SEARCH

SEARCH BY CITATION

Keywords:

  • Transplantation;
  • NOD/SCID model;
  • Umbilical cord blood;
  • Aldehyde dehydrogenase;
  • CD133

Abstract

  1. Top of page
  2. Abstract
  3. Introduction
  4. Materials and Methods
  5. Results
  6. Discussion
  7. Disclosure of Potential Conflicts of Interest
  8. Acknowledgements
  9. References

Transplanted adult progenitor cells distribute to peripheral organs and can promote endogenous cellular repair in damaged tissues. However, development of cell-based regenerative therapies has been hindered by the lack of preclinical models to efficiently assess multiple organ distribution and difficulty defining human cells with regenerative function. After transplantation into β-glucuronidase (GUSB)-deficient NOD/SCID/mucopolysaccharidosis type VII mice, we characterized the distribution of lineage-depleted human umbilical cord blood-derived cells purified by selection using high aldehyde dehydrogenase (ALDH) activity with CD133 coexpression. ALDHhi or ALDHhiCD133+ cells produced robust hematopoietic reconstitution and variable levels of tissue distribution in multiple organs. GUSB+ donor cells that coexpressed human leukocyte antigen (HLA-A,B,C) and hematopoietic (CD45+) cell surface markers were the primary cell phenotype found adjacent to the vascular beds of several tissues, including islet and ductal regions of mouse pancreata. In contrast, variable phenotypes were detected in the chimeric liver, with HLA+/CD45+ cells demonstrating robust GUSB expression adjacent to blood vessels and CD45/HLA cells with diluted GUSB expression predominant in the liver parenchyma. However, true nonhematopoietic human (HLA+/CD45) cells were rarely detected in other peripheral tissues, suggesting that these GUSB+/HLA/CD45 cells in the liver were a result of downregulated human surface marker expression in vivo, not widespread seeding of nonhematopoietic cells. However, relying solely on continued expression of cell surface markers, as used in traditional xenotransplantation models, may underestimate true tissue distribution. ALDH-expressing progenitor cells demonstrated widespread and tissue-specific distribution of variable cellular phenotypes, indicating that these adult progenitor cells should be explored in transplantation models of tissue damage.

Disclosure of potential conflicts of interest is found at the end of this article.


Introduction

  1. Top of page
  2. Abstract
  3. Introduction
  4. Materials and Methods
  5. Results
  6. Discussion
  7. Disclosure of Potential Conflicts of Interest
  8. Acknowledgements
  9. References

Potentially useful populations of adult stem cells for tissue repair or regenerative medicine include multipotent adult progenitors [1, [2]3] and primitive cells from hematopoietic [4], mesenchymal [5], or endothelial [6] lineages. All of these cell types were first identified in human bone marrow (BM), but several are also present in human umbilical cord blood (UCB) [7, [8], [9]10]. In the context of regenerative therapies, the contribution of adult stem cells is not limited to direct replacement of damaged cells. Introduction of transplanted adult cells can enhance survival, induce proliferation, and improve the function of damaged recipient cells in neural [11, 12], cardiac [13, [14]15], and endocrine tissues [16, 17]. Therefore, transplanted hematopoietic [18, [19]20], mesenchymal [15, 21], or endothelial progenitors [14, 22], from BM or UCB, may orchestrate the coordinated release of proangiogenic or prosurvival factors, resulting in improved cellular function. Regardless of the exact mechanism of tissue repair, the prospective isolation of human cells with documented regenerative function after transplantation has proven difficult because of the lack of sensitive models to detect tissue residence, particularly in solid organs.

Although human hematopoietic and endothelial progenitors can be identified by clonogenic assays in vitro, whether these cells seed and survive in peripheral tissues after transplantation in vivo is not well described. We have previously identified putative mixed progenitor populations according to conserved cytosolic aldehyde dehydrogenase (ALDH) activity [23], with or without further purification using expression of CD133, a cell surface marker expressed on hematopoietic and endothelial progenitors [24, 25]. Cytosolic ALDH is an enzyme that is highly expressed in hematopoietic progenitors [26] and implicated in the resistance of hematopoietic progenitor cells to alkylating agents [27]. Transplantation of lineage-depleted (Lin), ALDH-expressing cells into immune-deficient NOD/SCID mice produces robust, multilineage reconstitution in hematopoietic organs [23, 24]. To further characterize the distribution and survival of these progenitor cells in multiple tissues, we intravenously transplanted UCB-derived ALDHlo/hi and ALDHhiCD133−/+ cells into NOD/SCID/mucopolysaccharidosis type VII (MPSVII) mice, a model designed to accurately document donor/recipient cell interactions in peripheral tissues.

β-Glucuronidase (GUSB) is a lysosomal enzyme that is ubiquitously expressed. GUSB deficiency results in the lysosomal storage disease MPSVII [28], characterized by skeletal dysplasia, mental retardation, and reduced lifespan. GUSB-deficient mice [29] have been used to study disease progression and the localization of various transplanted murine cell types [30, [31], [32], [33]34]. By crossing the MPSVII mutation onto the NOD/SCID background [35], transplanted human cells can readily be visualized by virtue of their GUSB activity without reliance on the persistent expression of human-specific cell surface markers.

In this study, we used the NOD/SCID/MPSVII model to characterize the ability of human ALDH-expressing populations to reconstitute hematopoiesis and disseminate to nonhematopoietic tissues. After transplantation, ALDH-expressing cells were widely trafficked to peripheral organs and demonstrated variable distribution patterns. Human GUSB+ donor cells coexpressing hematopoietic (CD45) cell surface markers were the primary cell phenotype in vascular beds of organs, including the islet and ductal regions of mouse pancreata. Variable donor cell phenotypes were detected in the chimeric liver, with GUSB+ cells demonstrating reduced expression of both human and hematopoietic cell surface markers, indicating more widespread tissue distribution after xenotransplantation than had been previously detected.

Materials and Methods

  1. Top of page
  2. Abstract
  3. Introduction
  4. Materials and Methods
  5. Results
  6. Discussion
  7. Disclosure of Potential Conflicts of Interest
  8. Acknowledgements
  9. References

NOD/SCID/MPSVII Mice

The NOD/SCIDMPSVII mouse was produced by M.S.S. at Washington University School of Medicine (St. Louis, MO) by 10 backcrosses of the MPSVII mutation from its original strain (B6.C-H-2bml) onto the NOD/SCID mouse background (both mice were from Jackson Immunoresearch Laboratories, West Grove, PA, http://www.jacksonimmuno.com) [35]. Experimental NOD/SCID/MPSVII−/− mice, bred in our colony at Washington University in compliance with all regulatory committees, were identified by a GUSB sequence-specific polymerase chain reaction assay and confirmed by a lack of GUSB activity as described previously [35, 36]. Human cell reconstitution after the transplantation of human MSC, UCB-derived, or mobilized peripheral blood-derived CD34+ cells into NOD/SCID/MPSVII mice has been previously detailed [35, 37], with repopulating frequencies equivalent to those of the parental immune-deficient NOD/SCID mice.

Human Cell Purification by Aldehyde Dehydrogenase Activity

Human UCB samples were obtained from the cord blood banking facility at Cardinal Glennon Children's Hospital (St. Louis, MO) and used in accordance with the guidelines of local ethical and biohazard authorities at Washington University School of Medicine (St. Louis, MO). UCB samples were diluted with phosphate-buffered saline (PBS), and mononuclear cells (MNC) were isolated by Hypaque-Ficoll centrifugation. MNC were depleted of contaminating erythrocytes by red blood cell lysis in 0.8% ammonium chloride solution (StemCell Technologies, Vancouver, BC, Canada, http://www.stemcell.com). Human UCB MNC were enriched for Lin cells by magnetic bead separation as described previously [23, 24]. The resulting Lin population was further purified on the basis of ALDH activity by staining with Aldefluor substrate (Aldagen [formerly StemCo Biomedical], Durham, NC, http://www.aldagen.com) according to the manufacturer's specifications. This fluorescent substrate is metabolized by cytosolic ALDH and retained within the cells because of its negative charge [23, 26], and ALDHloLin or ALDHhiLin cells were selected by fluorescence-activated cell sorting (FACS) (MoFlo; Dako, Glostrup, Denmark, http://www.dako.com). Alternatively, Lin cells were also costained with anti-human CD133-APC (Miltenyi Biotec, Auburn, CA, http://www.miltenyibiotec.com) after incubation with the Aldefluor substrate, and ALDHhiCD133Lin or ALDHhiCD133+Lin cells were selected by FACS. Cell populations were routinely isolated to >95% purity and >95% viability by trypan blue staining and were screened for CD34 and CD45 (all antibodies from Becton, Dickinson and Company, San Jose, CA, http://www.bd.com) expression by flow cytometry.

Transplantation of Purified Cell Populations into NOD/SCID/MPSVII Mice

Unsorted Lin, ALDHloLin, ALDHhiLin, ALDHhiCD133Lin, or ALDHhiCD133+Lin cells were transplanted by tail vein injection into 8–12-week-old, sublethally irradiated (300 cGy) NOD/SCID/MPSVII−/− mice. NOD/SCID/MPSVII mice transplanted with PBS served as controls for analysis by flow cytometry and histochemical staining.

Analysis of Human Cell Distribution in the Tissues of Transplanted Mice

BM was flushed from the marrow compartments with PBS supplemented with 2% fetal calf serum (FCS). Spleen, liver, lung, and pancreas were mechanically dissociated without enzymatic digestion, sequentially filtered, and resuspended in PBS/2% FCS. Peripheral blood was collected from anesthetized mice by retro-orbital eye bleed as described previously [23]. BM and peripheral blood were lysed with a 0.8% ammonium chloride solution (StemCell Technologies). For FACS analysis, approximately 106 suspended cells were incubated with blocking solution and monoclonal antibodies for the human panleukocyte-specific marker CD45 in combination with HLA-A,B,C or isotype controls (all antibodies from Becton Dickinson). Dissociated solid tissues (liver, lung, and pancreas) were also costained with a viability marker, 7-aminoactinomycin D (7-AAD) (Becton Dickinson), immediately before analysis to exclude cells damaged by processing. Cells were analyzed by three-color flow cytometry on a Coulter FC-500 flow cytometer (Beckman Coulter, Miami, http://www.beckmancoulter.com). For flow cytometric analysis of GUSB positivity, dissociated cells from mouse spleen, liver, or pancreas were incubated with 100 μM ImaGene Green C12-FDGlcU GUSB substrate (Molecular Probes, Eugene, OR, http://probes.invitrogen.com) for 1–2 hours at 37°C. Cells were costained with anti-human CD45-APC or APC-isotype control and assayed on an FC-500 flow cytometer as described above.

GUSB-Specific Histochemical Assay

Liver, spleen, lung, pancreas, heart, kidney, brain, eye, skeletal muscle, sternum, and hip from NOD/SCID/MPSVII were frozen in optimal cutting temperature embedding medium (Sakura, Torrance, CA, http://www.sakuraus.com) for histochemical analysis. Serial sections were taken at 10-μm thickness and stained for GUSB activity as described previously, using naphthol AS-BI β-d-glucuronide (Sigma-Aldrich, St. Louis, http://www.sigmaaldrich.com) as a substrate [38, 39]. Slides were counterstained with methyl green.

Immunohistochemistry

Frozen sections from transplanted NOD/SCID/MPSVII mouse spleen, liver, and pancreas were fixed for 15 minutes in acetone at 4°C and blocked at room temperature with mouse-on-mouse reagent (Vector Laboratories, Burlingame, CA, http://www.vectorlabs.com) [35]. Unconjugated mouse anti-human CD45 primary antibody (anti-Hle-1; Becton Dickinson) diluted 1:200 was conjugated with an alkaline phosphatase goat anti-mouse IgG antibody (Sigma-Aldrich) followed by alkaline phosphatase development reagents (Vector Laboratories). Adjacent sections were sequentially immunostained for CD45 and GUSB activity as described above to detect donor-derived GUSB+ cells that coexpress CD45.

Results

  1. Top of page
  2. Abstract
  3. Introduction
  4. Materials and Methods
  5. Results
  6. Discussion
  7. Disclosure of Potential Conflicts of Interest
  8. Acknowledgements
  9. References

Isolation of Purified Human Populations

Increased ALDH activity is a characteristic of cells that possess documented neural and hematopoietic progenitor function in vivo [23, 26, 40, 41]. We have previously described the purification of UCB Lin cells according to ALDH activity by FACS, with the ALDHlo and ALDHhi populations making up 31.5% ± 1.8% and 50.1% ± 2.4% of total Lin cells, respectively [23]. Since endothelial progenitors can augment vascular regeneration (reviewed in [42]) and because CD133 expression is also conserved on progenitor cells from endothelial, hematopoietic, and neural lineages [9, 43, [44]45], ALDHhiLin cells were further purified on the basis of CD133 expression. As described previously [24], CD133 and CD133+ cells represented 33.7% ± 1.7% and 50.4% ± 2.5% of ALDHhiLin cells, or 14.7% ± 2.1% and 23.2% ± 4.3% of the total Lin cells, respectively. CD34 expression on the cells used in this study was similar to that in previous reports (ALDHloLin cells, 34.3% ± 8.3% CD34+; ALDHhiLin cells, 87.5% ± 13.7% CD34+; ALDHhiCD133Lin cells, 75.4% ± 7.5% CD34+; ALDHhiCD133+Lin cells, 92.7% ± 5.4% CD34+; and Lin cells, 55.4% ± 9.8% CD34+) [23, 24]. Starting populations demonstrated >95% expression of the panleukocyte marker CD45 by FACS (data not shown), indicating that UCB ALDH-expressing progenitors were primarily hematopoietic in origin.

Detection of Donor Cell Tissue Distribution After Transplantation of ALDH-Expressing Populations

Purified human ALDHloLin, ALDHhiLin, ALDHhiCD133Lin, and ALDHhiCD133+Lin cells (0.5–4 × 105 cells) were transplanted into NOD/SCID/MPSVII mice following 300 cGy radiation, without cotransplantation of supportive accessory cells [23, 46]. Human HLA-A,B,C (HLA, ubiquitously expressed on human nucleated cells) and CD45 (panleukocyte marker) coexpression was used to differentiate hematopoietic (HLA+/CD45+) from nonhematopoietic (HLA+/CD45) human cell detection by FACS (Fig. 1). Dead cells (7-AAD+) and debris were excluded from the analysis of liver and pancreas (data not shown).

thumbnail image

Figure Figure 1.. Fluorescence-activated cell sorting detection of human cells in the spleen, liver, and pancreas of transplanted NOD/SCID/mucopolysaccharidosis type VII (MPSVII) mice. Representative flow cytometric analysis of NOD/SCID/MPSVII mouse bone marrow (BM), liver, and pancreas after i.v. transplantation with 2 × 105 purified ALDHhiLin(A–C) or ALDHhiCD133+Lin(D–F) cells. At 5–6 weeks post-transplantation, human hematopoietic cells were detected in mouse tissues by coexpression of human HLA-A,B,C and the human panleukocyte marker CD45. Mice injected with ALDHhiLin cells showed robust repopulation in the BM (74.4% ± 7.9%; n = 4) and lower but consistent distribution in the liver (1.3% ± 0.3%; n = 4) and pancreas (0.3% ± 0.1%; n = 4). Mice transplanted with ALDHhiCD133+Lin cells showed lower chimerism in the BM (25.6% ± 10.4%; n = 4), liver (0.7% ± 0.2%; n = 4), and pancreas (0.2% ± 0.1%; n = 4) compared with ALDHhiLin cells. Human donor cells were not detected in mice transplanted with ALDHloLin (n = 3; [A], inset) or ALDHhiCD133Lin cells (n = 3; [D], inset). Abbreviation: 7-AAD, 7-aminoactinomycin D; HLA A,B,C, human leukocyte antigen A,B,C.

Download figure to PowerPoint

Intravenous injection of 2 × 105 ALDHhiLin cells produced a high frequency (74.4% ± 7.9%; n = 4) of human hematopoietic (HLA+/CD45+) reconstitution in NOD/SCID/MPSVII mouse BM, analyzed 5–6 weeks post-transplantation (Fig. 1A). An equal dose of ALDHhiCD133+Lin cells (Fig. 1D) reconstituted the BM at a significantly (p < .05) reduced level (25.6% ± 10.4%; n = 4). Corresponding ALDHloLin (Fig. 1A, inset; n = 3) and ALDHhiCD133Lin (Fig. 1D, inset; n = 3) cells were not detected in the BM or any other tissue. For comparison, transplantation of 2 × 105 unfractionated Lin cells produced 19.0% ± 7.0% (n = 3) human cell repopulation at 5–6 weeks post-transplantation (data not shown). Both ALDHhiLin and ALDHhiCD133+Lin cells also demonstrated consistent seeding in the liver and pancreas (Fig. 1B, 1C, 1E, 1F). As detected by flow cytometry to assess cell surface markers, cells seeding these tissues appeared to be primarily hematopoietic in origin, as nonhematopoietic (HLA+/CD45) cells were rarely detected in tissues using this particular assay (Fig. 1).

We transplanted a total of 46 NOD/SCID/MPSVII mice with 0.5–4 × 105 ALDHhiLin, ALDHloLin, ALDHhiCD133+Lin, or ALDHhiCD133Lin cells and quantified human cell chimerism by FACS in the BM (Fig. 2A), peripheral blood (Fig. 2B), liver (Fig. 2C), and pancreas (Fig. 2D). Human ALDHhiLin cells produced prolonged hematopoietic reconstitution in the BM of all transplanted mice at 5–6 (59.1% ± 7.8% HLA+/CD45+; n = 11) and 10–12 (66.0% ± 5.0% HLA+/CD45+; n = 8) weeks post-transplantation (Fig. 2A). In contrast, transplanted ALDHloLin cells demonstrated little or no reconstituting ability, with only 2 of 11 mice demonstrating <0.3% human cells after the injection of as many as 4 × 105 cells. Highly purified ALDHhiCD133+Lin cells repopulated the murine BM at approximately threefold lower levels (p < .05) compared with ALDHhiLin progenitors (Fig. 2A). Transplanted ALDHhiCD133Lin cells did not engraft. Transplantation of ALDHhiLin (n = 3) and ALDHhiCD133+Lin (n = 3) cells yielded primarily myeloid and B-lymphoid progeny (data not shown). Stable, multilineage hematopoietic repopulation was established in mice transplanted with ALDH-expressing cells, providing a potential reservoir of human cells for dispersal to peripheral tissues.

thumbnail image

Figure Figure 2.. Summary of human cell distribution in transplanted NOD/SCID/mucopolysaccharidosis type VII (MPSVII) mice. The bone marrow (BM) (A), peripheral blood (B), liver (C), and pancreas (D) of NOD/SCID/MPSVII mice transplanted intravenously with 0.5–4 × 105 purified ALDHhiLin (•; n = 19), ALDHloLin (○; n = 11), ALDHhiCD133+Lin (▪; n = 9), or ALDHhiCD133Lin (□; n = 7) cells were analyzed by fluorescence-activated cell sorting. Human cells were detected by coexpression of human HLA-A,B,C and CD45. Transplanted ALDHhiLin cells consistently produced a higher frequency of human progeny in the BM, peripheral blood, liver, and pancreas, as compared with ALDHhiCD133+Lin cells (p < .05). Human cells were not observed in any peripheral tissues after the transplantation of ALDHloLin or ALDHhiCD133Lin cells. Abbreviation: HLA A,B,C, human leukocyte antigen A,B,C.

Download figure to PowerPoint

Established hematopoietic chimerism established the transit of human cells in the peripheral blood of transplanted mice. Thus, high BM reconstitution by transplanted ALDHhiLin cells led to consistent detection of human cells in the peripheral blood at 5–6 (13.4% ± 6.0% HLA+/CD45+; n = 11) and 10–12 (12.9% ± 6.8% HLA+/CD45+; n = 8) weeks post-transplantation (Fig. 2B). After transplantation of ALDHhiCD133+Lin cells, human progeny were detected at lower levels in the peripheral blood at 5–6 (2.8% ± 1.5% HLA+/CD45+, p < .01) and 10–12 (0.4% ± 0.1% HLA+/CD45+, p < .05) weeks post-transplantation, suggesting that lowered BM chimerism reduced trafficking of human cells via the murine circulation.

Compared with BM and peripheral blood, lower levels of human cells were observed by FACS in the liver (Fig. 2C) and pancreas (Fig. 2D) of NOD/SCID/MPSVII mice transplanted with ALDHhiLin or ALDHhiCD133+Lin cells. For all transplanted cell populations, the frequency of human cells in the liver was approximately fivefold higher than in the pancreas (Fig. 2C, 2D). Human cell detection in these tissues was consistently higher after transplantation of ALDHhiLin cells, as compared with ALDHhiCD133+Lin cells. Taken together, these data suggest that increased hematopoietic chimerism led to increased peripheral tissue seeding and that human cells were distributed to various peripheral tissues via the circulation at variable frequencies after i.v. transplantation.

Distribution of GUSB+ Cells in the Tissues of Transplanted NOD/SCID/MPSVII Mice

The major advantage of the NOD/SCID/MPSVII model used in these studies is the sensitive detection of GUSB+ donor cells in tissues without reliance on the continued expression of human cell surface markers or in situ hybridization [35]. In addition, cell phenotype, distinct human cell localization within solid tissues, and interaction with recipient cells can also be documented by microscopy, without the use of FACS. Figure 3 shows the representative distribution of GUSB+ cells in situ for mice injected with ALDHloLin (Fig. 3A), ALDHhiLin (Fig. 3B), or ALDHhiCD133+Lin (Fig. 3C) cells in the spleen, liver, and pancreas at 6 weeks post-transplantation. Tissues from NOD/SCID/MPSVII mice not transplanted with human cells (PBS-injected) were completely devoid of GUSB staining (data not shown). Mice injected with nonrepopulating ALDHloLin (Fig. 3A) cells showed only rare GUSB+ cells in spleen (Fig. 3A, arrow), liver, or pancreas. Consistent with flow cytometric data, ALDHhiCD133Lin cells or related progeny were not detected in any analyzed tissues by GUSB histochemical expression (data not shown). In contrast, widespread distribution of GUSB+ donor cells was observed in tissues from ALDHhiLin (Fig. 3B) and ALDHhiCD133+Lin transplanted mice (Fig. 3C). As shown in Figure 4, high levels of human cell dispersal via the peripheral circulation were observed 6 weeks after i.v. injection of ALDHhiLin cells in a highly reconstituted (66.5% human HLA+/CD45+ cells in murine BM) representative mouse. Notably, GUSB+ human cells could be found in nearly all tissues examined, suggesting that transplanted ALDHhiLin progenitors from UCB display more widespread distribution in peripheral tissues than was previously recognized [23, 24].

thumbnail image

Figure Figure 3.. Donor-derived β-glucuronidase+ (GUSB+) cells in the spleen, liver, and pancreas of transplanted NOD/SCID/mucopolysaccharidosis type VII (MPSVII) mice. Spleen, liver, and pancreas sections from a NOD/SCID/MPSVII mouse 5–6 weeks after i.v. transplantation with 106 ALDHloLin(A), 105 ALDHhiLin(B), or 105 ALDHhiCD133+Lin(C) cells were analyzed for GUSB expression (naphthol-AS-BI-β-d-glucuronide staining, red) and counterstained using methyl green (magnification, ×200). Single GUSB-expressing cells were rarely detected in the spleen and liver (arrows) and were not detected in the pancreas after transplantation of ALDHloLin cells. GUSB+ cells were not detected in tissues after transplantation of up to 4 × 105 ALDHhiCD133Lin cells (not shown). Transplantation of ALDHhiLin or ALDHhiCD133+Lin cells produced widespread detection of GUSB+ cells throughout the spleen and liver and adjacent to ductal regions of the pancreas (arrowheads). Abbreviations: ALDH, aldehyde dehydrogenase; Lin, lineage-depleted; PBS, phosphate-buffered saline.

Download figure to PowerPoint

thumbnail image

Figure Figure 4.. Widespread tissue distribution of transplanted human ALDHhiLin cells in NOD/SCID/mucopolysaccharidosis type VII (MPSVII) mice. Tissue sections from a representative NOD/SCID/MPSVII mouse 5–6 weeks after i.v. transplantation with 2 × 105 ALDHhiLin cells were analyzed for β-glucuronidase (GUSB) expression (naphthol-AS-BI-β-d-glucuronide staining, red; magnification, ×200). Donor-derived GUSB+ cells were abundant in hematopoietic tissues, such as the sternum marrow (A) and spleen (B), whereas solid bone showed less staining, with occasional positive cells. Human cells uniformly infiltrated highly perfused tissues, such as the liver (C) and lung (D), at high frequencies. The kidney (E), heart (F), pancreas (G), and cartilage (H) showed consistent but comparatively lowered detection of GUSB+ cells. GUSB+ human cells were also detected in the brain (I) and were closely associated with the retinal pigment epithelial layer of the eye (J). Similar tissue distribution was observed after transplantation with ALDHhiCD133+Lin cells.

Download figure to PowerPoint

After transplantation of 2 × 105 ALDHhiLin cells, donor-derived GUSB+ cells were abundant in hematopoietic tissues such as the sternum marrow (Fig. 4A) and spleen (Fig. 4B). Transplanted mice that produced more modest murine BM repopulation showed clustering of GUSB+ cells along the endosteal surface of the bone, the site of primitive progenitor adhesion within the BM [47, 48]. Transplanted human ALDHhiLin cells uniformly infiltrated highly vascular tissues, such as the liver (Fig. 4C) and lung (Fig. 4D), whereas kidney (Fig. 4E), cardiac muscle (Fig. 4F), and cartilage (Fig. 4G) showed less dense scattered distribution. Patterned distribution by ALDHhiLin cells was observed in the pancreas, with donor-derived GUSB+ cells selectively found neighboring both ductal and islet regions but not highly represented in exocrine areas (Fig. 4H). Scattered GUSB+ cells were present at reduced intervals in the brain (Fig. 4I) and in close association with the retinal pigment epithelial layer of the eye (Fig. 4J). A similar pattern of GUSB-expressing cells was observed at 5–6 and 10–12 weeks post-transplantation. Mice transplanted with ALDHhiCD133+Lin cells demonstrated a similar pattern of distribution, albeit with proportionately fewer cells (data not shown).

Transplanted Human ALDHhiLin Cells Localized to Ductal and Islet Regions in the Pancreas

Tissues such as the liver and pancreas may provide amenable targets for the development of cell and gene therapy applications. Distinct, patterned distribution of human ALDH-expressing progenitors (ALDHhiLin and ALDHhiCD133+Lin cells) was consistently observed in the pancreas of transplanted NOD/SCID/MPSVII mice (Fig. 5). Specifically, GUSB+ cells or progeny were consistently detected adjacent to ductal (Fig. 5A) and islet (Fig. 5B) structures in the proximal portion of the endocrine pancreas. GUSB+ donor cells clearly surrounded the borders of islet structures (Fig. 5B). This distribution was observed in all pancreata examined by GUSB histochemistry (n = 4 for ALDHhiLin transplants). Donor-derived GUSB+ cells located in both ductal and islet regions consistently costained with human CD45 (Fig. 5C), indicating that these cells were primarily hematopoietic in origin. The arrows in Figure 5D clearly mark CD45+ (brown) and GUSB+ (red) donor-derived cells, confirming flow cytometric data that human cells in the pancreas were primarily hematopoietic. Furthermore, this distribution occurred without prior chemical damage to β cells and did not result in obvious inflammation. Insulin staining and control of blood glucose in these mice were normal prior to and throughout the post-transplantation period (data not shown).

thumbnail image

Figure Figure 5.. Human hematopoietic cells distributed to pancreatic duct and islet regions. Pancreatic sections from representative NOD/SCID/mucopolysaccharidosis type VII mice 10–12 weeks after i.v. transplantation with 1–4 × 105 ALDHhiLin cells were analyzed for β-glucuronidase (GUSB)-expressing cells (naphthol-AS-BI-β-d-glucuronide [ASBI] staining, red), with nuclei counterstained using methyl green (magnification, ×100–×200). Donor-derived GUSB+ cells displayed patterned distribution adjacent to acinar ducts (A) and surrounding islets (arrowheads, [B]). Peri-islet cells were commonly found adjacent to ductal regions. (C, D): Colocalization of GUSB activity (ASBI, red) and human panleukocyte CD45 (brown) confirmed that donor-derived GUSB+ cells detected within the pancreas were primarily hematopoietic (CD45+) in origin (arrows). Abbreviations: d, duct; i, islet.

Download figure to PowerPoint

Characterization of Donor Cells Detected in the Liver

Because the fusion of mouse and human BM cells with murine hepatocytes after xenotransplantation has been established [49, 50] and because preliminary GUSB screening in the liver suggested more widespread human cell residence than predicted by FACS (Fig. 1B), we further investigated the phenotypes of cells detected in the liver (Fig. 6). Representative photomicrographs of representative liver sections from chimeric mice transplanted with human UCB ALDHhiLin cells at low (Fig. 6A, ×100) and high (Fig. 6B, ×400) magnifications demonstrated the diffuse dispersal of human cells throughout the liver and also visualized the cytoplasmic localization of GUSB staining in donor-derived human cells ranging in size from 5 to >20 μm diameter. The majority of human cells detected within liver tissue 5–6 weeks post-transplantation were hematopoietic (CD45+) (Fig. 6C) in origin and coexpressed mature lineage markers for human myeloid cells (Fig. 6D, CD45+CD33+), human B-lymphocytes (Fig. 6E, CD45+CD19+), and human macrophages (Fig. 6E, CD45+CD11b+). Although these analyses were performed without tissue perfusion prior to euthanasia to remove blood vessel-resident hematopoietic cells, these data demonstrate the widespread distribution of donor hematopoietic cells with access to a hepatic microenvironment. Other highly vascularized tissues, such as the myocardium (Fig. 4F) and neural tissue (Fig. 4I), did not achieve such widespread donor cell distribution.

thumbnail image

Figure Figure 6.. Hematopoietic cells distributed to the liver demonstrated β-glucuronidase (GUSB) and human hematopoietic cell surface marker expression. Liver sections from mice injected with 2 × 105 ALDHhiLin(A, B) were analyzed at 5–6 weeks post-transplantation for the presence of GUSB activity. A low-power photomicrograph (A) (magnification, ×100) demonstrated the widespread distribution of donor GUSB+ cells. A high-power photomicrograph (B) (magnification, ×400) demonstrated the variable morphology and size of cells donor cells dispersed throughout the liver tissue. (C–E): Fluorescence-activated cell sorting analysis was performed on human hematopoietic cells (CD45+; R1) distributed in the liver at 5–6 weeks post-transplantation. Hematopoietic cells coexpressed mature lineage markers for myeloid cells (CD33, CD15), B-lymphocytes (CD19), and macrophages (CD11b) (n = 3). Abbreviation: FSC, forward scatter.

Download figure to PowerPoint

At 10–12 weeks post-transplantation, two distinct donor-derived cell phenotypes demonstrating different morphology and intensity of GUSB-staining were distributed throughout the livers of mice transplanted with either ALDHhiLin (Fig. 7A, 7B) or ALDHhiCD133+Lin (Fig. 7C, 7D) cells at 10–12 weeks post-transplantation. Other tissues, such as the lung (Fig. 4D) and kidney (Fig. 4E), also showed various intensities of GUSB expression in donor cells, whereas GUSB intensities were consistently high in individual cells detected within the pancreas (Fig. 4G), brain (Fig. 4I), and eye (Fig. 4J). Although different cell types can vary in their expression of endogenous GUSB, cells detected in the livers of transplanted mice revealed the presence of small, round, intensely red-stained cells (Fig. 7A–7C, arrowheads) and larger, moderately red-stained cells (Fig. 7A–7D, arrows) throughout the liver. These larger, moderately stained donor cells were present in mouse livers at 5–6 and 10–12 weeks post-transplantation.

thumbnail image

Figure Figure 7.. Donor cells in the liver demonstrate variable GUSB, CD45, and hepatocyte-specific albumin expression. Liver sections from mice injected with 2 × 105 ALDHhiLin(A, B) or 2 × 105 ALDHhiCD133+Lin(C, D) cells contained small, dark red-stained GUSB+ cells (arrowheads) and larger, more diffusely stained cells with typical hepatocyte morphology (arrows). (E–G): Human donor cells were detected using a fluorescent substrate of GUSB and human panleukocyte marker CD45 antibody at 10–12 weeks post-transplantation. GUSBhi cells (E) were detected at 15.0% ± 4.9% (R1) of total cells (n = 3), whereas human CD45 (F) was expressed at only 0.7% ± 0.4% (R2) of total cells (n = 3). (G): Using dual-color fluorescence-activated cell sorting, all human CD45+ cells also showed high GUSB fluorescent intensity (GUSBhi) and represented 1.1% ± 0.6% of total cells (n = 3). The remainder of GUSB-expressing cells displayed lowered fluorescence intensity and did not express human CD45 (13.4% ± 4.5%; n = 3). (H, I): Histochemical staining for GUSB activity (ASBI, red), and CD45 expression (brown) confirmed the presence of numerous GUSB+ cells that were negative for CD45 cell surface expression (arrowheads). (J, K): Costaining for human hepatocyte-specific albumin using immunofluorescence and for GUSB expression using immunohistochemistry performed on liver-resident donor cells revealed the presence of rare clusters and individual human cells that possess a typical hepatocyte function (arrows). The majority of donor-derived GUSB+ human cells did not costain for human albumin (arrowheads; n = 3). Abbreviations: ALDH, aldehyde dehydrogenase; ASBI, naphthol-AS-BI-β-d-glucuronide; GUSB, β-glucuronidase.

Download figure to PowerPoint

Flow cytometric analysis of dissociated liver tissue at 10–12 weeks after transplantation with ALDHhiLin cells revealed a relatively high frequency of cells (R1, 15.0% ± 4.9%) that metabolized a fluorescent substrate of GUSB (Fig. 7E). In contrast, these mice (n = 4) showed low frequencies of human HLA-A,B,C (0.9% ± 0.5%) and CD45 (0.7% ± 0.4%) cell surface expression by flow cytometry (Fig. 7F). Dual-color flow cytometry for GUSB activity and human cell surface CD45 expression revealed a small but distinct cluster of cells intensely stained for GUSB activity that coexpress human cell surface CD45 (1.1% ± 0.6% GUSBhiCD45+) and a larger population of donor-derived, GUSB-expressing cells (13.4% ± 4.5%) with reduced expression of CD45 (Fig. 7G). The CD45+ cells always coexpressed human HLA-A,B,C, whereas human HLA-A,B,C expression was not detected on any CD45 cells detected in the liver (data not shown), suggesting that both human HLA-A,B,C and CD45 expression was reduced by donor-derived cells within the murine liver. CD45 cells were not evident in other nonhematopoietic tissues, such as pancreas (Fig. 5C), where there was concordance between CD45 and GUSB staining. In addition, CD45 cells were not observed in previous studies of human UCB CD34+ cell transplantation in the NOD/SCID/MPSVII model by Hofling et al., where liver-resident cells always coexpressed CD45 [35]. Thus, donor cells in the liver of mice transplanted with ALDHhiLin cells revealed more variable phenotypes than previously realized with Lin or CD34+ cells.

To further confirm these findings, serial sections from livers were stained for CD45 expression alone (Fig. 7H), or costained to detect GUSB activity (red) and CD45 (brown) expression in the same cell (Fig. 7I). Cells expressing human CD45 were rarely found dispersed through the liver parenchyma; however, occasional clusters of CD45+ cells were observed in perivascular regions (Fig. 7H, 7I). Numerous GUSB+ cells that did not express human CD45 (Fig. 7I, arrowheads) were observed throughout the liver. These liver-resident GUSB+CD45 cells were more numerous at 10–12 weeks post-transplantation, were widely dispersed, were in close contact with murine hepatocytes, and were distinct from the GUSB+CD45+ human cells within liver blood vessels (Figs. 6, 7). Thus, detection of GUSB expression using the NOD/SCID/MPSVII model provides a unique method of detecting these persistent donor-derived cells within nonhematopoietic tissues.

To uncover the identity or lineage-restriction of CD45 donor cells detected within the liver, we stained serial sections for human hepatocyte-specific albumin expression using immunofluorescence (Fig. 7J), followed by routine GUSB expression using colorimetric immunohistochemistry (Fig. 7K). This analysis revealed the presence of rare clusters and individual human cells that possess a human albumin expression, a well-documented hepatocyte-specific function in vivo (arrows) [49, 50]. However, the majority of donor-derived GUSB+ human cells did not costain for human albumin (arrowheads, Fig. 7K; n = 3), suggesting that some but not all liver-resident progeny of UCB ALDHhiLin cells possessed the capacity to adopt typical hepatocyte morphology and function after i.v. transplantation, hematopoietic reconstitution, and transport to the liver via the peripheral circulation.

Discussion

  1. Top of page
  2. Abstract
  3. Introduction
  4. Materials and Methods
  5. Results
  6. Discussion
  7. Disclosure of Potential Conflicts of Interest
  8. Acknowledgements
  9. References

Highly enriched human UCB-derived cells expressing high ALDH activity demonstrated robust hematopoietic reconstitution and efficient dissemination to multiple peripheral tissues of transplanted NOD/SCID/MPSVII mice via the peripheral circulation. Hematopoietic reconstitution by the ALDHhiLin population in the BM and spleen exceeded both unfractionated Lin cells and further purified ALDHhiCD133+Lin population after transplantation of equal cell doses and confirmed the strong hematopoietic repopulating function of ALDH-expressing cells from UCB [23, 24]. We have previously reported that the ALDHhiCD133+ population contained a higher SCID-repopulating cell (SRC) frequency compared with more heterogeneous ALDHhiLin cells when transplanted into immune-deficient recipients at limiting dilution [24]. Although fewer ALDHhiCD133+Lin cells are able to establish hematopoiesis in NOD/SCID and NOD/SCID β-2-microglobulin null mice, this highly purified population actually induced less robust repopulation, primarily because of the removal of accessory cells that support the efficient proliferation and expansion of ALDHhi SRC. A similar phenomenon has been previously reported when comparing BM chimerism by CD34+ and CD34+CD38 cells transplanted without accessory cell support [7, 46, 51]. Nonetheless, stable BM reconstitution was achieved by each cell population within 6 weeks post-transplantation; and reconstitution remained stable for at least 12 weeks. Increased levels of human hematopoietic reconstitution led to increased transit of human cells in the peripheral circulation and increased detection of donor cells in tissues. Thus, strong hematopoietic repopulation and widespread tissue dispersal underscore the potential of human ALDH-expressing progenitor populations in the future development of cellular therapies for hematopoietic replacement and/or tissue repair.

Established human cell chimerism in the murine BM following transplantation initiated increased trafficking of human cell progeny within the peripheral circulation. In turn, donor cell distribution to tissues rich in capillary beds, such as the lung and the liver, was increased compared with less-vascularized tissues. Thus, increased tissue perfusion may lead to efficient exposure of potentially regenerative cells to diseased or damaged tissues. However, widespread distribution of nonhematopoietic, human albumin-expressing cells in the liver cannot be fully explained by the initial seeding of cells in the microvasculature alone because other highly vascularized tissues, such as the brain and myocardium, do not demonstrate increased human cell distribution and variation in phenotype. In addition, quantitative detection by GUSB-expressing donor cells by FACS in the livers of chimeric animals at 10–12 weeks post-transplantation routinely outweighed the frequency of human cells detected by human CD45 or human HLA-A,B,C expression in the circulation, suggesting the accumulation of human cells that did not express human hematopoietic markers specifically in the liver.

Detailed analysis of tissue-resident cell phenotype revealed selective detection of human hematopoietic cells in vascular regions of tissues, including the endocrine pancreas surrounding both ducts and islets. Immediate and persistent transit of these cells to the pancreatic microenvironment may be important for the regeneration of β-cell function after streptozotocin treatment, as murine and human BM-derived progenitor cells have been shown to initiate proliferation and restore insulin content in hyperglycemic recipients [16, 17]. Thus, human ALDH-expressing cells represent putative progenitors with endocrine-delivery characteristics necessary for potential endogenous regeneration of damaged pancreatic tissue. Future experiments will address the potential of these cells for tissue repair.

The GUSB deficiency in our NOD/SCID/MPSVII model allowed for specific and sensitive detection of donor-derived, GUSB+ cells. In most tissues, including the BM, spleen, and pancreas, detectable GUSB activity correlated closely with human HLA-A,B,C and CD45 cell surface marker expression. However, flow cytometric detection of intracellular GUSB activity combined with human cell surface marker detection in the liver revealed GUSB-positive cells with low expression of human CD45. However, since the starting ALDH-expressing populations were >95% CD45+ and since HLA+/CD45 cells were rare in other tissues, our data suggest that GUSB+/HLA/CD45 cells in the liver may be the result of an overall reduction in human cell surface marker expression to the point where human HLA-A,B,C and CD45 were not detected in cells with considerable GUSB expression. GUSB expression within the liver was also distinct and confined within the membrane boundaries of cells with a hepatocyte morphology, indicating that the potential uptake of GUSB enzyme by murine hepatocytes by a phenomenon known as cross-correction was not observed in contrast to the previous report using transplanted CD34 cells [35]. In that report, there was a perfect concordance between CD45 and GUSB expression in the liver after transplantation of human CD34+ cells into the NOD/SCID/MPSVII mouse. These data demonstrated that human cells that express physiological levels of GUSB do not contribute sufficient enzyme to surrounding cells to be detected histochemically. Cross-correction to murine cells is also unlikely in the current studies, because in order for a GUSB murine cell to endocytose enough enzyme to be detected histochemically, a cell expressing high levels would need to be nearby. This is clearly not the case (Fig. 7D, 7E).

Cellular fusion between hematopoietic cells and hepatocytes has been well documented by several groups [49, 50, 52]. Furthermore, murine BM cell mobilization or human UCB cell transplantation in models of murine hepatic damage after CCl4 administration has demonstrated that these hematopoietic cells can undergo cell fusion events in vivo [53, 54]. Thus, in our system that did not use chemical-induced hepatocyte damage, the presence of moderate GUSB activity in CD45 cells of the liver may have occurred via cellular fusion and nuclear reprogramming, resulting in the dilution of GUSB activity, downregulation of human cell surface marker expression, and possible upregulation of human albumin production (Fig. 7J, 7K). However, we cannot rule out the possibility that donor UCB-derived nonhematopoietic (CD45) cells, namely putative mesenchymal or endothelial progenitors, which also demonstrate reduced expression of human HLA but lacked the expression of human albumin (Fig. 7), may have established residence specifically in murine liver [37]. In previous studies, transplanted human CD34+ cells did not demonstrate reduced human CD45 expression in the liver of NOD/SCID/MPSVII mice [35], indicating that this expression pattern may be unique to ALDH-expressing cells or circulating progeny. Regardless of lineage restriction, these liver-resident cells were donor-derived (GUSB+), and traditional models, which rely solely on the continued expression of cell surface markers, to detect tissue residence after transplantation may have underestimated donor cell distribution in peripheral organs.

In summary, ALDH-expressing populations from human UCB are an easily procured, rich source of repopulating multipotent progenitors for a variety of potential clinical cellular transplantation applications. Although ALDH cells from UCB are composed primarily of hematopoietic progenitors [23, [24], [25]26], ALDH-expressing cells from alternate adult sources, such as human BM, may also contain enriched mesenchymal or endothelial progenitors [55], which have recently been implicated in important regenerative processes [16, 17]. The diverse localization, variable donor cell phenotypes, and distinct distribution patterns characteristic of this population have not previously been described for intravenously transplanted progenitors. The ability of ALDH-expressing cells to distribute to multiple organs after transplantation provides an efficient means to deliver primitive progenitor cells to potentially mediate regenerative processes after tissue damage.

Disclosure of Potential Conflicts of Interest

  1. Top of page
  2. Abstract
  3. Introduction
  4. Materials and Methods
  5. Results
  6. Discussion
  7. Disclosure of Potential Conflicts of Interest
  8. Acknowledgements
  9. References

The authors indicate no potential conflicts of interest.

Acknowledgements

  1. Top of page
  2. Abstract
  3. Introduction
  4. Materials and Methods
  5. Results
  6. Discussion
  7. Disclosure of Potential Conflicts of Interest
  8. Acknowledgements
  9. References

We thank Jackie Hughes for flow cytometer operation, Marie Roberts for assistance with GUSB staining, and Krysta Levac for critical review of the manuscript. D.A.H. designed and performed research, analyzed data, and wrote the paper. T.P.C. performed research and wrote the paper. L.W. performed research. S.H. performed research. W.C.E. performed fluorescence-activated cell sorting. M.H.C. provided umbilical cord blood samples. M.S.S. designed research, provided NOD/SCID/MPSVII mice, and edited the manuscript. J.A.N. designed research and edited the manuscript. This work was supported by National Institute of Diabetes and Digestive and Kidney Diseases (NIDDK) Grant R01DK61848-01 and National Heart, Lung, and Blood Institute Grant R01-HL073256-01 (to J.A.N.), by NIDDK Grant R01-DK57586 (to M.S.S.), by a grant from the Krembil Foundation, and by Juvenile Diabetes Research Foundation Regeneration of β-cell Function Initiative Grant 1-2005-1173 (to D.A.H.). D.A.H. is currently affiliated with the Department of Physiology and Pharmacology, University of Western Ontario, London, Ontario, Canada, and the Vascular Biology Group, Krembil Centre for Stem Cell Biology, Robarts Research Institute, London, Ontario, Canada; J.A.N. is currently affiliated with the Stem Cell Program, University of California at Davis, Sacramento, CA.

References

  1. Top of page
  2. Abstract
  3. Introduction
  4. Materials and Methods
  5. Results
  6. Discussion
  7. Disclosure of Potential Conflicts of Interest
  8. Acknowledgements
  9. References