The long-term stability in the incidence of bone cells of donor origin is consistent with the stable bone marrow chimerism that we have previously observed over this period of time (18).

The data support the suggestion by Pereira et al. (19) and Prockop (20) that transplanted cells with bone forming capabilities engraft in the marrow of recipient mice and participate in normal biological turnover, in which the marrow provides a continual source of osteoblastic progenitor cells. The data also supports the hypothesis proposed by Horwitz et al. (21) that whole marrow contains enough of these cells to be able to replace sufficient osteoblasts to be clinically useful in diseases such as osteogenesis imperfecta, where marrow ablation is standard preparative treatment. Other strategies for the therapeutic use of putative “mesenchymal stem cells” have involved the direct injection of isolated and culture-expanded bone marrow cells into bone lesions or joints. These techniques were shown to promote lesion repair and joint resurfacing at the site of injection but did not have systemic effects (7, 22).

Marrow stromal cells produce bone matrix when placed in diffusion chambers and implanted into the peritoneum (9, 10). Other studies have shown that fibroblast colonies, or clones originating from single stromal cells (CFU-f  ), have high osteogenic potencies (21) and can form bone (23). However, stromal cells are not repopulated in radiochimeras reconstituted by marrow transplantation (24–28). If marrow fibroblasts are to be used for clinical therapies, transplanting them so that they engraft and are active may be a major problem (29). Piersma et al. (30) have subsequently demonstrated that CFU-f could be repopulated in an irradiated host when infused intravenously as part of a whole marrow transplant, although the exact population being assayed remains controversial. Keating et al. (31) showed that the hemopoietic stromal microenvironment, defined as the adherent layer of human long-term marrow cultures grown from patients who received normal marrow transplants, progressively became donor with time after transplant. However, Simmons et al. (32) and Lennon and Micklem (33) more recently showed that the stromal fibroblastic cells in marrow cultures made from human allogeneic transplant recipients and long-term murine chimeras were mainly or entirely host derived. All these studies were done in irradiated recipients, involved in vitro manipulation in order to discriminate donor versus host, and did not go on to show functional feed-out of transplanted donor cells in vivo. Therefore, it remains to be determined whether stroma, which has been considered non-transplantable via intravenous infusion, lodges within the marrow when part of a whole bone marrow transplant.

Recently, Pereira et al. (19) showed that 1–5 mo after a transplant into irradiated hosts of “mesenchymal cells” enriched by adherence to plastic, expanded in culture, and marked with the human COL1A1 minigene, there was widespread tissue distribution of connective tissue cells containing and expressing the gene. Along with many other tissues, such as brain, spleen, marrow, and lung, the study showed positive signal for and expression of the gene by PCR and reverse transcriptase PCR from ground bone. However, these positive results could be due to many different cell types, and they do not definitively prove the presence of donor bone cells. In the studies of Pereira et al. (19), the irradiation of recipient mice was necessary to ensure engraftment, and hence limits clinical application of this strategy. Our data demonstrate that the need for isolation and purification of osteoblastic cells can be circumvented, and preparative marrow ablation treatments are not required. In addition, the data show conclusively that not only do these cells successfully engraft when included as part of a normal marrow transplant, but they are then functionally active, as only osteoblasts, which are actively producing bone, can become encapsulated within it and differentiate into osteocytes.

Previous studies have shown that marrow chimerism can be achieved in mice without the need for marrow ablation (14, 18). This study provides the first clear demonstration of the in vivo derivation of osteocytes from a whole marrow transplant into nonablated recipients. With the development of new methodologies and transgenic mouse models, future studies will reveal further critical information at the individual cell level, as well as possibilities in the treatment of bone disorders by bone marrow transplantation.