While some studies have suggested that hematopoietic stem cells might give rise to other tissue types, others indicate that transdifferentiation would have to be an extremely rare event. We have now exploited smooth muscle type α-actin (αSMA) promoter– driven green fluorescent protein (GFP) transgenic mice (αSMA-GFP mice) for bone marrow transplantation to evaluate their potential to generate donor-type tissues in irradiation chimeras. There was a highly restricted pattern of GFP expression in the transgenic mice, marking bone marrow stromal cells and mesangial cells in the kidney. However, these characteristics were not transferable to wild-type animals given transgenic marrow cells even though hematopoietic cells were largely replaced. Our findings support earlier studies suggesting that the bone marrow microenvironment is difficult to transplant and indicate that hematopoietic stem cells are unlikely to give rise to αSMA-expressing progeny.
Bone marrow contains hematopoietic stem cells (HSCs) and mesenchymal stem cells (MSCs) [1, 2]. Bone marrow cells are capable of differentiating into various lineages including hepatocytes [3–5], vascular endothelial cells , biliary epithelial cells , adipocytes , cardiomyocytes , neuronal cells , and skeletal muscle cells [10, 11]. This has raised the hope of developing new stem cell strategies for tissue repair and regeneration.
Recent studies with nonspecific green fluorescent protein (GFP) transgenic mice have also suggested that bone marrow HSCs may give rise into other tissue/cell types, including vascular smooth muscle cells  and glomerular mesangial cells . However, many other studies argue against HSC plasticity [14–19]. Cell fusion is the central issue in this controversy . Macrophages have an ability to fuse among themselves and with other cell types  and are the major participants in cell fusion [18, 22]. Because any tissue injury results in extensive recruitment of inflammatory cells, the evidence for stem cell plasticity is often confounded by the results of cell fusion. There is another technical issue in bone marrow transplantation. Bone marrow stromal cells, including MSCs [1, 23], are generally not replaced with systemic bone marrow transplantation [24, 25]. To address these issues, the specific identification of the cells of interest and minimizing the potential of false-positive results by cell fusion are crucial.
Recently, we have developed a new transgenic smooth muscle type α-actin (αSMA)-GFP mouse that carries a GFP gene coupled with αSMA regulatory sequence . By utilizing this αSMA-GFP mouse as a bone marrow donor, the smooth muscle cells originating from transplanted bone marrow cells can be specifically identified by the expression of GFP. Here, we report that bone marrow of αSMA-GFP mouse contains GFP+ (αSMA-expressing) cells. However, a progenitor for αSMA-expressing cells was not transplantable in systemic bone marrow chimeras even when hematopoiesis was successfully reconstituted by transplanted stem cells. Developmental plasticity of HSCs seems to be strictly limited, and the transdifferentiation of HSCs into mesenchymal cells unlikely occurs in vivo.
Materials and Methods
αSMA-GFP and Tie2-GFP Mice
The αSMA-GFP mice (C57BL6) that express GFP under the control of αSMA promoter were produced in the Transgenic Mice Facility of the National Eye Institute, NIH. The αSMA promoter was provided by Dr. James A. Fagin of Ohio State University (see Wang et al.  for details). The regulatory sequence of αSMA gene used contains −1074 bp of 5′ flanking region, the transcription start site, 48 bp of exon 1, the 2.5-kb intron 1, and the 15-bp exon 2 of mouse αSMA. GFP is specifically expressed in both vascular and nonvascular smooth muscle cells [27, 28].
Tie2-GFP mice that express GFP under endothelial-specific receptor tyrosine kinase (Tek, also called Tie-2) promoter were first described by Motoike et al. . The mice used in this study were obtained through Jackson Laboratory (Bar Harbor, ME, http://www.jax.org) (stock No. 003658). Because the mouse strain FVB/N, the original background of Tie2-GFP mouse, carries the rd mutant gene that causes retinal degeneration , the mice were backcrossed to C57BL6. The removal of rd mutant gene was confirmed by polymerase chain reaction (PCR) amplification followed by digestion with the restriction enzyme Dde 1 that recognizes the mutant sequences .
All experimental procedures involving animals were reviewed by the Institutional Animal Care and Use Committee of University of Oklahoma Health Sciences Center, and animals were cared for in accordance with its guidelines.
Bone Marrow Transplantation
Bone marrow transplantation was performed with αSMA-GFP and Tie2-GFP mice as donors and Rag2−/− mice (T- and B-cell deficient , C57BL6 background) as recipients. Immediately after the mice were killed by carbon dioxide suffocation, bone marrow cells were flushed from femurs and tibiae, pooled, and washed twice with Ca2+-, Mg2+-free phosphate-buffered saline (PBS). The recipient mice received a lethal dose (950 rads) of total body irradiation from a 137Cs source. After 5 to 6 hours of irradiation, 107 bone marrow cells were injected to the recipient mice through the tail vein, as described [13, 33]. Unless specifically mentioned, all experiments in this study were conducted at 6 months after the bone marrow transplantation was performed.
Vascular Smooth Muscle Cells
Aortas were obtained from αSMA-GFP mice immediately after they were euthanized with carbon dioxide suffocation. The aortas were carefully dissected, and all contaminating tissues were removed under a microscope. The obtained aortas were minced into small pieces with two blades and cultured in Dulbecco's modified Eagle's medium (DMEM) supplemented with 10% fetal calf serum (FCS) (Biofluids Inc., Rockville, MD, http://www.biofluids.com) in collagen-coated culture plates (Becton Dickinson Labware, Bedford, MA, http://www.bd.com) under a 5% carbon dioxide atmosphere. The cells were grown from the pieces of aorta attached to the culture plate within a few days and expanded at least up to four to five passages. In this study, the cells were used at four passages. The identity of cells was confirmed by detecting the expression of GFP and immunostaining with antibody against αSMA (Sigma-Aldrich, St. Louis, http://www.sigmaaldrich.com). Retinal capillary pericytes were cultured as previously reported .
Isolation of Renal Glomeruli and Mesangial Cell Cultures
Immediately after the mice were killed, kidneys were dissected into cortex and medulla regions. The cortex was minced into small pieces with blades and passed through a 50-mesh stainless steel sieve (sieve opening, 300 μm; VWR International, Bristol, CT, http://www.vwr.com) and a 100-μm nylon cell strainer (BD Biosciences, Franklin Lakes, NJ, http://www.bdbiosciences.com). In this differential sieving, the intact glomeruli retained on the 100-μm nylon cell strainer, whereas many small-tissue debris/fragments went through the strainer. The glomeruli were washed and resuspended in PBS. The purity of glomeruli was confirmed by microscopy.
To culture mesangial cells, the glomeruli were incubated with 750 U/ml collagenase (type IA) (Sigma-Aldrich) in PBS at 37°C for 25 minutes. After being washed with PBS, the digested glomeruli were cultured in DMEM media containing 10% FCS. The identity of cells was confirmed by immunostaining with antibody against αSMA.
Antibodies and Cell Sorting
The antibodies phycoerythrin (PE)-conjugated Flk-1 and Sca-1 (Ly6A/E; D7) monoclonal antibodies (mAbs) and allophycocyanin (APC)-conjugated CD44, Mac-1, and c-kit (2B8) mABs were all purchased from BD PharMingen (San Diego, http://www.bdbiosciences.com/pharmingen). PE-conjugated anti-Tie2 (TEK) was purchased from eBioscience (San Diego, http://www.ebioscience.com). The stained cells with PE- or APC-conjugated antibodies were subjected to sorting on a MoFlo (DakoCytomation, Fort Collins, CO, http://www.dakocytomation.us). Flow cytometric analysis was performed with a FAC-Scalibur and the Cellquest program (Becton Dickinson Labware).
Isolation of Bone Marrow Hematopoietic Stem Cells and Peripheral Mononuclear Cells
Bone marrow cells collected from mice were suspended with Hanks' medium supplemented with 3% FCS. Cells were incubated with antibodies to lineage markers (Gr-1 and Mac-1 for myeloid cells, anti-CD19 and anti-CD45RA for B lineage cells, and TER-119 for erythroid cells), followed by incubation with goat anti-rat immunoglobulin G-coated magnetic beads (Miltenyi Biotec, Bergisch Gladbach, Germany, http://www.miltenyibiotec.com). Cells attached to beads were removed with a magnetic separator using negative selection columns. The lineage-depleted bone marrow cells were then incubated with a mixture of labeled antibodies to the lineage markers (fluorescein isothiocyanate-conjugated anti-CD3, anti-CD8, anti–Gr-1, anti–Mac-1, anti–TER-119, anti-CD2, and anti-CD45R) and APC-conjugated anti–c-kit antibody and PE-conjugated anti–Sca-1 antibody to sort lineage marker–negative (Lin−) c-kit+ and Sca-1+ populations. In this study, Lin− Sca-1+ c-kit+ cells were used as HSC-enriched cells.
Peripheral mononuclear cells were separated from blood by centrifugation on Ficoll/Hypaque (lymphocyte separation medium; Cellgro, Mediatech, Herndon, VA, http://www.cellgro.com).
DNA Preparation and Polymerase Chain Reaction
DNA isolation was performed with a QIAshredder (Qiagen, Valencia, CA, http://www1.qiagen.com) and a DNA isolation kit (QIAprep mini spin kit, Qiagen). All experiments were conducted according to the manufacturer's instructions.
PCR was conducted using PuReTaq Ready-To-Go PCR Beads (0.2 ml tubes/plate, Amersham Biosciences, Piscataway, NJ, http://www.amersham.com) on a GeneAmp PCR System 9700 (Applied Biosystems, Foster City, CA, http://www.appliedbiosystems.com). Approximately 50 ng of isolated genomic DNA was amplified in a total volume of 25 μl with PCR beads that contained 2.5 units of Tag DNA polymerase, 10 mM Tris-HCl (pH 9.0), 50 mM KCl, 1.5 mM MgCl2, 200 μ M of dATP, dCTP, dGTP, and dTTP, and RNase/DNase-free bovine serum albumin. The amplification was performed by a total of 32 cycles of denaturation at 95°C for 1 minute, annealing at 55°C for 1 minute, and elongation at 72°C for 30 seconds. The products were visualized by ethidium bromide after electrophoresis on 1% agarose gel. The primers used were AGAACATCATCCCTGCATCC (sense primer) and CTGGGATGGAAATTGTGAGG (antisense primer) for glyceraldehyde-3-phosphate dehydrogenase (GAPDH) (product size, 500 bp) and AAGGGCGAGGAGCTGTTCAC (sense primer) and TTCTGCTGGTAGTGGTCGGC (anti-sense primer) for GFP (product size, 495 bp).
Real-time PCR was conducted on a Smart Cycler II system (Cepheid, Sunnyvale, CA, http://www.cepheid.com) using SYBER Green reverse transcription–PCR reagents (Applied Biosystems). The experiments were performed according to the manufacturer's instructions.
αSMA-GFP Mice Express GFP in Vascular Smooth Muscle Cells Including Renal Glomerular Mesangial Cells
αSMA-GFP transgenic mice carry a GFP gene regulated by the αSMA promoter, and GFP expression is strictly limited to the cells that express αSMA [27, 28]. No hematopoietic cells express GFP. In retina, GFP was detectable only in retinal vasculature (Fig. 1A) because vascular smooth muscle cells and capillary pericytes are the only cells that express αSMA. GFP was also detectable in cultured capillary pericytes (Fig. 1B). Similarly, vascular smooth muscles are the major αSMA-expressing cells in the kidney. In glomeruli, however, mesangial cells are the only cell type that expresses αSMA. Although the αSMA levels in normal mesangial cells are generally low in vivo , GFP was still detectable in glomeruli isolated from αSMA-GFP mice (Fig. 2A). Moreover, in cell culture, mesangial cells all displayed the high levels of GFP expression (Fig. 2B). This indicates that, by using αSMA-GFP mice as bone marrow donors, newly formed smooth muscle cells and glomerular mesangial cells of bone marrow origin can be specifically identified.
Bone Marrow of αSMA-GFP Transgenic Mice Contains a Small Population of GFP+ Cells
Using an isolation method for authentic GFP+ cells, a small GFP+ population (0.02%–0.03% of bone marrow mononuclear cells) could be specifically detected and sorted from bone marrow . No GFP+ cells were detectable in wild-type mice. Those marrow-derived αSMA-GFP+ cells were heterogeneous with respect to Sca-1, c-kit, and Mac-1 expression (Fig. 3). Their surface phenotype was Sca-1− in 82.2%, c-kit− in 70.4%, and Mac-1+ in 55.4%, whereas CD44 was detectable on more than 99% of the cells. Those GFP+ cells were distinct from endothelial lineage cells because two vascular endothelial cell markers, Tie-2 and Flk-1 (VEGFR2, KDR), were negative. The marrow-derived GFP+ cells were also distinct from vascular smooth muscle cells. The cells isolated from aorta of αSMA-GFP mice were brighter in GFP intensity and exclusively negative for Mac-1 expression (Fig. 4). Importantly, similar GFP+ cells to ones in bone marrow were also detectable in peripheral blood of αSMA-GFP mice at very low frequencies (approximately 0.15% of mononuclear cells).
Bone Marrow Stromal Cells Derived From αSMA-GFP Mice Express GFP
Although αSMA is a well-established cell marker for smooth muscle cells and differentiated myofibroblasts , bone marrow stromal cells are also known to express αSMA [38–40]. Stromal cells are adherent and clonogenic in vitro and are suggested to contain a putative smooth muscle progenitor . To confirm the αSMA expression in stromal cells, we examined whether GFP+ cells were detectable in bone marrow cell cultures established from αSMA-GFP mice (Fig. 5).
Bone marrow cells obtained from αSMA-GFP mice were cultured in DMEM media supplemented with 10% FCS that had been widely used for the culture of vascular smooth muscle cells and capillary pericytes . Although no obvious GFP+ cells were observed at the initiation of culture, cells with high GFP intensity became conspicuous among the cells attached to the culture plate within 1 to 2 weeks. When the culture was continued, GFP+ cells gradually dominated among the attached cells. The bone marrow cells were also cultured in Dexter media that can support in long-term the growth of hematopoietic cells as well as stromal cells. GFP expression was detected only in the stromal cells that were strongly attached to culture plates. No hematopoietic cells grown in the same culture showed GFP expression. This indicates that bone marrow contains stromal progenitors that differentiate into cells expressing αSMA.
Because αSMA-GFP mice contain GFP+ bone marrow cells, there is a possibility that this small population of GFP+ cells can be the source for αSMA-expressing (GFP+) stromal cells. To test this possibility, GFP+ cells sorted from bone marrow cells of αSMA-GFP mice were cultured in DMEM media containing 10% FCS. However, no GFP+ stromal cells were detected in culture even after more than 4 weeks. It is also possible that, to differentiate into stromal cells, these cells may require the interaction with some other cell types. To test this possibility, the same GFP+ cells were also cultured together with the whole bone marrow cells from wild-type mice. However, stromal cells grown in this culture were all GFP negative. This suggests that αSMA-expressing (GFP+) cells are unlikely to be the progenitors for αSMA-expressing stromal cells.
αSMA-GFP Chimeric Mice Reconstitute Hematopoiesis With Stem Cells Carrying a αSMA-GFP Transgene
To test the hypothesis that bone marrow cells can be a source of progenitors for vascular smooth muscle cells or glomerular mesangial cells, we performed bone marrow transplantation to create a chimeric model in which marrow cells from αSMA-GFP mice were transferred into wild-type mice (Rag2−/− mice) that received a lethal dose (950 rads) of irradiation. Because neither granulocytes nor monocytes express αSMA, the possible artifact by cell fusion [17, 18] could be minimized by using αSMA-GFP mice as bone marrow donors.
The long-term reconstitution of hematopoiesis by donor HSCs was confirmed by detecting the GFP transgene in bone marrow cells at 6 months after bone marrow transplantation. In PCR analysis (Fig. 6), the GFP transgene was detected in whole bone marrow cells, sorted Lin− c-kit+ Sca-1+ cells, and peripheral mononuclear cells (MNCs) obtained from the recipients of αSMA-GFP marrow transplant. The intensity of PCR product amplified for GFP gene in chimeric mice was comparable to that of αSMA-GFP donor mice. Without receiving bone marrow transplants, no PCR products for the GFP gene were detected in the cells of wild-type mice. To estimate the efficiency of transplantation, the GFP transgene was quantified by real-time PCR. The transplantation efficiency (percent) based on the copy number of GFP transgene adjusted by a house-keeping GAPDH gene was 86.8 ± 11 (mean ± standard deviation, n = 6). This confirms that, at least, HSCs of αSMA-GFP transgenic mice were successfully transplanted and the hematopoietic cells were largely replaced with cells carrying the αSMA-GFP transgene.
To further confirm the replacement of HSCs, we performed bone marrow transplantation using Tie2-GFP mice that express GFP under the control of a Tie-2 promoter. Because HSCs are known to express Tie-2 [43, 44], Tie2-GFP+ cells can be detectable in bone marrow of Tie2-GFP chimeric mice if HSCs are successfully transplanted. In flow cytometric analysis (Fig. 7), Tie2-GFP mice displayed a small population of GFP+ cells. This GFP+ cell population was also clearly detectable in Tei2-GFP chimeric mice, confirming that HSCs are transplantable with the procedure used for αSMA-GFP mice in this study.
αSMA-GFP Chimeric Mice Generate Neither GFP+ Bone Marrow Cells nor GFP+ Stromal Cells
Because reconstitution of hematopoiesis with stem cells carrying the αSMA-GFP gene was confirmed, we examined whether αSMAs expressing bone marrow cells and stromal cells were also reconstituted with the cells carrying the αSMA-GFP transgene in αSMA-GFP chimeric mice.
In flow cytometric analyses, αSMA-GFP mice displayed a small population (0.02%–0.03%) of GFP+ bone marrow cells whereas no GFP+ cells were detected in wild-type mice. In αSMA-GFP chimeric mice, however, there was no significant increase in the GFP+ cell population (Fig. 8). This was true with 20 chimeric mice that successfully received the αSMA-GFP bone marrow transplant. The bone marrow cells were also cultured in DMEM containing 10% FCS using six-well plates, and the GFP expression was examined under the fluorescent microscope continuously twice a week until cells reached 60%–80% confluence (approximately 106 cells per well). However, no single GFP+ cell was detected at any time throughout cell culture (Fig. 9). This means that no single animal among 20 mice showed an increase in GFP+ bone marrow cells and no single GFP+ cell was detected among 2 × 107 stromal cells examined. This suggests that a progenitor for αSMA-expressing (αSMA-GFP+) cells is distinct from that for hematopoietic cells. It is also not transplantable.
No GFP+ Glomerular Mesangial Cells Were Generated in αSMA-GFP Chimeric Mice
A previous study using nonspecific GFP mice suggested that bone marrow HSCs can generate glomerular mesangial cells . If HSCs differentiate into αSMA-expressing mesangial cells, GFP+ cells should be detectable in glomeruli of αSMA-GFP chimeric mice where hematopoiesis was reconstituted with HSCs carrying the αSMA-GFP gene.
Unlike αSMA-GFP donor mice, however, no GFP expression was detected in glomeruli of αSMA-GFP marrow–recipient mice. It is still possible that, since the expression levels of αSMA in normal glomeruli in vivo are generally low, newly formed mesangial cells would not express detectable levels of GFP even though they carried the αSMA-GFP gene. To more rigorously explore whether glomeruli contained mesangial cells with the αSMA-GFP gene, the isolated glomeruli were digested with collagenase and cultured in DMEM media containing 10% FCS because mesangial cells highly express αSMA in vitro  (Fig. 2). If HSCs differentiate into mesangial cells, GFP+ cells should have been detectable among the cells grown from glomeruli even if GFP was undetectable in intact glomeruli. With the procedure described here, approximately 10,000 glomeruli were generally recovered from one kidney. GFP expression was also routinely examined under fluorescent microscope until the cells reached near confluence (approximately 106 cells) in six-well plates. This means that nearly 4 × 105 isolated glomeruli and 4 × 107 cultured mesangial cells from 20 mice that successfully received αSMA-GFP bone marrow transplants were examined. However, no GFP+ cells were detected in either isolated glomeruli or cultured mesangial cells from αSMA-GFP chimeric mice despite the fact that the same cells from αSMA-GFP donor mice all displayed strong GFP signals (Fig. 10). These findings strongly suggest that the competence of marrow HSCs to generate αSMA-expressing cells in the periphery is extremely low. The transdifferentiation of HSCs in vivo, if any, must be an extremely rare event.
Creating chimeric mice by bone marrow transplantation has been widely used to identify cells of marrow origin, but discrimination of donor-type cells is the key step in this approach. The present study uses a new transgenic αSMA-GFP mouse that expresses GFP under control of the αSMA promoter as a bone marrow donor to specifically identify donor-type αSMA-expressing cells. Because GFP expression is limited to cells that express αSMA, GFP+ cells are detectable only when the cells from donor mice differentiate into αSMA-expressing cells. The principal findings are as follows. First, murine bone marrow contains a small population of αSMA-expressing cells, and αSMA is present in cultured marrow stromal cells. Second, the progenitor for these αSMA-expressing cells is distinct from HSCs and is not systemically transplantable. Third, the plasticity of HSCs is strictly limited, and HSCs unlikely differentiate into mesenchymal cells under physiological conditions.
Bone marrow cells have been shown to differentiate into various mesenchymal cells, although the originating progenitors have not been well defined. Nonspecific GFP mice that express GFP under control of the chicken α-actin promoter with a cytomegarovirus enhancer  have been widely used as bone marrow donors to create chimeric mice. GFP+ cells have been detected in choroidal angiogenesis  and in renal glomeruli [13, 33, 47, 48] using this chimeric model. With this evidence, bone marrow cells were suggested to contribute to regeneration of vascular smooth muscle cells in peripheral tissues. However, because all hematopoietic cells were GFP positive, it is difficult to conclusively identify the particular GFP+ cells. Moreover, monocytes/macrophages carrying the GFP protein often fuse to host cells. Mesenchymal stem cells are also known to include stromal cells [25, 49] that are capable of differentiating into αSMA-expressing cells . However, in general, stromal cells are not replaced with systemic bone marrow transplantation [24, 25]. Whether bone marrow contains a transplantable mesenchymal progenitor is still controversial. Our study clearly demonstrated that no αSMA-expressing stromal cells expressed GFP in αSMA-GFP chimeric mice even though hematopoiesis was successfully reconstituted with cells carrying a αSMA-GFP transgene. This indicates that stem or progenitor cells for αSMA-expressing cells and stem/progenitors (HSCs) for hematopoiesis are distinct and that stromal progenitors are not transplantable.
The plasticity of HSCs is another controversial issue . Sata et al.  reported that αSMA-expressing cells were generated from transplanted HSC-enriched Lin− c-kit+ Sca-1+ cells in injured arteries. Masuya et al.  also reported that glomerular mesangial cells were generated from Lin− c-kit+ Sca-1+ CD34− cells from a bone marrow donor even without particular tissue injuries, suggesting that HSCs can contribute to physiological turnover of mesangial cells. Our data directly contradict those studies. In this study, the αSMA-GFP transgene was clearly detectable in both Lin− c-kit+ Sca-1+ bone marrow cells and peripheral mononuclear cells of αSMA-GFP chimeric mice. If HSCs were the source of αSMA-expressing cells, some αSMA-GFP+ cells should have been detectable in either intact glomeruli or cultured mesangial cells. However, not even one GFP+ cell was detected in many mice that successfully received αSMA-GFP marrow transplants. This indicates that transdifferentiation of HSCs to nonhematopoietic cell types is unlikely to occur in vivo. This also indicates the importance of the use of specific mesenchymal markers for the study of mesenchymal stem/progenitors. The infiltration of inflammatory cells is inevitable in many cases and particularly in injured tissues. Even though “false positives” by cell fusion are carefully excluded, the identification of a small population of real GFP+ cells is difficult if many GFP+ inflammatory cells are present.
The present study also demonstrated that bone marrow contains a small population of GFP+ (αSMA-expressing) cells and similar GFP+ cells were also detectable in peripheral mononuclear cell preparations. These cells were negative for the endothelial cell markers Tie-2 and KDR. They were also different from vascular smooth muscle cells with respect to Mac-1 expression and the intensity of GFP signals. However, these cells are unlikely to be progenitors for αSMA-expressing cells. At the very least, they are not the source of αSMA-expressing stromal cells in bone marrow cell culture in vitro.
The stromal stem/progenitor for αSMA-expressing cells is clearly difficult to transplant. In this study, we used Rag2−/− mice as bone marrow recipients to enhance the engraftment of such bone marrow minor populations by combining immunodeficiency of Rag2−/− and lethal doses of irradiation. However, the progenitor for αSMA-expressing cells was not transplantable. Whether the efficiency could be improved to transplant the stromal progenitor is an important question relevant to developing new tissue regeneration therapies. New strategies such as direct infusion into the bone cavity may be needed to effectively engraft such cells [51–53]. αSMA-GFP mice should provide an excellent tool for that type of investigation.
T.Y. and Y.K. contributed equally to this study. Supported by NIH grant AI 20069 (to P.W.K.).
The authors indicate no potential conflicts of interest.