SEARCH

SEARCH BY CITATION

Keywords:

  • Bone marrow transplantation;
  • Cellular therapy;
  • Busulfan;
  • Cystic fibrosis transmembrane conductance regulator

Abstract

  1. Top of page
  2. Abstract
  3. Introduction
  4. Materials and Methods
  5. Results
  6. Discussion
  7. Conclusion
  8. Disclosures
  9. Acknowledgements
  10. References

Bone marrow-derived cells (BMDCs) can engraft as epithelial cells throughout the body, including in the lung, liver, and gastrointestinal (GI) tract following transplantation into lethally irradiated adult recipients. Except for rare disease models in which marrow-derived epithelial cells have a survival advantage over endogenous cells, the currently attained levels of epithelial engraftment of BMDCs are too low to be of therapeutic benefit. Here we tested whether the degree of bone marrow to epithelial engraftment would be higher if bone marrow transplantation (BMT) were performed on 1-day-old mice, when tissues are undergoing rapid growth and remodeling. BMT into newborn mice after multiple different regimens allowed for robust hematopoietic engraftment, as well as the development of rare donor-derived epithelial cells in the GI tract and lung but not in the liver. The highest epithelial engraftment (0.02%) was obtained in mice that received a preparative regimen of two doses of busulfan in utero. When BMDCs were transplanted into myelosuppressed newborn mice that lacked expression of the cystic fibrosis transmembrane conductance regulator (CFTR) protein, the chloride channel that is not functional in patients with cystic fibrosis, the engrafted mice showed partial restoration of CFTR channel activity, suggesting that marrow-derived epithelial cells in the GI tract were functional. However, BMT into newborn mice, regardless of the myeloablative regimen used, did not increase the number of bone marrow-derived epithelial cells over that which occurs after BMT into lethally irradiated adult mice.


Introduction

  1. Top of page
  2. Abstract
  3. Introduction
  4. Materials and Methods
  5. Results
  6. Discussion
  7. Conclusion
  8. Disclosures
  9. Acknowledgements
  10. References

We and others have shown that bone marrow (BM)-derived cells can differentiate into epithelial cells of the lung, liver, gastrointestinal (GI) tract, kidney, and skin following BM transplantation (BMT) into myeloablated adult recipients [1, [2], [3]4]. Other than fumarylacetoacetate hydrolase (FAH−/−) mice, in which wild-type (WT) BM-derived hepatocytes have a survival advantage and can clonally expand [4], the levels of BM-derived epithelial cells are typically less than 1% [2] and therefore are too low to be clinically relevant [5]. Several studies suggest that tissue injury promotes the migration and differentiation of BM-derived cells (BMDCs) into functional epithelia [5, [6], [7]8].

Here we tested whether engraftment of BMDCs as epithelial cells in the lung, GI tract, and liver would be higher if BMT were performed in newborn mice, when tissues are undergoing tremendous expansion. This hypothesis is supported by recent data showing that the kinetics of BM to pancreatic epithelial engraftment is different when mice are transplanted as newborns versus as adults [9]. Transplantation in utero of mesenchymal stem cells (MSCs) into mice [10], sheep [11, [12]13], and humans [14] leads to widespread and long-term engraftment of MSC progeny throughout the body. These MSCs differentiate into multiple mature cell types in the developing organs, including hepatocytes, osteocytes, skeletal myocytes, and cardiomyocytes [13]. Also, BMT into newborns has been reported to increase the engraftment of BM-derived neurons in the brains of mice [15] and humans (transplanted at 9 months of age) [16].

Based on prior studies in which mice were transplanted at birth [17], we first determined the optimal preparative regimen for hematopoietic engraftment after BMT into WT newborn mice. We then determined the levels of epithelial engraftment in multiple tissues of these mice. To assess the functionality of the BM-derived epithelial cells, WT BMDCs were transplanted into newborn mice lacking the cystic fibrosis transmembrane conductance (CFTR) gene (CFTR−/−). Lack of activity of CFTR, which is a chloride transporter expressed on the apical membrane of epithelial cells, causes cystic fibrosis (CF) in humans [18]. In addition to the importance of evaluating BMT as a possible approach for ameliorating CF pathophysiology, CFTR−/− mice provide an excellent model for testing epithelial function, as CFTR activity can be monitored by both in vitro and in vivo electrophysiological tests. We have previously demonstrated that BMT of wild-type (CFTR-positive) BM into irradiated adult CFTR−/− recipients results in BM-derived epithelial cells that provide CFTR activity in the GI tract and nasal epithelium [5]. Here, we test whether BMT at birth would confer even greater rescue to the CFTR−/− recipients.

Materials and Methods

  1. Top of page
  2. Abstract
  3. Introduction
  4. Materials and Methods
  5. Results
  6. Discussion
  7. Conclusion
  8. Disclosures
  9. Acknowledgements
  10. References

Mouse Colonies

Transgenic CFTR−/− (B6.129P2-Cftrtm1Unc) mice [19], B6.SJL-Ptprca Pep3b/BoyJ mice, C57Bl/6-Tg(ACTB-enhanced green fluorescent protein [EGFP])10sb/J mice that express GFP on the chicken β-actin promoter, and C57Bl/6J mice were purchased from the Jackson Laboratory (Bar Harbor, ME, http://www.jax.org), bred in the Yale University Animal Facility, and genotyped with standard protocols. To allow the animals to reach adulthood, CFTR−/− mice were fed Harlan Teklad 9F food (Harlan Teklad, Madison, WI, http://www.teklad.com) and drinking water supplemented with 17.5 g/250 ml Colyte (Schwarz Pharma, Milwaukee, WI, http://www.schwarzusa.com). Mice that underwent BMT were fed a liquid diet (Peptamen; Nestle, Deerfield, IL, http://www.nestleusa.com) as previously described [20]. All procedures were performed in compliance with relevant laws and institutional guidelines and were approved by the Yale University Institutional Animal Care and Use Committee.

BM Isolation, Transplantation, and Engraftment

BM was harvested as previously described [21] and resuspended in phosphate-buffered saline (PBS) at 0.5–1 × 107 cells per 50–100 μl. Mice were myeloablated with one of four different myeloablative regimens as previously described [17]. Regimen 1: For double exposure to busulfan, 15 mg/kg busulfan was administered via intraperitoneal injection in 0.02% dimethyl sulfoxide (DMSO) (B-2635; Sigma-Aldrich, St. Louis, http://www.sigmaaldrich.com) to pregnant females on days 17 and 18 of pregnancy. Regimen 2: For single exposure to busulfan and total body irradiation (TBI), busulfan was injected i.p. to pregnant females on day 19 of pregnancy, and 1-day-old pups were exposed to 400 cGy using a cesium irradiator. Regimens 3 and 4: For TBI, 1-day-old pups were exposed to 400 or 750 cGy. Approximately 12 hours after birth (and after TBI), the pups were tail-clipped for genotyping (for the CFTR−/− colony) and injected via the temporal vein with 50–100 μl of male green fluorescent protein (GFP)-positive or CD45.1 whole BM cells (0.5–1 × 107 BMDCs per pup). For analysis of hematopoietic engraftment, GFP-positive cells or CD45.1 cells from blood and BM were assessed by flow cytometry using standard protocols. For analysis of nonhematopoietic engraftment, donor-derived cells were tracked by the combination of Y chromosome fluorescence in situ hybridization (Y FISH) and immunofluorescence for different cell type-specific markers. In all experiments, BM donor mice were male and transgenic for EGFP under the β-actin promoter. Recipient mice were females, except in the experiments performed with CFTR−/− mice, in which three of nine recipients were male. In this case the donor-derived cells were assessed by GFP expression.

Complete Blood Cell Counts

On days 1 and 14 of life, 20–50 μl of blood was obtained by tail clipping. For analysis on days 30 and 80, blood was collected from the retro-orbital sinus. Blood was diluted 1:10 in PBS and white blood cell (WBC), red blood cell (RBC), and platelet (PLT) counts were performed with an automated analyzer using mouse blood cell parameters (Baker 9110+; BioChem Immuno Systems, Allentown, PA, http://www2.siri.org/msds/f2/bzg/bzgbh.html).

Electrophysiology

The rectal potential difference (RPD) was assessed as described previously [5, 22]. Briefly, mice were anesthetized with 110 mg/kg ketamine and 10 mg/kg xylazine. The RPD was sensed with a 3 M KCl-agar bridge inserted ∼0.5 cm into the rectum of the mice and connected through an Ag-AgCl electrode to a digital voltmeter. The potential difference (PD) was measured with perfusion of two consecutive solutions: 1) Cl-free/5 × 10−3 M barium/10−4 M amiloride for 15 minutes and 2) Cl-free/5 × 10−3 M barium/10−4 M amiloride solution supplemented with 10−5 M forskolin for an additional 15 minutes.

Short-circuit current and potential difference measurements were obtained by Ussing chamber analysis as described [5, 23]. Briefly, the GI tract was removed, cut longitudinally, washed, and mounted in the Ussing chamber with an aperture size of 0.3 cm2 (VCC MC2 multichannel voltage-current clamp; Physiology Instruments, San Diego, http://www.physiologicinstruments.com). Both surfaces (serosal/apical) were equilibrated in Krebs bicarbonate Ringer's solution for 20 minutes, and baseline PD values were recorded. Then, the apical membrane of the tissue was exposed to Cl-free/5 × 10−3 M barium/10−4 M amiloride bath solution for 15 minutes. Then, we added 10−5 M forskolin bilaterally, and the PD was recorded for 15 minutes. For the CFTRinh-172 inhibitor studies, after the tissue reached the maximum response to the forskolin (8–10 minutes), 10−5 M CFTRinh-172 was added bilaterally.

RNA Isolation and CFTR mRNA Analysis

After DNase I-RNase-free treatment (Roche Applied Science, Indianapolis, http://www.roche-applied-science.com), 2 μg of total RNA was reverse-transcribed using Superscript II RNase H reverse transcriptase (Invitrogen, Grand Island, NY, http://www.invitrogen.com) with 100 ng of random hexamers. Two μl of cDNA were amplified with primers mCF7 (5′-CAGTGGAAGAGTTTCATTCTG-3′) and mCF12 (5′-CCTTCTCCAAGAACTGTGTTGTC-3′) from exons 10 and 11 of the murine CFTR mRNA, respectively. As a cDNA quality control, we amplified β-actin cDNA using primers spanning exons 2 and 3. Reactions without reverse transcriptase were performed as controls for RNA purity. β-Actin and CFTR polymerase chain reaction (PCR) were performed with 30 and 40 cycles of PCR, respectively. Primers were designed to avoid amplification of the CFTR mRNA that is transcribed from the knockout CFTR−/− allele.

Y FISH and Immunofluorescence on Tissue Sections

Isotype, serum, and no-primary-antibody controls were included for each sample in the staining protocols. Negative and positive control tissues were processed in each staining run. For Y FISH/CD45/CDX2 and FISH/CD45/cytokeratin, deparaffinized 3-μm sections were incubated in BD Retrievagen A solution (Becton, Dickinson and Company, San Diego, CA, http://www.bd.com) for 15 minutes at 100°C and then 20 minutes at room temperature. Slides were washed, incubated in 0.2 N HCl for 12 minutes and 1 M NaSCN for 20 minutes at 80°C, and Y FISH was performed as described [24]. Slides were washed and incubated in 1:20 anti-CD45RB (Santa Cruz Biotechnology Inc., Santa Cruz, CA, http://www.scbt.com) mixed with 1:100 F4/80 (eBiosciences, San Diego, CA, http://www.ebioscience.com) for 1 hour at room temperature, washed, and incubated with anti-rat alexa 647 (Molecular Probes Inc., Eugene, OR, http://probes.invitrogen.com).

For CDX2 staining, the slides were incubated with avidin/biotin block (Vector Laboratories, Burlingame, CA, http://www.vectorlabs.com) for 10 minutes at room temperature, washed, and incubated with anti-CDX2 (BioGenex, San Ramon, CA, http://www.biogenex.com) diluted 1:50 for 1 hour at room temperature using a mouse-on-mouse kit (Vector Laboratories), and signal was detected with streptavidin-fluorescein isothiocyanate (FITC) (15 minutes at room temperature). For the pan-cytokeratin (CK) staining, slides were washed, fixed in 2% buffered paraformaldehyde for 8 minutes, digested with 0.125% trypsin, for 1 minutes at 37°C, washed with 5% fetal calf serum in PBS to inactivate the trypsin, and incubated with 1:200 anti-pankeratin (Dako, Carpinteria, CA, http://www.dako.com) or 1:200 anti-CK18 (kind gift from Scott Randell, University of North Carolina) overnight at 4°C. Slides were washed, incubated in 1:500 anti-rabbit-FITC (Molecular Probes) for 1 hour at 37°C, and coverslipped using Vectashield containing 4,6-diamidino-2-phenylindole (Vector Laboratories). For CFTR staining, deparaffinized slides were biotin-blocked, incubated in 3% goat serum, incubated in 1:200 anti-CFTR antibody (R3195; kind gift of Dr. Christopher Marino, University of Tennessee, Memphis, TN) overnight at 4°C, incubated in 1:100 biotinylated anti-rabbit antibody (Chemicon, Temecula, CA, http://www.chemicon.com), and detected with 1:500 streptavidin-Texas Red (Molecular Probes). For each tissue, at least three slides were analyzed.

Statistical Analysis

Statistical analyses were conducted via paired t tests, with p values of less than .05 (two-tailed) used to reject the null hypothesis.

Results

  1. Top of page
  2. Abstract
  3. Introduction
  4. Materials and Methods
  5. Results
  6. Discussion
  7. Conclusion
  8. Disclosures
  9. Acknowledgements
  10. References

Testing Different Myeloablative Methods for BMT in Newborn Mice

Myeloablative conditioning methods are well described for adult mice, but protocols for myelosuppression of newborn mice are not as well-established. Based on a pioneering study performed by Yoder et al. [17], we investigated the effects of in utero busulfan exposure and/or TBI of day 1 newborns (Tables 1, 2). As described in Materials and Methods, we tested four different myeloablative regimens: Regimen 1, two doses of busulfan (Bu/Bu); Regimen 2, one dose of busulfan, followed by 400 cGy of TBI (Bu/TBI); Regimen 3, 400 cGy of TBI (TBI 400); and Regimen 4, 750 cGy of TBI (TBI 750). Control mice received no conditioning. Approximately half of the newborns from each group underwent BMT (Tx). To assess the relative myeloablative effects of the described regimens, complete blood cell counts (CBCs) were obtained on days 1, 14, 20–30, and 60–80 after birth.

Table Table 1.. Assessing newborn myeloablation via complete blood cell counts
Thumbnail image of
Table Table 2.. Assessing hematopoietic engraftment via percentage of chimerism
Thumbnail image of

As seen in Table 1, there was no difference in the CBC between control mice and untransplanted Bu/Bu mice. In contrast, untransplanted Bu/TBI mice had a significant decrease in the day 1 WBC and PLT counts (Table 1). These mice showed a failure to thrive, with poor food intake, and were sacrificed prior to day 14 of life. These data suggest that, unlike Bu/Bu, Bu/TBI is very toxic. There was no significant difference in the CBC of transplanted Bu/Bu mice compared with untransplanted control or Bu/Bu mice. However, even though the blood counts were not significantly affected, the Bu/Bu regimen allowed for significant hematopoietic engraftment as assessed by flow cytometry of recipient peripheral blood (PB) and BM (40–41%; n = 4; Table 2). In transplanted Bu/TBI mice, the WBC was nearly normal at all times tested, and the mean RBC was only slightly lower than normal at day 14. Despite the fact that the PLT was significantly lower than controls, mice in this group achieved a safe PLT count of 4.3 × 105 cells per microliter within 2 weeks, and adequate platelet numbers were maintained long-term, reaching normal levels 60–80 days post-transplant (Table 1). Therefore, BMT after Bu/TBI prevented the lethal toxicity seen in the untransplanted group. Moreover, the Bu/TBI regimen showed the highest levels of hematopoietic engraftment (81–89%; n = 6; Table 2). Although the animals survived Bu/TBI if transplanted at day 1, transplanted Bu/TBI mice had a body weight that was 65% of that of transplanted Bu/Bu mice at day 14, and the difference in weight became more pronounced after 2 months (56%). In all other respects, these mice appeared healthy.

Hematopoietic engraftment in the PB at different times and in the BM at the time of sacrifice were also assessed in mice that received only 400 cGy of TBI (n = 3) or 750 cGy of TBI (n = 4) (Table 2). TBI 400 mice had high PB (76%) and BM (82%) engraftment at 80–100 days post-transplant. As in the transplanted Bu/TBI mice, TBI alone led to impaired growth. Although Bu/TBI without BMT was lethal, untransplanted TBI 400 mice survived to adulthood. However, some of these untransplanted mice developed limb tremors, suggesting that TBI alone causes growth reduction and, in some cases, CNS damage as reported previously for newborn mice exposed to irradiation [25].

TBI 750 mice also had high PB (62%) and BM engraftment (81%) at day 25. However, both untransplanted and transplanted mice were sacrificed by days 14 and 30, respectively, due to failure to thrive, with little weight gain, poor food intake, tremors, and poor neuromuscular coordination. CBC of untransplanted TBI 750 mice obtained at the time of sacrifice (14 days) revealed decreased WBC, RBC, and PLT counts compared with control mice. Thus, 750 cGy is highly toxic for 1-day-old pups [17, 25]. When BM was transplanted into 1-day-old mice that had received no preconditioning, there was 1% hematopoietic engraftment on average (Table 2).

BM-Derived Epithelial Cells in the GI Tract and Lung of Transplanted Pups

Eighty to 100 days post-BMT, the GI tract, lungs, and liver of transplanted mice from each group were analyzed for BM-derived epithelial cells. Donor-derived lung epithelial cells and hepatocytes were assessed by costaining for the Y chromosome (donor origin) and CK18 (epithelial cells). The lack of staining for CD45 (a hematopoietic cell marker) served as additional confirmation of the epithelial phenotype. Donor-derived GI epithelial cells were costained for the presence of the Y chromosome and the GI-specific transcription factor CDX2 and again confirmed by their lack of CD45 expression. Figure 1A and 1B shows representative images in which BM-derived epithelial cells were found (arrows) for GI tract (Y-chromosome-positive/CDX2-positive/CD45-negative) and lung (Y-chromosome-positive/CK18-positive/CD45-negative), respectively. From each experimental tissue, the frequency of donor-derived epithelial cells in the lung and GI tract was assessed by counting at least 4 × 104 CDX2-positive/CD45-negative (∼80% of total cells in the small GI tract) or CK18 positive CD45-negative (∼ 54% of the total cells in the lung) and determining the percentage of these that were donor-derived. The frequency of donor-derived cells from each mouse was express as the percentage of CDX2-positive or CK18-positive cells that were also positive for the Y-chromosome (or GFP) and negative for CD45. Table 3 reports the mean BM-derived epithelial cell frequencies for each experimental group. Epithelial donor-derived cells were detected in the small GI tract of at least one animal in all of the experimental groups analyzed (Table 3). No BM-derived epithelial cells were detected in the liver of transplanted mice under any of the experimental conditions tested.

thumbnail image

Figure Figure 1.. Assessing epithelial engraftment by Y-chromosome/CDX2/CD45 immunofluorescence and Y-chromosome/cytokeratin/CD45 immunofluorescence. (A): Staining for Y chromosome (red), CDX2 (green), and CD45 (magenta) in intestinal tissue of a female bone marrow transplantation (BMT) recipient. An arrow points to a Y positive, CDX2 positive epithelial cell. Note that the majority of blood cells in the lumen are endogenous, as they do not stain for the Y chromosome. (B): Staining for Y chromosome (red), cytokeratin-18 (green), and CD45 (magenta) in lung tissue of a female BMT recipient. An arrow points to a Y-positive, cytokeratin-positive, CD45-negative epithelial cell in the lung alveolus. Arrowheads point to Y-positive, CD45-positive donor-derived blood cells in the lung. (C, D): Staining for Y chromosome (red) and cytokeratin (green) shows bone marrow-derived epithelial cells in the jejunum (C) and colon (D) of transplanted female cystic fibrosis transmembrane conductance regulator (CFTR)−/− mice. Arrows indicate Y chromosome-positive, cytokeratin-positive, CD45-negative donor-derived epithelial cells, whereas arrowheads show Y chromosome-positive, CD45-positive blood cells in the jejunum (C) and colon (D) of BMT recipients. (E, F): Immunofluorescence for CFTR (red). Isolated CFTR high-expressing cells were found in wild-type mice (E) and transplanted wild-type into cystic fibrosis (CF) mice (F) but not in transplanted CF into CF mice (data not shown). For all images, nuclei are stained with 4,6-diamidino-2-phenylindole (blue).

Download figure to PowerPoint

Table Table 3.. Nonhematopoietic engraftment as assessed by immunofluorescence
Thumbnail image of

Unexpectedly, rare donor-derived epithelial cells were also detected in the GI tract and lung in one of four mice that had been transplanted without any myelosuppression, with a frequency of 0.002%. In the small GI tract of Bu/TBI or TBI 400 mice, the frequency of donor-derived epithelial cells was also very low (0.001%), and only one of five and one of three mice, respectively, had BM-derived epithelial cells. In contrast, Bu/Bu mice had the highest occurrence of donor-derived epithelial cells in the small GI tract (six of eight mice), and the frequency of BM-derived epithelial cells was 10-fold higher then other experimental groups (0.01%). In addition, this was the only experimental group in which we found rare donor-derived epithelial cells in the large GI tract (one of eight mice). Similar data were obtained for the lung tissues, in which the highest occurrence of BM-derived epithelial cells was in the Bu/Bu mice (five of seven mice), and the frequency of donor-derived epithelial cells was 0.02% (Table 3).

Functional CFTR Is Expressed in CFTR−/− Mice Transplanted as Newborns

To test whether the donor-derived GI epithelial cells were functional, and to provide an additional epithelial specific protein for detection, we transplanted WT(CFTR+) male GFP+ BM into newborn CFTR−/− mice (three male and six female). Using in vivo and in vitro CFTR functional assays, we showed previously that marrow-derived epithelial cells express functional CFTR after BMT of adult CFTR−/− mice [5]. We now tested whether transplantation into newborn CFTR−/− mice also restored functional CFTR channel activity.

In initial experiments, CFTR−/− mice were transplanted after Bu/TBI. However, the CFTR−/− pups did not survive with this regimen (data not shown). Therefore, in subsequent studies, CFTR−/− mice (n = 9) were myelosuppressed with the Bu/Bu protocol and transplanted on postnatal day 1. These CFTR−/− mice were monitored by in vivo rectal potential difference (RPD) at 12 and 24 weeks post-transplantation (Fig. 2A; Table 4). Untransplanted CFTR+/− mice (n = 14), which have very similar electrophysiology to wild-type mice [22, 26], and untransplanted CFTR−/− (n = 15) mice were used as positive and negative controls, respectively. CFTR activity was measured as the change (Δ) in RPD in response to exposure to forskolin [5]. Forskolin induces a hyperpolarization –5.4 ± 1.0 mV in CFTR+/− mice and causes a depolarization (the PD becomes more positive) of 3.0 ± 0.5 mV in CFTR−/− animals (Fig. 2A; Table 4). This forskolin-induced depolarization in CFTR−/− mice is due other cAMP-responsive channels and the lack of functional CFTR in the mucosa of the distal colon.

thumbnail image

Figure Figure 2.. Analyses of CFTR−/− mice transplanted as newborns. (A): Mean ± SD of forskolin-induced ΔRPD in CFTR+/− (white bar; n = 14), CFTR−/−(black bar; n = 15), CFintoCF (striped bars; n = 2), and WTintoCF mice (gray bars; n = 9) 12 and 24 weeks post-transplant, as indicated. (B): Forskolin-induced distal colon ΔIsc/cm2 for CFTR+/− (white bar; n = 7), CFTR−/− (black bar; n = 7), CFintoCF (n = 5; striped bar; n = 2) and WTintoCF mice (gray bars; n = 9) 24 weeks post-transplantation. (C): ΔIsc in response to CFTRinhibitor-172. ∗, statistically significant difference between CFintoCF and WTintoCF groups (p < .05). Abbreviations: CFintoCF, CFTR−/− mice transplanted with CFTR−/− bone marrow; CFTR, cystic fibrosis transmembrane conductance regulator; ΔIsc, change in short-circuit current; ΔRPD, change in rectal potential difference; wks, weeks; WTintoCF, CFTR−/− mice transplanted with wild-type bone marrow.

Download figure to PowerPoint

Table Table 4.. Data summary for CFTR−/− mice myeloablated in utero by busulfan and transplanted at birth
Thumbnail image of

In the CFTR−/− mice that were transplanted with WT BM (WTintoCF), forskolin induced a ΔRPD of 1.1 ± 0.2 mV at 12 weeks and 1.3 ± 0.3 mV at 24 weeks post-BMT, which represents a 1.8-mV change compared to untransplanted CFTR−/− mice (3.0 ± 0.5 mV), confirming the presence of CFTR activity in the GI tract (Fig. 2A). At both time points, the difference in PD between WTintoCF transplanted mice and untransplanted CFTR−/− mice was statistically significant (p = .0005 and p = .002, respectively). To evaluate whether Bu/Bu itself might induce electrophysiological changes, CFTR−/− mice conditioned with the Bu/Bu regiment were transplanted on postnatal day 1 with BM from CFTR−/− mice (CFintoCF; n = 2). The forskolin induced ΔRPD in CFintoCF mice was similar to that of untransplanted CF mice at each time point (2.9 ± 0.0 and 2.8 ± 0.4 mV). When the transplanted Bu/Bu CFTR−/− mice were sacrificed 5 months post-transplant, the hematopoietic engraftment in the BM ranged from 37% to 95%, with an average of 70% (Table 4).

In vitro CFTR activity in the distal colon from untransplanted CFTR+/− and CFTR−/− mice, as well as WTintoCF (n = 9) and CFintoCF (n = 2) mice, was assessed in Ussing chambers. After forskolin stimulation, the change (Δ) in short-circuit current (Isc) in the distal colon of WTintoCF mice was –1.9 ± 2.7 μAmp, which is a statistically significant increase in Isc of approximately 12 μAmp compared with untransplanted CFTR−/− mice (p = .004). In contrast, the ΔIsc from colonic tissues of CFintoCF mice was –18 ± 1.3 μAmp, which was not significantly different from untransplanted CFTR−/− mice (p = .18) (Fig. 2B; Table 4). WTintoCF mice transplanted as newborns showed restoration of approximately 1.7% of the Isc of CFTR+/− mice (mean = 99.8 μAmp/cm2). To further investigate whether the forskolin-dependent ΔIsc in transplanted mice was truly due to CFTR, in four of eight colonic tissues from WTintoCF and two from CFintoCF mice, we assessed for inhibition of this current by the highly specific thiazolidinone-type CFTR inhibitor, CFTRinh-172, which blocks CFTR activity by interacting with its NBD1 domain [27, 28]. Consistent with our previously published data [5], bilateral addition of CFTRinh-172 produced an Isc decay in CFTR+/− tissue and had no effect on tissues of CFTR−/−. Similarly, CFTRinh-172 had no effect on the PD of CFintoCF mice transplanted as newborns, but in the tissues of WTintoCF mice, the inhibitor produced either a small decay in Isc (ΔIsc/cm2 = −0.6 μAmp) or no effect (Table 4). Statistical analysis revealed that the CFTRinh-172 inhibition in WTintoCF mice was significantly different from that of CFintoCF and CFTR−/− mice (Fig. 2C), confirming that some of the forskolin induced current in WTintoCF mice is due to functional CFTR.

CFTR expression, as assessed by reverse transcription-polymerase chain reaction (RT-PCR) (Table 5), was more consistently detected in the colon (seven of nine mice) and rectum (six of nine) of WTintoCF mice than the small GI tract. In the respiratory tract, six of nine lungs, three of nine tracheas and three of three nasal epithelial samples from WTintoCF mice had detectable CFTR mRNA. CFTR is mainly express in epithelial cells. Therefore, the presence of CFTR expression in the tissues of transplanted mice supports the presence of BM-derived CFTR+ epithelial cells. To determine whether nonepithelial BM-derived cells in the blood may be responsible for the CFTR mRNA detected, RT-PCR was performed on total white blood cells from the blood. No CFTR mRNA was ever detected in the PB of any WT or transplanted mouse. However, very low expression of CFTR mRNA was detected in the BM of four of nine WTintoCF mice transplanted as newborns and some WT controls (Table 5). This expression was markedly lower than that detected in the tissues of transplanted mice. The expression of CFTR in the BM may reflect the promiscuous gene expression that is frequently observed in BM-derived stem cells [29] or the presence of rare CFTR+ epithelial cells. Interestingly, it has been suggested that a population of epithelial progenitors may “hide out” in the BM [30, 31]. No CFTR expression was detected in any tissue of either untransplanted CFTR−/− mice or the CFintoCF group.

Table Table 5.. Data summary of CFTR cDNA detection by reverse transcription-polymerase chain reaction in different tissues of transplanted CFTR−/−micea
Thumbnail image of

We next assessed for donor-derived epithelial cells by costaining for Y chromosome, cytokeratin, and CD45 (Fig. 1C, 1D; Table 6) and for CFTR (Fig. 1E, 1F; Table 6). Donor-derived epithelial cells (Y-positive/CK-positive/CD45-negative or GFP-positive/CK-positive/CD45-negative) were detected in the small GI tracts of most of (eight of nine) the mice, which is similar to what we observed for WT recipients that were myeloablated with two doses of busulfan (six of eight mice; Table 3). In contrast, a high frequency (six of eight) of transplanted CFTR−/− mice had donor-derived epithelial cells in the large GI tract, which was not observed in WT recipients who had been myeloablated with the same regimen (one of eight mice; Table 3). This suggests that there may be abnormalities in the large GI tract of CF mice that promote engraftment of BM-derived epithelial cells. In three of nine mice, donor-derived epithelial cells were detected in all tissues analyzed (Table 6). However, the frequency of these cells in the GI tract of transplanted mice was very low (approximately 1 in 10,000 to 1 in 100,000 epithelial cells), and they were not uniformly distributed. Immunofluorescent analysis for CFTR protein revealed rare GI epithelial cells expressing CFTR (Fig. 1E, 1F; Table 6). Since the donor mice are CFTR−/−, the presence of this protein, which is clearly expressed on the apical membrane of epithelial cells, provides further evidence for the presence of functional donor-derived epithelial cells.

Table Table 6.. Data summary of detection of donor-derived epithelial cells in the GI tract of transplanted CFTR−/−micea
Thumbnail image of

Discussion

  1. Top of page
  2. Abstract
  3. Introduction
  4. Materials and Methods
  5. Results
  6. Discussion
  7. Conclusion
  8. Disclosures
  9. Acknowledgements
  10. References

Although BMDCs can engraft as epithelial cells in the lung, liver, GI tract, and skin following administration into lethally irradiated recipients [3], much remains to be learned before BMT for engraftment of epithelial cells could be feasible as a therapeutic approach for diseases involving epithelial damage. The mechanisms by which these BMDCs migrate to a specific tissue and integrate as epithelial cells are still unclear, and it is not known to what extent tissue injury promotes the occurrence of BM to epithelial engraftment. Also, the degree of donor-derived cells is generally too low to be of therapeutic value (<1%) [5]. We hypothesized that epithelial engraftment requires the tissue remodeling machinery [6, 7]. Here, we tested whether the degree of BM to epithelial engraftment would be higher if BMT were performed at the newborn stage, when tissues are undergoing rapid and massive growth and remodeling, as was recently shown for pancreatic duct epithelium [9].

We compared different myeloablative regimens to determine which method is least toxic to the newborn mice while allowing hematopoietic engraftment in our hands. The level of toxicity was as follows, from highest to lowest: TBI 750 > Bu/TBI 400 > TBI 400 > Bu/Bu > no myeloablation. The scale of hematopoietic chimerism after BMT was as follows, from highest to lowest: Bu/TBI 400 > TBI 750 = TBI 400> Bu/Bu > no myeloablation. Not surprising, the highest toxicity occurred in newborns irradiated with 750 cGy, and these mice were sacrificed early post-transplant due to severe growth retardation and lack of activity, suggesting CNS damage as previously reported [17]. Just prior to sacrifice, CBCs demonstrated a decrease in the WBC, RBC, and PLT. The Bu/TBI 400 combination was toxic and impaired growth as well. However, in this case, the transplantation of BM rescued the mice. The CBC of transplanted Bu/TBI mice reached normal levels by 60–80 days post-BMT. In contrast, TBI (400 Gy) alone was sublethal. Although they had impaired growth, untransplanted mice survived long-term with recovery of their CBCs, and with BMT, the hematopoietic chimerism of these mice was quite high.

Myelosuppression with exposure to two doses of busulfan in utero was less toxic than the other regimens used. Bu/Bu mice had nearly normal growth rates, and the CBCs were almost normal whether or not BMT was performed. Our data on untransplanted Bu/Bu mice differ from previously published findings by Yoder et al., which showed that 70% of untransplanted Bu/Bu mice die after 30 days due to marrow aplasia [17]. The differences are likely due to the busulfan preparation and the route of administration. We dissolved busulfan in 0.02% DMSO, whereas Yoder et al. used 1% carboxymethylcellulose [17]. Also, we administered the cells intravenously, and they used subcutaneous injections. With the approach used here, exposure to busulfan in utero was less toxic to the newborn mice than direct injection into adult mice [32], allowing long-term survival and high hematopoietic chimerism without toxicity. Low-level (1%) long-term, hematopoietic engraftment was also attained after transplantation of 1 × 107 BM cells into nonmyeloablated 1-day-old pups. These data are consistent with previous data showing that in nonmyeloablated mice, the percentage of engraftment is approximately 0.17% per 1,000,000 cells infused [33].

BM-derived epithelial cells were rare and isolated in all experimental groups, suggesting that BMDCs do not engraft as GI tract or lung stem cells. Due to the rarity of these cells, it was difficult to rigorously compare the engraftment levels between the groups. In general, the frequency of these cells was between 0.001% (Bu/TBI, TBI 400, and control mice) and 0.02% (Bu/Bu). Consistent with the higher frequency of epithelial engraftment, BM-derived epithelial cells were present in almost all mice that received Bu/Bu myeloablation, whereas for the other experimental groups, donor-derived epithelial cells were found in a lower percentage of the mice. Our data contrast with recent work by Wang et al. [9], who demonstrated that BMT into newborn mice increases the contribution of BM cells to the pancreatic β cells. Therefore, our data suggest that this is not true for all organs.

Surprisingly, the frequency of marrow-derived epithelial cells did not correlate strictly with the degree of toxicity or with hematopoietic engraftment. Mice myelosuppressed with busulfan had lower hematopoietic engraftment than mice myeloablated with regimens containing TBI, and one of four mice from the nonmyeloablated group, in which hematopoietic engraftment was extremely low (1%), had donor-derived epithelial cells in the small GI tract and lung. These data are provocative, since they suggest that the cells engrafting as epithelial cells may not need to first engraft in the BM. Loi et al. [34] have shown that plastic-adherent MSCs administered to naïve nonirradiated (and therefore not myeloablated) mice resulted in engraftment of donor-derived airway epithelial cells of approximately 0.025%, which is similar to what was found when mice were myeloablated and transplanted with WBM after total body irradiation. Therefore, it is plausible that hematopoietic engraftment is not necessary for derivation of epithelial cells and that MSCs may be the population of cells in the BM able to differentiate into epithelial cells. In addition, the data suggest that tissue damage may not be necessary for the development of epithelial cells from BMDCs, at least when transplantation is performed in newborns.

Different studies have reported a wide range of BM-derived epithelial cell frequency in the lung, from none [35] up to 20% [3]. Some of these differences are due to differences in the experimental methods used, including, for example, the subpopulation of BM donor cells, mouse models, and age of donors and recipients. However, the method used for the detection of the donor-derived epithelial cells is most likely the primary reason underlying the apparent discrepancies. First, in our hands, use of FISH for the Y chromosome in thin sections can overestimate, and GFP detection can underestimate, the frequency of BM to epithelial engraftment. (Of note, some investigators report that Y FISH can underestimate the number of marrow-derived cells [36].) Second, issues of overlay plague many of the approaches used. Improved methods for ruling out CD45-positive cells and ruling in epithelial cells in addition to use of confocal or deconvolution microscopy will reconcile these discrepancies.

No BM-derived hepatocytes were detected using any of the preparative regimens tested, suggesting that the liver microenvironment may not be permissive for the development of marrow-derived hepatocytes when the cells are transplanted into newborn mice.

In our electrophysiology studies to assess the function of BM-derived epithelial cells in the GI tract, BMT into newborn CFTR−/− mice yielded data very similar to what we observed in adult CFTR−/− BMT recipients [5]. BM-derived epithelial cells can establish chloride transport across the epithelia even when the level of gut engraftment is 1 in 10,000 to 1 in 100,000 cells. One possible explanation for this unexpectedly high level of CFTR activity comes from the nature of the CFTR channel activity itself. It is well documented that 5% of normal levels of CFTR expression is sufficient to restore levels of chloride transport to those of WT tissues [37, [38]39]. Therefore, there is not a linear correlation between the number of CFTR-expressing cells and the amount of chloride transport. This concept is also supported by a recent study of the electrophysiological characteristics of epithelia generated from mixtures of CFTR+/+ and CFTR−/− human airway cells in vitro [39]. Our data also show that the restoration of the chloride activity in transplanted CFTR−/− is slightly higher when the electrophysiological measurements are performed in vivo rather than in vitro. We believe that during the in vivo perfusion, in which the probe is inserted 0.5 cm into the colon of the mouse, we are sampling an extended area of tissue and therefore an extremely high number of cells. In contrast, during the in vitro electrophysiological measurements, we analyze an area of only 0.3 cm2. An additional explanation could be that using the Y chromosome approach as a host-specific marker may underestimate the extent of BM-derived cells, as recently suggested by Rizvi et al. [36].

The level of restoration of CFTR-dependent chloride secretion was very similar between CFTR−/− mice transplanted as newborns or adults. Using CFTR−/− mice that had been myelosuppressed and transplanted with CFTR−/− BM as controls, we excluded the possibility that detection of CFTR activity was due to the transplantation process rather than the presence of epithelial cells expressing functional CFTR.

A recent study suggests that the mechanism by which BMDCs populate the epithelial compartment of the GI tract is by cell-cell fusion [36]. However, this is a controversial observation, since other studies failed to detect fusion in the GI tract [40, 41]. Our study was not design to assess fusion; therefore, we cannot exclude this possibility.

Conclusion

  1. Top of page
  2. Abstract
  3. Introduction
  4. Materials and Methods
  5. Results
  6. Discussion
  7. Conclusion
  8. Disclosures
  9. Acknowledgements
  10. References

Factors that can affect epithelial derivation from BMDCs are not known. We postulated that BM transplantation into newborn mice could increase the level of epithelial engraftment as compared with that in adult recipients. We used different myeloablative regimens associated with different degrees of hematopoietic engraftment. Our data show that although BMT in newborn mice gave rise to rare donor-derived epithelial cells in the GI tract and lung, transplantation into newborns, regardless of the preparative regimen used, did not lead to significantly higher levels of BM-derived epithelial cells than in mice transplanted in adulthood. This was confirmed at the functional level, when transplantation was performed in newborn CFTR−/− mice, in which the BMT re-established some chloride transport across the epithelia. At least in newborns, the level of hematopoietic engraftment, as well as the toxicity of the myelosuppressive regimens, did not correlate with the degree of epithelial engraftment. Therefore, we conclude that BMDCs can engraft as functional epithelial cells during the physiological growth and expansion of tissues in newborn mice and that these levels are similar to those observed during the tissue repair process in adult life.

Acknowledgements

  1. Top of page
  2. Abstract
  3. Introduction
  4. Materials and Methods
  5. Results
  6. Discussion
  7. Conclusion
  8. Disclosures
  9. Acknowledgements
  10. References

We thank Buqu Hu, Stephanie Donaldson, Elisa Ferreira, and Christina Caputo for technical assistance; Dr. Christopher Marino for providing the R3195 anti-CFTR antibody; Dr. Scott Randell for providing the anti-CK-18 antibody; Dr. Alan Verkman for providing the CFTRinh-172 blocker; Dr. Michael Speicher for providing the murine Y chromosome template; and Dr. Neil Theise, Dr. Li Chai, Gino Galietta, and Scott Johnson for helpful discussions. This work was supported by NIH Grants DK61846, HL073742, and DK053428.

References

  1. Top of page
  2. Abstract
  3. Introduction
  4. Materials and Methods
  5. Results
  6. Discussion
  7. Conclusion
  8. Disclosures
  9. Acknowledgements
  10. References