Tissue-Specific Stem Cells Formation of Pancreatic Duct Epithelium from Bone Marrow During Neonatal Development
a
Division of Research Immunology/Bone Marrow Transplantation, bDepartment of Pathology, cCongressman Julian Dixon Cellular Imaging Core, dDivision of Gastroenterology, Hepatology, and Nutrition, and eDevelopmental Biology Program, Childrens Hospital Los Angeles, Los Angeles, California, USA Key Words. Bone marrow transplantation • Epithelial stem cells • Pancreas • Ducts • Neonate • Mice
ABSTRACT Recent reports suggest that bone marrow– derived cells engraft and differentiate into pancreatic tissue at very low frequency after pancreatic injury. All such studies have used adult recipients. The aim of our studies was to investigate the potential of bone marrow to contribute to the exocrine and endocrine components of the pancreas during the normal rapid growth of the organ that occurs during the neonatal period. Five to ten million bone marrow cells from adult, male, transgenic, green fluorescent protein (GFP) mice were injected into neonatal nonobese diabetic/severely compromised immunodeficient/2microglobulin-null mice 24 hours after birth. Two months after bone marrow transplantation, pancreas tissue was analyzed with fluorescence immunohistochemistry and fluorescence in situ hybridization (FISH). Co-staining of GFP, with anticytokeratin antibody, and with FISH for the presence of donor Y chromo-
some indicated that up to 40% of ducts (median 4.6%) contained epithelial cells derived from donor bone marrow. In some of these donor-derived ducts, there were clusters of large and small ducts, all comprised of GFPⴙ epithelium, suggesting that whole branching structures were derived from donor bone marrow. In addition, rare cells that coexpressed GFP and insulin were found within islets. Unlike pancreatic damage models, no bone marrow– derived vascular endothelial cells were found. In contrast to the neonatal recipients, bone marrow transplanted into adult mice rarely generated ductal epithelium or islet cells (p < .05 difference between adult and neonate transplants). These findings demonstrate the existence in bone marrow of pluripotent stem cells or epithelial precursors that can migrate to the pancreas and differentiate into complex organ-specific structures during the neonatal period. STEM CELLS 2006;24:307–314
INTRODUCTION
bone marrow or cord blood transplantation. Although the donor cells express tissue-specific markers, the low frequency of such events has created technical difficulties in proving whether true transdifferentiation is being observed rather than contamination of circulating, mature hematopoietic cells. In addition, in almost all studies, some form of tissue injury is present, raising the possibility that trafficking of inflammatory or endothelial pre-
A debate has raged in recent years over whether stem cells within the bone marrow have the capacity to “transdifferentiate,” i.e., to generate nonhematopoietic cell types [1–3]. Most reports that support this concept have described cells of bone marrow origin present at low frequency in various organs, such as liver [4 –7], lung [8], brain [9, 10], and muscle [11–13], after
Correspondence: Gay M. Crooks, M.D., Division of Research Immunology/BMT, Childrens Hospital Los Angeles, 4650 Sunset Blvd., M.S. #62, Los Angeles, CA 90027, USA. Telephone: 323-669-5690; Fax: 323-906-8193; e-mail:
[email protected] Received February 4, 2005; accepted for publication July 25, 2005; first published online in STEM CELLS EXPRESS August 18, 2005. ©AlphaMed Press 1066-5099/2006/$20.00/0 doi: 10.1634/stemcells.2005-0052
STEM CELLS 2006;24:307–314 www.StemCells.com
Downloaded from www.StemCells.com at University of California on March 9, 2009
XIULI WANG,a SHUNDI GE,a IGNACIO GONZALEZ,b GEORGE MCNAMARA,c C. BARTH ROUNTREE,d KENNY KEZHE XI,a GRACE HUANG,a ANIL BHUSHAN,e GAY M. CROOKSa
Pancreatic Ducts from Bone Marrow
308
MATERIALS
AND
METHODS
5-Bromo-2ⴕ-deoxyuridine Analysis of Cell Cycling in Neonatal Mice To detect the presence of cycling cells in the pancreas, neonatal and adult animals were administered 5-Bromo-2⬘-deoxyuridine (BrdU) (25 mg/kg; Sigma-Aldrich, St. Louis, http://www. sigmaaldrich.com) by i.p. injection. Pancreas tissue was harvested from animals sacrificed 2 hours after BrdU and processed for fluorescence immunohistochemistry. BrdU was detected with BrdU sheep polyclonal antibody conjugated with Cy3TM, and beta cells and ductal epithelial cells were detected with fluorescein isothiocyanate (FITC)–insulin or FITC– cytokeratin (CK), respectively, by using fluorescence immunohistochemistry staining as below.
Bone Marrow Transplantation into Neonatal Mice Adult (10- to 12-week-old) male hemizygous mice transgenic for enhanced green fluorescent protein (GFP) gene (C57BL/ 6-TgN, ACEbEGFP 1Osb/J; The Jackson Laboratory, Bar Harbor, ME, http://www.jax.org) were used as donors for bone marrow transplantation (BMT). Bone marrow cells from femurs and tibia of donors were flushed, washed twice with phosphate-buffered saline (PBS) without Ca and Mg
(Mediatech, Inc., Herndon, VA, http://www.cellgro.com), and frozen down in 10% Cryoserv (Edwards Lifesciences, Irvine, CA, http://www.edwards.com) without further manipulation. Nonobese diabetic/severely compromised immunodeficient/ 2microglobulin-null (Nod/Scid/2Mnull) mice were used as recipients of transplanted bone marrow under a protocol approved by the Institutional Animal Care and Use Committee at Childrens Hospital Los Angeles (CHLA). Five to ten million thawed bone marrow cells from GFP⫹ mice suspended in 50 l PBS were injected into neonatal mice within 24 hours after birth, through the superficial temporal vein as described [18]. No irradiation or other conditioning regimen was given. Animals were housed in sterile conditions and weaned at 4 weeks of age. Because gender typing of neonatal mice could not be performed at the time of transplant, both male and female recipients were transplanted and analyzed. Recipient mice were sacrificed 2 months post transplantation for analysis. Engraftment of donor-derived cells in peripheral blood was confirmed at the time of sacrifice by identifying GFP⫹ cells by flow cytometry (FACSCalibur; BD Biosciences–Immunocytometry Systems, San Jose, CA, http://www.bdbiosciences.com/ immunocytometry_systems). In separate experiments, adult Nod/Scid/2Mnull mice (8 to 10 weeks old) were transplanted via tail vein injection with 5–10 ⫻ 106 adult GFP⫹ bone marrow cells without irradiation and were sacrificed for analysis 2 months post transplantation.
Histochemistry After sacrifice, pancreas tissue was dissected and fixed in 10% neutral buffered formalin (Richard-Allan Scientific, Kalamazoo, MI, http://www.rallansci.com) for 6 hours, embedded in paraffin (Leica, Nussloch, Germany, http://www.leica.com), sectioned using a microtome (Leica). Slides of 5-m thickness were dewaxed with 100% Toluene (Sigma-Aldrich), and rehydrated. Antigen unmasking was performed with Vector unmasking buffer (Vector Laboratories, Inc., Burlingame, CA, http:// www.vectorlabs.com) for 12 minutes. Nonspecific binding was blocked with 250 –300 l 100 mM Tris-Buffered Saline (TBS) (pH 7.5) containing 0.1% Tween-20, 3% bovine serum albumin, and 5% normal donkey serum (Immunoglobulin G–free, Protease-free; Jackson ImmunoResearch Laboratories, Inc., West Grove, PA, http://www.jacksonimmuno.com) for at least 30 minutes until primary antibody was added. Pancreas tissues from nontransplanted Nod/Scid/2Mnull mice and GFP transgenic mice were used as negative and positive controls for GFP staining, respectively. The following anti-mouse primary antibodies were used for fluorescence immunohistochemistry: BrdU sheep polyclonal antibody conjugated with Cy3TM (Abcam Inc., Cambridge, MA, http://www.abcam.com), platelet/endothelial cell adhesion molecule-1 (PECAM-1) (CD31) goat polyclonal antibody (Santa Cruz Biotechnology Inc., Santa Cruz, CA, http://www.scbt. com); GFP rabbit polyclonal antibody (Novus Biologicals, Inc., Littleton, CO, http://www.novus-biologicals.com), insulin guinea pig polyclonal antibody (DakoCytomation Inc., Carpinteria, CA, http://www.dakocytomation.com), pan-CK mouse monoclonal antibody (Sigma), and CD45 rat monoclonal antibody (SouthernBiotech, Birmingham, AL, http://www. southernbiotech.com). Secondary antibodies used in the study
Downloaded from www.StemCells.com at University of California on March 9, 2009
cursors (with or without cell fusion) is responsible for at least some of the observations. With these caveats accepted, the obvious possible therapeutic implications necessitate the further careful study of bone marrow as a source of tissue repair and regeneration. Reports of bone marrow differentiation into pancreatic tissue have been few and contradictory. Some studies have shown rare cells of donor bone marrow origin that coexpress insulin lodged within the pancreatic islets, raising the possibility that bone marrow might be used to regenerate beta cells in the treatment of type 1 diabetes [14, 15]. Other studies have found little if any evidence of beta-cell differentiation [16, 17]. In all previous studies, whether negative or positive, some degree of pancreatic injury was either induced specifically (using the beta-cell toxin streptazotocin) [15–17] and/or nonspecifically and subclinically by total body irradiation prior to the bone marrow transplant [14 –17]. All experiments were performed with adult donors and recipients. In contrast to previous reports, the current study explored the potential of bone marrow to contribute to the development of the exocrine and endocrine components of the pancreas in the neonatal period during which rapid tissue growth and differentiation occur in the absence of injury. Data presented here show that bone marrow transplanted into immune-deficient neonatal mice contributed to a high percentage of epithelial cells within the pancreas, in some cases forming entire branching ductal structures. Bone marrow transplanted into adult recipients did not generate pancreatic ducts, indicating that factors present specifically during neonatal development are responsible for the recruitment of bone marrow cells with epithelial potential to areas of rapid epithelial growth. These data have important implications for the study of epithelial differentiation in the developing animal and support the existence of significant epithelial potential in the bone marrow that might be harnessed for clinical use.
Wang, Ge, Gonzalez et al.
Fluorescence In Situ Hybridization In cases in which female recipients were used, fluorescence in situ hybridization (FISH) analysis was performed to identify donor Y chromosomes in tissue. For dual staining of CK and mouse Y-paint chromosome, slides were processed with antigen unmasking, as described above, followed by digestion with 0.16% trypsin in diluent (Zymed Laboratories, San Francisco, http://www.zymed.com) for 10 minutes at 37°C and washing with TBST. After blocking, sections were incubated with antibody against mouse CK, followed by incubation with biotinylated anti-mouse antibody, and visualized by fluorescein (dichlorotriazinyl aminofluorescein)– conjugated streptavidin (Jackson ImmunoResearch Laboratories, Inc). Post fixation in 4% paraformaldehyde, slides were sequentially dehydrated in 70%, 90%, and 100% alcohol. One microliter of Y-paint probe (Cy3TM) in 9 l hybridization buffer (Cambio Ltd., Cambridge, U.K., http://www.cambio. co.uk) was applied on a 22 ⫻ 22 mm coverslip. Slides were sealed with rubber cement, denatured in 60°C for 10 minutes, and hybridized overnight at 37°C in humidified container. After washing, the slides were mounted with Vectashield medium containing DAPI. Fluorescein and Cy3 filters were used to reveal CK and Y chromosomes.
Quantification of Donor-Derived Ducts in Recipient Pancreas Six to 24 sections of pancreas tissue from each mouse were analyzed by fluorescence microscopy. These sections were derived from the 10th, 20th, 30th, 40th, 50th, 60th, and 70th 5-m sections that spanned the length of the pancreatic tissue, thus providing representative analysis of the entire organ. Ducts were defined in each section based on CK staining of definite duct structures. GFP⫹ ducts were defined by coexpression of GFP and CK in at least two cells in each clearly defined ductal structure. www.StemCells.com
Statistical Analysis
Means ⫾ SEM of BrdU⫹ cells obtained from adult and neonatal mice were determined, and a two-sample t-test was performed. Medians and confidence intervals of GFP⫹ ducts from recipient animals were given, as these data were nonparametric. Comparison between the neonatal and adult recipients was performed with a Mann-Whitney test. p ⬍ .05 was considered significant (two-tailed).
RESULTS Neonatal Pancreas Contains a High Frequency of Cycling Epithelial Ductal and Beta Cells We hypothesized that the neonatal model may represent a unique scenario in which pancreatic epithelium undergoes rapid cell division without injury, thus providing signals that assist in recruitment of epithelial precursors from the bone marrow. To explore this possibility, BrdU labeling in the pancreas was analyzed 2 hours after administration to otherwise untreated neonatal and adult Nod/Scid/2Mnull mice. Pancreas tissue from neonatal mice displayed irregularly shaped islets and very small ducts compared with their adult counterparts. BrdU⫹ cells could be seen in both exocrine and endocrine compartments of the pancreas and were significantly more frequent overall in the neonatal than in adult pancreas (37.1 ⫾ 1.9 versus 2.7 ⫾ 0.3 per microscope field, magnification 20⫻, p ⬍ .0001). When ducts and beta cells were specifically analyzed by BrdU staining, a low background level of cell cycling was seen in both cell types from adult mice (6.8% ⫾ 1.4% of ducts and 3.0% ⫾ 1.0% of islets contained at least one BrdU⫹ cell). The frequency of cycling was significantly higher in the neonatal pancreas in both ductal epithelium (48.1% ⫾ 5.5% of ducts contained BrdU⫹ cells) and beta cells (26.8% ⫾ 2.2% of islets contained BrdU⫹ cells) than in the adult pancreas (p ⬍ .0001) (Figs. 1A–1D). Because epithelial tissue in the neonatal pancreas is rapidly dividing, we reasoned that this scenario may provide a unique stimulus required to recruit epithelial precursors from bone marrow to participate in epithelial growth and differentiation.
Hematopoietic Engraftment After Bone Marrow Transplantation into Nonirradiated Neonatal Recipients
Hematopoietic engraftment of GFP⫹ donor bone marrow was determined by analyzing GFP expression in peripheral blood at the time of sacrifice of recipient mice (2 months post transplantation) (Fig. 2A). Bone marrow and peripheral blood in donor GFP transgenic mice did not uniformly express the GFP marker (39.0% ⫾ 0.7% of nucleated cells expressed GFP in bone marrow and 88.6% ⫾ 2.1% in peripheral blood) (data not shown). Nontransplanted mice had no background GFP staining in peripheral blood (Fig. 2B). Engraftment of donor cells was obtained in all transplanted mice, with 22.3% ⫾ 4.9% GFP⫹ cells observed in the peripheral blood (n ⫽ 9 mice) (Fig. 2C). Consistent with the fluorescence-activated cell sorting data, hematopoietic cells of donor origin were also occasionally observed in the blood vessels of recipient pancreas, using fluorescence immunohistochemistry staining (Fig. 2D).
Downloaded from www.StemCells.com at University of California on March 9, 2009
were anti-goat–FITC, anti-rabbit–Cy3, and anti-mouse–FITC (Jackson ImmunoResearch Laboratories, Inc.). Slides were mounted with Vectashield medium with 4⬘6-diamidino-2-phenylindole (DAPI) (Vector Laboratories, Inc.) after washing three times with TBS containing 0.1% Tween-20 (TBST). Images were viewed with a Leica DMRXA microscope (Bannockburn, IL) using a Plan Apo 20 or 40⫻/1.25 NA phase 3 DIC or Plan Apo 63⫻/1.32 oil immersion objective lens. Filter sets used were DAPI, chroma 31000; fluorescein, 41001; and Cy3, 41007a (Chroma Technology Corp., Rockingham, VT, http:// www.chroma.com). An LS300W ozone-free xenon arc lamp (Sutter Instrument Co., Novato, CA, http://www.sutter.com) was coupled to the microscope with a liquid light guide. Images were acquired from EasyFISH software with a SkyVision–2/ VDS-1300 12-bit digital camera (Applied Spectral Imaging Ltd., Migdal Ha’emek, Israel, http://www.spectral-imaging. com) and printed using Microsoft PowerPoint (Redmond, WA, http://microsoft.com). Hematoxylin and eosin staining of skin and liver from transplanted and nontransplanted mice and analysis by a blinded pathologist were performed to rule out the presence of graftversus-host disease (GVHD).
309
310
Pancreatic Ducts from Bone Marrow
Bone Marrow Cells Contribute to Pancreatic Duct Formation When Transplanted into Neonatal Mice To investigate the capacity of bone marrow cells to contribute to pancreatic tissue during normal neonatal development, GFP⫹ bone marrow cells were intravenously infused into mice 24 hours after birth without inducing pancreas injury. Pancreas tissues from all animals were analyzed at 2 months post transplantation. Numerous hollow structures lined with uniform, low columnar, tightly spaced cells were partially or completely derived from donor bone marrow cells based on GFP expression. The identity of the structures was determined using the epithelial marker CK and the vascular endothelial marker PECAM-1; no cross-reaction between the antibodies was seen, and pancreatic ducts were clearly distinguished from blood vessels in both transplanted and nontransplanted mice (data not shown). Bone marrow contribution to epithelial ducts was analyzed in a total of 165 pancreas sections from nine mice (in different experiments) sacrificed 2 months after neonatal BMT. Among 4,034 ductal structures identified on the basis of CK staining, a median of 4.6% (range 0.9%– 43.1%) of ducts contained cells
Figure 1. Cell cycling of ductal epithelium and beta cells is increased during the neonatal period. BrdU expression (shown as red nuclei) in (A) ductal cells (CK shown as green) and (B) beta cells (insulin shown as green) of pancreas tissue from neonatal (1-day-old) and adult (8-
week-old) mice. DAPI (blue) stains all nuclei. Scale bars ⫽ 50 m. The percentage of ducts (C) and islets (D) containing cycling cells (i.e., at least one BrdU⫹ cell) was significantly increased in neonates compared with adults (*p ⬍ .0001 with two-sample t-test). Means and SEM are shown. Abbreviations: BrdU, 5-Bromo-2⬘-deoxyuridine; CK, cytokeratin; DAPI, 4⬘6-diamidino-2-phenylindole.
Downloaded from www.StemCells.com at University of California on March 9, 2009
Figure 2. Hematopoietic engraftment after neonatal BMT. (A): Whole BM cells from adult male GFP transgenic mice were injected intravenously into newborn Nod/Scid/2Mnull mice 24 hours after birth, and recipient mice were sacrificed 2 months post BMT for analysis. Detection of GFP⫹ donor cells in PB by FACS analysis of (B) nontransplanted Nod/Scid/ 2Mnull mouse, and (C) Nod/Scid/ 2Mnull mouse transplanted with GFP⫹ BM. (D): Fluorescence immunohistochemistry staining of pancreas from transplanted animal demonstrating GFP⫹ blood cells (red) inside vessel stained with PECAM (green). DAPI (blue) stains all nuclei. Scale bars ⫽ 50 m. Abbreviations: BM, bone marrow; BMT, bone marrow transplantation; CK, cytokeratin; DAPI, 4⬘6-diamidino-2-phenylindole; FACS, fluorescence-activated cell sorting; FISH, fluorescence in situ hybridization; GFP, green fluorescent protein; IHC, immunohistochemistry; Nod/Scid/2Mnull, nonobese diabetic/severely compromised immunodeficient/2microglobulin-null; PB, peripheral blood; PECAM, platelet/endothelial cell adhesion molecule.
Wang, Ge, Gonzalez et al.
ence of male donor cells contributing to the epithelial lining of the pancreatic ducts (Fig. 5). Analysis of serial sections taken along the length of the pancreas showed that the bone marrow– derived ducts tended to be found toward the center of the organ. These clusters of predominantly GFP⫹ ducts represented interlobular ducts (large ducts typically surrounded by loose mesenchymal tissue); in contrast, the small intralobular ducts surrounded by acinar tissue rarely contained GFP⫹ epithelial cells. In some areas, approximately 50% of the interlobular ducts were GFP⫹. To rule out the presence of contaminating hematopoietic cells, CD45 staining was applied to the sections from transplanted and nontransplanted mouse pancreas. No duct epithelial cells coexpressed CD45 (data not shown). Of note, no GFP-expressing cells were found in 1,300 blood vessels (identified by expression of PECAM-1) in any of the nine animals transplanted during the neonatal period. Thus, bone marrow cells did not contribute to the formation of pancreatic blood vessels in the neonatal bone marrow transplant model.
Bone Marrow–Derived Cells Occurred at Low Frequency in Islets Co-staining of pancreatic tissue sections showed occasional bone marrow– derived cells within the islets that expressed both GFP and insulin (data not shown). Among a total 438 islets identified in pancreas sections from five mice analyzed 2 months after BMT, 2.5% of islets appeared to contain one or more donor-derived cells coexpressing GFP and insulin. However, only rare cells (1.3%) within each positive islet expressed GFP (Table 2). To further confirm the donor origin of these GFP⫹insulin⫹ cells, pancreas sections stained with GFP/insulin were processed for FISH staining and found to have single X and Y chromosome FISH signals, demonstrating diploidy (data not shown).
Ductal Generation from Bone Marrow Is Rare After Transplantation Performed in Adult Recipients
Figure 3. Formation of pancreatic ducts derived from GFP⫹ donor bone marrow. Fluorescence immunohistochemistry of pancreas tissue from transplanted animals with CK (green), GFP (red) antibodies, and DAPI (blue) (A–C). Different patterns of donor engraftment are shown. (A): Clusters of ducts, all of which contain 100% of epithelium of donor origin. (B): Ducts derived completely from donor cells are seen adjacent to recipient-derived ducts. (C): Less than 50% of cells lining duct are of donor origin. (D): Pancreatic ducts from GFP transgenic donor (positive control). (E): Pancreatic ducts from nontransplanted animal (negative control) show CK expression (green) but not GFP. Scale bars ⫽ 50 m. Abbreviations: CK, cytokeratin; DAPI, 4⬘6-diamidino-2-phenylindole; GFP, green fluorescent protein.
www.StemCells.com
To test whether a similar pattern of ductal engraftment from bone marrow could be seen when transplantation occurred outside the neonatal period, GFP⫹ bone marrow cells were infused into nonirradiated adult Nod/Scid/2Mnull mice. High levels of hematopoietic engraftment were seen at the time of sacrifice (33.9% ⫾ 5.8% GFP⫹ cells in peripheral blood). Among 2,352 ductal structures from four mice sacrificed after 2 months, a median of 0.5% (range 0%–1.6%) of ducts contained GFP⫹ cells, 0.1% (range 0%– 0.2%) of ducts contained more than 50% donor GFP⫹ epithelial cells, and 0.4% (range 0%–1.4%) of ducts contained less than 50% donor GFP⫹ epithelial cells (Table 1). Analysis of the data from all animals studied confirmed that the number of ducts containing donor-derived cells was significantly higher in neonatal than adult recipients (p ⬍ .05). No beta cells of donor bone marrow origin were found in animals transplanted as adults.
DISCUSSION The diverse array of the body’s epithelial cells is developed and maintained by mechanisms that are largely unknown [19]. Although many organs are believed to contain epithelial stem cells responsible for the regeneration of mature special-
Downloaded from www.StemCells.com at University of California on March 9, 2009
that coexpressed GFP and CK. Two patterns of ductal engraftment by donor cells could be seen. In the first pattern, ducts were completely or mostly derived from GFP⫹ donor cells (Figs. 3A, 3B). These were recorded as ducts with greater than 50% of epithelial cells coexpressing GFP (Table 1). In these cases, clusters of large and small ducts, all expressing GFP, were often found, as if whole branching structures were derived from donor bone marrow. Coexistence of donor- and hostderived ducts in the same section could be seen (Fig. 3B), demonstrating the specificity of staining. In the second pattern, only small numbers of epithelial cells of donor origin were found scattered within each of the ducts (Fig. 3C). The existence of two or more isolated GFP⫹ cells per duct with this pattern was recorded as ducts with less than 50% donor cells (Table 1). Ducts containing only one GFP⫹ cell were not recorded. The pattern and specificity of GFP staining in ductal epithelium were validated with pancreas tissues from GFP transgenic mice (Fig. 3D) and nontransplanted Nod/Scid/2Mnull mice (Fig. 3E). In the nine neonatal transplanted mice analyzed at 2 months, a median of 0.2% (range 0%–31.3%) of all ducts had more than 50% donor cells and 4.1% (range 0.9%–11.8%) of all ducts had less than 50% donor cells (Table 1). The colocalization of CK and GFP in the donor-derived ductal epithelial cells can be clearly shown at higher magnification (Figs. 4A– 4C). FISH and CK fluorescence immunohistochemistry analysis of the sections from female recipients showed donor-derived ductal epithelial cells with Y chromosome, confirming the pres-
311
Pancreatic Ducts from Bone Marrow
312 Table 1. Frequency of donor-derived pancreatic ducts after neonatal BMT
Median percentages of ducts containing donor-derived cells (confidence interval) 关range兴 With >50% GFPⴙ cell Nontransplanted (1,029 ducts, n ⫽ 4 mice) Neonatal BMT (4,034 ducts, n ⫽ 9 mice) Adult BMT (2,352 ducts, n ⫽ 4 mice)
With
