Canine cells engraft into regenerating mouse muscle
The lower right hindlimb of each NOD/SCID mouse was exposed to 12 Gy of ionizing radiation, the lowest dose that prevented host muscle regeneration (data not shown). The tibialis anterior (TA) muscle of the same hindlimb was injected with barium chloride to induce muscle regeneration, and the following day, mononuclear cells isolated from a wild-type canine muscle biopsy were injected directly into the same TA muscle, along the length of the muscle, from the proximal to the distal end. The injected muscle was harvested 28 days after injection, and cryosections were immunostained using antibodies specific for canine dystrophin and canine lamin A/C. Muscle injected with 1 × 104 (Figure and ) or 4 × 104 (Figure and ) canine cells displayed a significant number of nuclei expressing canine lamin A/C (Figure and ) and fibers expressing canine dystrophin (Figure and ), indicating canine donor cell engraftment.
Figure 1 Canine muscle cell engraftment into mouse muscle is quantifiable and consistent. (A) Cryosections from NOD/SCID mouse muscle injected with 1 × 104 (a, b) or 4 × 104 (c, d) canine muscle-derived mononuclear cells were immunostained with (more ...)
Canine donor cell engraftment is quantifiable and consistent
NOD/SCID mice were injected as described above, with freshly isolated canine muscle-derived mononuclear cells from three different donor dogs, with cell doses ranging from 2 × 103 to 5 × 104 cells per injection. The number of fibers expressing canine dystrophin and the number of nuclei expressing canine lamin A/C per cross-section of muscle were determined using a minimum of three cross-sections from each injected muscle, from within the region of highest engraftment and covering a distance of approximately 800 to 1200 μm. Muscles from at least two separate experiments were analyzed; each experiment representing a single canine muscle biopsy and cell preparation, and a minimum of three mice for each cell dose.
Initially, we expressed canine donor cell engraftment as a percentage of total fibers expressing canine dystrophin or a percentage of total nuclei expressing canine lamin A/C. However, the amount of host mouse muscle remaining after BaCl2
induced degeneration was not consistent between recipients, resulting in variable values when expressing engraftment as a percentage. However, there was a reproducible positive correlation between the number of cells injected from canine donor G982 and the number of fibers expressing canine dystrophin per cross-section (Figure ) and the number of nuclei expressing canine lamin A/C per cross-section (Figure ). The correlation between cell dose and engraftment was also observed for canine donors G993 and H299 (Additional file 1
). A logarithmic curve was determined to be the best fit curve for the data shown in Figures and , and S1 as the r2
value was more than 0.95 for all curves shown. Moreover, a linear relationship between the number of donor nuclei, as determined by expression of canine lamin A/C, and the number of fibers expressing canine dystrophin, was seen for all donors, with an average of approximately 1.75 ± 0.07 canine nuclei per myofiber expressing canine dystrophin per cross-section (Figure , Additional file 1
Notably, each donor's muscle-derived cell population had a different capacity for reconstitution as measured by the number of fibers expressing canine dystrophin, the number of nuclei expressing canine lamin A/C, and the number of nuclei expressing Pax7 and canine lamin A/C. Yet, different muscle cell preparations from the same donor displayed similar levels of engraftment (Figures and , Additional file 1
). Therefore, the canine-to-murine xenotransplantation model provides a sufficiently robust platform to quantitatively assess engraftment potential of different populations of canine muscle cells.
Canine donor cells consistently engraft to the murine satellite cell niche
Nuclei expressing Pax7 and canine lamin A/C were detected at the outer periphery of muscle fibers and underneath laminin of the extracellular matrix, suggesting that canine cells had engrafted into the niche normally occupied by murine satellite cells (Figure ). In addition, a small number of Pax7-positive nuclei not expressing lamin A/C were detected, indicating that irradiated and injected muscles maintained a minimal population of murine satellite cells (data not shown). Importantly, the number of canine-derived Pax7-positive cells increased with the number of donor cells injected, and engraftment at all cell doses was consistent for each donor (Figure , data not shown).
Figure 2 Canine derived Pax7-positive cells persist in injected mouse muscle. (A) Cryosections from mouse muscle injected with canine muscle-derived mononuclear cells were immunostained with anti-Pax7 (green), anti-lamin A/C (red), and anti-laminin (blue). (B) (more ...)
To confirm that mononuclear cells expressing canine lamin A/C present in transplanted muscle were maintained in the myogenic lineage, mouse TA muscles harvested 28 days after canine muscle cell injection were digested with collagenase, and the resulting mononuclear cells were cultured in growth medium containing 20% fetal bovine serum and 2.5 ng/ml bFGF for 7 to 10 days. Cells were fixed and stained for expression of canine lamin A/C to detect cells of donor canine origin, and for Pax7 and myogenin to detect all myogenic cells. We chose to stain cells for expression of both Pax7 and myogenin to include both undifferentiated and differentiated muscle cells. Cells expressing Pax7 or myogenin isolated and cultured from muscle injected with 2,000 or 10,000 canine donor cells were exclusively canine lamin A/C-positive, indicating that a subpopulation of transplanted canine cells was capable of generating mononuclear muscle cells in vitro (Figure ). Muscles injected with 400 canine donor cells did not yield detectable numbers of canine muscle cells in vitro, perhaps due to the small number of donor cells engrafted.
A subpopulation of canine muscle-derived cells express CXCR4
Recently, a subpopulation of satellite cells expressing CXCR4 have been identified in mice, and cells sorted for expression of CXCR4 and β1-integrin efficiently engraft into regenerating muscle of mdx
]. We detected expression of CXCR4 transcript in cultured proliferating and differentiating canine satellite cell derived myoblasts, and the freshly isolated population of mixed canine muscle-derived cells (Figure ). SDF-1, the sole ligand for CXCR4, was detected in RNA from whole canine skeletal muscle and the freshly isolated population of mixed canine muscle-derived cells, but not cultured myoblasts, suggesting that satellite cells or another population of cells were the source of SDF-1. Notably, Pax7 transcript, but not MyoD transcript, was present in freshly isolated muscle-derived cells, indicating that this population of cells has a more progenitor-like phenotype and has not initiated myogenic commitment and differentiation.
Figure 3 Sorting for CXCR4-positive cells does not enhance engraftment. (A) RNA isolated from whole canine skeletal muscle (SkM), proliferating canine satellite cell-derived myoblasts (PrMbs), differentiating myoblasts (DfMbs), and freshly isolated canine muscle-derived (more ...)
FACS sorting of cultured canine proliferating primary myoblasts demonstrated that 98.5% of viable cells were CXCR4-positive, whereas 88% of viable cells were β1-integrin-positive (data not shown). We sorted freshly isolated canine muscle-derived cells from various donors and observed that approximately 1 to 3.5% of canine muscle-derived mononuclear cells were CXCR4-positive, and less than 0.5% of cells were CD45-positive; however, we were unable to detect a significant population of β1-integrin-positive cells (Figure , data not shown).
Sorting for CXCR4-positive cells does not increase engraftment efficiency
NOD/SCID mice were injected with freshly isolated mixed canine muscle-derived cells and CXCR4-positive cells sorted from the mixed cell population. Injection of 1 × 104 CXCR4-positive sorted cells resulted in a greater number of fibers expressing canine dystrophin (Figure ), nuclei expressing canine lamin A/C (Figure ), and Pax7/canine lamin A/C double-positive nuclei (Figure ) per cross-section of muscle when compared to injection of 1x104 mixed cells. However, for this experiment, 1% of the parent population of mixed cells was CXCR4-positive, and as such, 1 × 106 mixed cells were required to obtain 1 × 104 CXCR4-positive sorted cells. Therefore, the true comparison is 1 × 104 mixed cells to 100 sorted cells, and 5 × 104 mixed cells to 500 sorted cells. Using this relationship, the freshly isolated mixed canine muscle-derived cells resulted in a greater number of fibers expressing canine dystrophin (Figure ), nuclei expressing canine lamin A/C (Figure ), and nuclei expressing both Pax7 and canine lamin A/C (Figure ).
Recombining CXCR4-positive and CXCR4-negative cells after sorting did not restore engraftment to the level observed with the mixed cell population (data not shown). Moreover, depleting the parent mixed canine muscle-derived cell population of CXCR4-expressing cells also resulted in a 10-fold decrease in the ability to generate fibers expressing canine dystrophin within recipient mouse muscle (Figure ), indicating that the CXCR4-positive cells within the mixed cell population are likely responsible for the majority of the observed engraftment, and that sorting may have a negative effect on cell engraftment.
CXCR4 and SDF-1 play an important role in donor cell engraftment
The process of sorting cells may adversely affect engraftment by reducing cell viability; however, it is also possible that binding of the antibody to the cell surface CXCR4 receptor interferes with function. Muscle injected with freshly isolated mixed canine muscle-derived cells incubated with α-CXCR4 antibody, but not subjected to FACS, displayed significantly fewer myofibers expressing canine dystrophin (Figure ), fewer nuclei expressing canine lamin A/C (Figure ), and fewer nuclei expressing Pax7 and canine lamin A/C (Figure ) than muscle injected with cells alone or cells incubated with control antibody. This may be due to antibody-mediated internalization of the receptor or blocking of SDF-1 binding to the CXCR4 receptor [14
]. Either mechanism argues for an important role for CXCR4 in engraftment and/or differentiation of donor cells.
Figure 4 Anti-CXCR4 antibody and SDF-1 inhibit engraftment. (A, B, C) Cryosections from mouse muscle injected with 1 × 104 freshly isolated mixed canine muscle-derived mononuclear cells alone, cells pre-incubated with anti-CXCR4, cells pre-incubated with (more ...)
The function of the CXCR4 receptor requires binding of its ligand, SDF-1. We compared engraftment of the freshly isolated mixed canine muscle-derived cells alone to mixed cells incubated before injection with 10 ng/ml of SDF-1 or 10 ng/ml of FGF-2. SDF-1, but not FGF-2, specifically reduced the number of fibers expressing canine dystrophin (Figure ), the number of nuclei expressing canine lamin A/C (Figure ), and the number of nuclei expressing Pax7 and canine lamin A/C (Figure ). Therefore, exogenously added SDF-1 did not improve donor cell engraftment, and indeed, appeared to mimic the effect of the anti-CXCR4 antibody. It is intriguing to hypothesize that exogenous SDF-1 blocked binding of CXCR4 on canine donor cells to SDF-1 in recipient mouse muscle.
Diprotin A treatment of donor cells enhances engraftment through CXCR4
Binding of SDF-1 to CXCR4 is negatively regulated by CD26/DPP-IV, a cell surface peptidase that cleaves SDF-1 at the N-terminus [18
]. CD26 is expressed on hematopoietic stem cells, and inhibition of peptidase activity with diprotin A enhances engraftment of donor hematopoietic cells to the recipient bone marrow niche, presumably by strengthening the interaction between CXCR4 on the surface of donor stem cells and SDF-1 present in the bone marrow niche [20
RT-PCR demonstrated that cultured proliferating and differentiating canine satellite cell derived myoblasts, and the freshly isolated population of mixed muscle-derived cells expressed CD26/DPP-IV transcript (Figure ). Freshly isolated canine bone marrow cells, a positive control for CD26/DPP-IV activity, and muscle-derived cells were incubated with Gly-Pro-p-nitroanilide, a substrate of CD26/DPP-IV, and production of the cleavage product, p-nitroaniline, was monitored by measuring absorbance at 405 nm. Muscle-derived mononuclear cells displayed CD26/DPP-IV activity at a level comparable to canine bone marrow cells, and this activity was inhibited with diprotin A (Figure ).
Figure 5 Diprotin A enhances engraftment of donor cells. (A) 1 × 105 canine bone marrow cells (BMC), and 1 × 105 freshly isolated skeletal muscle-derived cells (SkMC) alone, or with 5 mM diprotin A (SkMC+dpA), were incubated with Gly-Pro-p-nitroaniline (more ...)
Diprotin A treatment of donor cells before injection significantly increased the number of fibers expressing canine dystrophin (Figure ), the number of nuclei expressing canine lamin A/C (Figure ), and the number of Pax7/canine lamin A/C double-positive nuclei detected in injected muscle (Figure ). To confirm the increase in donor nuclei present, a PCR-based assay that distinguishes between canine and mouse DNA showed that mouse muscle injected with diprotin A treated donor canine cells had three-fold more canine DNA content than muscle injected with cells alone (data not shown).
However, CD26/DPP-IV cleaves other chemokines, such as IP-10/CXCL10 and MIP1β/CCL4 [22
]. To ensure that the increase in engraftment we observed with diprotin A was specific for SDF-1/CXCR4, canine muscle-derived cells were incubated with diprotin A alone or with diprotin A and α-CXCR4 antibody before injection. As shown in Figures and , the increase in engraftment in the presence of diprotin A was prevented by α-CXCR4 antibody, indicating that diprotin A specifically affected binding of SDF-1 to CXCR4.
Diprotin A increases engraftment of donor derived Pax7-positive cells
NOD/SCID mice were injected as described above with 1 × 104 untreated freshly isolated mixed canine muscle-derived cells, or 1 × 104 diprotin A treated cells, and engraftment was measured weekly for 4 weeks (Figure ), or at weeks 3, 5, and 7 after injection (Figure ). Different donors were used for the experiment shown in Figure and the experiment shown in Figure , which resulted in a different level of engraftment. However, the trend in engraftment as a function of time, and the effect of diprotin A is similar.
Figure 6 Diprotin A enhances proliferation of donor cells in vivo. Mouse muscle injected with 1 × 104 canine muscle-derived mononuclear cells alone, or cells treated with 5 mM diprotin A, were harvested weekly for 4 weeks after injection (A, B, C), or (more ...)
In muscle injected with untreated cells, the number of canine dystrophin-positive fibers reached a plateau 3 weeks after injection, and remained constant from weeks 3 through 7 (Figure and ). In contrast, muscle injected with diprotin A-treated cells displayed a continuous increase in canine dystrophin-positive fibers for no less than 7 weeks after donor cell injection, and beyond week 3, displayed a significantly greater number of canine dystrophin-positive fibers compared to muscle injected with untreated cells (Figure and ).
A similar pattern is observed for the number of nuclei expressing canine lamin A/C per cross-section (Figure and ). However, the number of Pax7/canine lamin A/C double-positive cells are significantly increased in muscle injected with diprotin A-treated cells at all time points (Figure and ). Moreover, there is a significant increase in the number of Pax7/canine lamin A/C double-positive cells between weeks 1 and 2 in muscle injected with diprotin A-treated cells, suggesting that Pax7/canine lamin A/C double-positive cells proliferated in vivo.
Diprotin A maintains donor cell proliferation
To quantitatively assess proliferation of donor cells in vivo, the nucleoside analog, EdU, was administered daily for 7 days, during weeks 1, 2, 3, or 4 after injection of untreated cells or diprotin A-treated cells into NOD/SCID mice. During weeks 1 and 2, both EdU-positive and EdU/canine lamin A/C double-positive nuclei were observed; however, EdU-positive cells that did not co-express canine lamin A/C were rare in weeks 3 and 4, suggesting that proliferating cells present during weeks 3 and 4 were mainly donor-derived (data nor shown).
The proportion of EdU/canine lamin A/C double-positive cells decreased in muscle injected with untreated cells and reached a plateau of approximately 10% by week 3 (Figure ). In contrast, diprotin A treatment of donor cells resulted in an increase in the proportion of donor cells incorporating EdU at weeks 3 and 4 after injection, reaching approximately 25 to 30% of cells. Less than 1% of cells co-expressed canine lamin A/C and caspase-3 at weeks 1 and 4 after injection, and diprotin A treatment did not affect the number of canine donor cells expressing caspase-3 (data not shown). This is consistent with the view that the majority of donor cell death occurs in the first 24 h after injection [24
]. Together, these data indicate that diprotin A treatment maintained a greater proportion of donor cells in proliferation.
The proportion of EdU/canine lamin A/C double-positive cells that also express Pax7 remained constant at approximately 16% in muscle injected with untreated cells (Figure ). In muscle injected with diprotin A-treated cells, this proportion increased to 35% at week 3, and remained significantly higher than in muscle injected with untreated cells. This indicates that diprotin A treatment maintains a greater proportion of proliferating donor derived Pax7-positive myogenic cells in vivo.
Diprotin A enhances engraftment of donor cells in immune tolerant cxmd canines
Together, these current data indicate that the most effective regimen involves intramuscular transplantation of a freshly isolated mixed population of canine muscle-derived cells treated with diprotin A before injection. Therefore, to determine if these results can be translated to a large animal model of muscular dystrophy, we chose to compare engraftment of freshly isolated mixed muscle-derived cells alone to cells treated with diprotin A using the immune tolerant cxmd canine model of DMD.
canines underwent hematopoietic stem cell transplantation (HSCT), as previously described [9
]. Briefly, each cxmd
recipient was exposed to 200 cGy of total body irradiation, and infused with bone marrow and G-CSF mobilized peripheral blood mononuclear cells from a DLA-identical littermate donor. For the first cxmd
recipient, H376, analysis of donor chimerism 4 weeks after transplant demonstrated that 80% of granulocytes and 46% of lymphocytes were donor derived. However, 12 weeks after transplant, donor hematopoietic chimerism increased to 99% for granulocytes and 80% for lymphocytes and remained constant thereafter. The donor muscle cell injection was performed 15 weeks after bone marrow transplantation. A muscle biopsy was obtained from the hematopoietic stem cell donor, and muscle-derived cells were isolated as described above for canine-to-murine transplantation experiments.
The biceps femoris (BF) muscle on one side of the cxmd recipient was marked with non-dissolvable sutures to identify six sites of cell injection. In an attempt to mimic the xenotransplant regimen, three sites were injected with 5 × 105 cells resuspended in 1.2% barium chloride, and three sites with 5 × 105 diprotin A-treated cells resuspended in 1.2% barium chloride. The sites injected with untreated cells were separated from the sites injected with diprotin A-treated cells by a minimum of 5 cm to prevent crossover or contamination.
Injected muscle was biopsied 8, 16, and 24 weeks after injection. The average number of dystrophin-expressing fibers per cross-section of muscle was determined using five cryosections from each biopsy, covering a total distance of 1000 μm, and normalized to cross-sectional area. Diprotin A treatment of donor cells resulted in a dramatic 6.8-fold increase in the number of fibers expressing dystrophin 24 weeks after injection, recapitulating the effect observed in canine-to-murine xenotransplantation experiments (Figure ).
Figure 7 Diprotin A enhances engraftment of donor cells in canine allogeneic transplantation. (A) Skeletal muscle cryosections from a chimeric cxmd canine (H376) biopsied 8, 16, and 24 weeks after injection with 5 × 105 donor cells in 1.2% BaCl2, or 5 (more ...)
To determine if the active muscle regeneration induced by barium chloride was important for donor cell engraftment, a second chimeric cxmd recipient, H220, was injected with cells resuspended in PBS alone. Analysis of donor chimerism in H220 8 weeks after HSCT demonstrated that 51% of granulocytes and 16% of lymphocytes were donor derived. However, 12 weeks after transplant, donor hematopoietic chimerism dropped to 3% and remained constant thereafter. The donor muscle cell injection was performed 37 weeks after HSCT. The biceps femoris (BF) muscles on each side of the cxmd recipient were marked with non-dissolvable sutures to identify six sites of cell injection. The left BF muscle was injected with 4 × 105 cells alone, and the right BF muscle injected with 4 × 105 diprotin A-treated cells.
Biopsies of injected muscle were obtained 8, 16, and 24 weeks after injection. Diprotin A treatment of donor cells resulted in a 2.6-fold increase in the number of dystrophin-positive fibers 24 weeks after injection (Figure ). The number of dystrophin-positive fibers 24 weeks after injection was 15.3 per mm2 for H220 as compared to 90.3 per mm2 for H376, suggesting that co-injection of cells with barium chloride enhanced engraftment of donor cells. Together, these data indicate that diprotin A successfully enhanced engraftment of donor cells to dystrophic skeletal muscle, and that results from the xenotransplant model accurately predicted results from the canine-to-canine allogeneic model.