We previously demonstrated that erythropoietin (EPO)-secreting mesenchymal stromal cells (MSC) can be used for the long-term correction of renal failure-induced anemia. The present study provides evidence that coimplantation of insulin-like growth factor I (IGF-I)-overexpressing MSC (MSC-IGF) improves MSC-based gene therapy of anemia by providing paracrine support to EPO-secreting MSC (MSC-EPO) within a subcutaneous implant. IGF-I receptor RNA expression in murine MSC was demonstrated by RT-PCR. Functional protein expression was confirmed by immunoblots and MSC responsiveness to IGF-I stimulation in vitro. IGF-I was also shown to improve MSC survival following staurosporin-induced apoptosis in vitro. A cohort of C57Bl/6 mice was rendered anemic by right kidney electrocoagulation and left nephrectomy. MSC-EPO were subsequently admixed in a bovine collagen matrix and implanted, in combination with MSC-IGF or MSC null, by subcutaneous injection in renal failure mice. In mice receiving MSC-EPO coimplanted with MSC-IGF, hematocrit elevation was greater and enhanced compared with control mice; heart function was also improved. MSC-IGF coimplantation, therefore, represents a promising new strategy for enhancing MSC survival within implanted matrices and for improving cell-based gene therapy of renal anemia.
- chronic kidney disease
- bone marrow-derived progenitor cells
gene and cell therapy are attractive approaches for the treatment of chronic diseases such as chronic kidney disease (CKD). Cell-based gene therapy, which couples ex vivo gene modification to cell therapy, provides a much safer alternative to traditional gene therapy. In fact, a variety of genetically engineered cell types have been employed in various clinical settings (2, 16, 17, 35); the mesenchymal stromal cell (MSC) is considered particularly promising. MSC are nonhematopoietic progenitor cells native to the bone marrow, and their inherent potential for expansion in culture and amenability to genetic modification make them ideal candidates for cell-based gene therapy. Previous preclinical studies employing gene-modified MSC demonstrated that these progenitors are indeed ideal delivery vehicles for the in vivo administration of therapeutic transgenic products, such as erythropoietin (EPO) (12, 22, 39, 45). We previously reported that murine MSC genetically modified to secrete EPO (MSC-EPO) are capable of generating a dose-dependent increase in hematocrit when admixed in a bovine collagen matrix and implanted by subcutaneous injection in mice (12–14). However, despite the relative successes reported in our studies and others, limited long-term cellular viability compromises the efficacy of MSC-based cell therapy interventions (25). Ameliorating the viability and survival of implanted cells would therefore be instrumental in improving the outcome of cell therapy studies involving MSC.
Several well-documented characteristics identify insulin-like growth factor I (IGF-I) as an attractive candidate for improving the in vivo viability of implanted cells, particularly MSC. Briefly, IGF-I regulates a number of cellular processes through activation of the MAPK and phosphatidylinositol 3-kinase/Akt signaling pathways, including cell proliferation and migration, cell growth, angiogenesis, and apoptosis (8). In addition, it has been shown that murine and human bone marrow stromal cells express the IGF-I receptor (IGF-IR) (9, 18, 42), suggesting that murine MSC are likely to be responsive to IGF-I stimulation. IGF-I has long been known to stimulate erythroid progenitor development (4, 26, 36) and has been shown to synergistically increase hematocrit in mice with CKD-associated anemia when used in conjunction with EPO (7). We therefore hypothesized that IGF-I secretion by gene-modified MSC (MSC-IGF) within a subcutaneous implant could serve as a protective microenvironment providing paracrine support to MSC-EPO, leading to improved long-term delivery of EPO for the treatment of renal anemia. In the present study, we demonstrate that, in a mouse model of CKD, the coimplantation of MSC-IGF with MSC-EPO induces a greater and enhanced hematocrit elevation, as well as improved cardiac function, compared with animals receiving MSC-EPO alone.
MATERIALS AND METHODS
Isolation, culture, and retroviral transduction of primary murine bone marrow-derived MSC.
All animal experimentation was performed according to the guidelines of the Canadian Council on Animal Care with prior approval of the Institutional Animal Care and Use Committee. MSC isolation and culture were performed as previously described (12). Briefly, female C57Bl/6 mice (Charles River, St. Constant, PQ, Canada; ∼15–20 g body wt) were euthanized by CO2 asphyxiation. Whole bone marrow was harvested in cell culture dishes by flushing resected femurs and tibias with complete culture medium [DMEM supplemented with 10% FBS and 48 μg/ml gentamicin (Sandoz Canada)] and kept at 37°C with 5% CO2. Nonadherent hematopoietic cells were discarded 5 days later, and adherent stromal cells were maintained in culture for 7–10 passages to generate a polyclonal wild-type (WT) MSC population.
A retroviral construct harboring murine EPO (mEPO) cDNA was previously generated in our laboratory (14). We additionally generated a retroviral plasmid construct harboring the murine IGF-1A gene. We transfected 293-GP2 viral packaging cells with one of three plasmids, the construct harboring mEPO, the construct harboring murine IGF-1A, or the empty vector, to generate MSC-EPO, MSC-IGF, or MSC null, respectively. IGF-I and EPO secretion levels in supernatants of null and gene-modified MSC were subsequently assessed by ELISA specific for mouse IGF-I (R & D Systems, Minneapolis, MN) or human EPO (Roche Diagnostics).
Subcutaneous implantation of MSC-IGF for determination of long-term systemic effect.
The long-term effect of MSC-IGF implantation was assessed by subcutaneous implantation of 2 × 106 MSC-IGF in a Matrigel plug. Briefly, MSC were suspended in 50 μl of PBS and then admixed with 400 μl of Matrigel (BD Biosciences, Mississauga, ON, Canada) at 4°C in single-dose 1-ml syringes. Syringes were subsequently warmed until no longer cool to the touch, and the MSC suspended in Matrigel was implanted by subcutaneous injection in the right flank of recipient C57Bl/6 mice (n = 5; Charles River). A total of 2 × 106 MSC-IGF C49, a clonal population of IGF-I-secreting MSC, were implanted by subcutaneous injection in normal mice (n = 5). Polyclonal MSC-IGF PC1 (n = 5) and MSC null PC1 (n = 5) were used as controls. Animals were monitored for 30 days after implantation: blood samples were collected from the saphenous vein for determination of hematocrit (see below), and animals were weighed once per week. Average percent weight gain was assessed as follows: (body weight at a given time point ÷ weight at day 0) × 100. Plasma IGF-I levels were assessed by ELISA specific for mouse IGF-I.
Subcutaneous coimplantation of gene-modified MSC and analysis of hematocrit in normal mice.
We previously described the subcutaneous implantation of MSC in Contigen (C. R. Bard, Covington, GA), a US Food and Drug Administration-approved bovine collagen matrix (14). Briefly, MSC were washed with PBS and suspended in 50 μl of serum-free medium and then admixed with 400 μl of Contigen in single-dose 3-ml syringes. The matrix-embedded MSC were subcutaneously implanted in the right flank of recipient mice: two groups of mice (n = 5 in each group) received 2 × 105 MSC-EPO + 106 MSC-IGF or 2 × 105 MSC-EPO + 106 MSC null. Blood samples were collected from the saphenous vein approximately once per week via heparinized microhematocrit tubes, and hematocrit was measured by the standard microhematocrit centrifugation method. Plasma IGF-I and EPO levels were assessed by ELISA specific for mouse IGF-I or human EPO.
Matrigel plug assay for histological analysis of angiogenesis.
For histological assessment of angiogenesis induced by MSC-IGF implantation, the same groups described in Subcutaneous implantation of MSC-IGF for determination of long-term systemic effect were used. MSC-IGF C32, a clonal population of MSC secreting levels of IGF-I comparable to MSC null (0.5 ng of IGF-I per 106 cells per 24 h), were also implanted as an additional control (n = 3). Matrigel plugs were surgically retrieved after 14 days, cut in half, and processed for histology. The implant was embedded in paraffin, with the cut side of each half-implant facing the proximal edge of the block, and immunohistochemistry was performed on 5-μm-thick Matrigel plug sections. Endogenous biotin activity was blocked using a kit (Zymed Laboratories, Markham, ON, Canada), and the sections were blocked with 2.5% BSA in PBS and then incubated with a rabbit polyclonal anti-mouse von Willebrand factor antibody (Neomarkers, Fremont, CA). The sections were incubated with biotinylated goat anti-rabbit IgG antibody (BD PharMingen) and probed with streptavidin-peroxidase (Vector Laboratories, Burlingame, CA), and 3,3′-diaminobenzidine chromogenic substrate was added (Vector Laboratories). Standard hematoxylin counterstaining was performed on all sections. Microvessels were defined as circular structures formed by von Willebrand factor-positive cells with a well-defined lumen. Total microvessels were counted on each section, from every animal in each of the above-described groups, using a Zeiss Standard 25 ICS transmitted-light microscope (Carl Zeiss Canada, Toronto, ON, Canada). Sections were photographed, and Scion Image software (version 184.108.40.206, Scion, Frederick, MD) was used to calculate the surface area of each section. Microvessel density was determined as follows: average number of vessels per implant section ÷ surface area of the section (vessels/mm2).
Flow cytometric analysis of angiogenesis induced by coimplantation.
For assessment of angiogenesis induced by MSC-EPO and MSC-IGF coimplantation, 2 × 105 MSC-EPO were admixed with 106 MSC-IGF or 106 MSC null in growth factor-reduced Matrigel (BD Biosciences), as described above. At 14 days after implantation, implants were removed and cells were retrieved by enzymatic digestion. Briefly, implants were cut into small fragments in a solution of PBS supplemented with 1.6 mg/ml type IV collagenase and 200 μg/ml DNase I (Sigma-Aldrich) and incubated at 37°C for 30 min. Cells were dissociated from Matrigel particles by repeated pipetting and reincubated at 37°C for an additional 20 min. The dissociated implants were suspended in staining buffer (3% FBS in PBS) and filtered through a 40-μm nylon mesh cell strainer (BD Biosciences Discovery Labware, Bedford, MA). The filtrate was retained, washed with staining buffer, and stained using biotin-conjugated rat anti-mouse CD31 and FITC-labeled rat anti-mouse CD45 antibodies (BD Biosciences, San Diego, CA), and phycoerythrin (PE)-conjugated streptavidin was added. Fc block was performed before antibody incubation using a purified anti-mouse CD16/Cd32 antibody (anti-Fcγ III/II receptor, clone 2.4G2, BD Biosciences). Isotypic controls were biotin-conjugated rat IgG2a and FITC-labeled rat IgG2b (BD Biosciences). Cells were then acquired using a fluorescence-activated cell sorter (FACS) Coulter flow cytometer, and data were analyzed with Cellquest Pro software (BD Immunocytometry Systems, San Jose, CA).
Generation of a mouse model of chronic renal failure and correction of renal failure-induced anemia.
A murine model of chronic renal failure was generated as previously described (15). Briefly, moderate renal failure was induced in C57Bl/6 mice (Harlan, Indianapolis, IN) by electrocoagulation of the right kidney surface, leaving the hilum intact (day 0); animals were allowed to recover for 22 days, and then the left kidney was surgically ablated (day 22). Hematocrits remained at normal physiological levels after right kidney injury but declined dramatically after left nephrectomy, stabilizing at ∼40% after 30 days. On day 52, anemic mice were implanted with 2 × 105 MSC-EPO + 106 MSC-IGF (n = 9), 2 × 105 MSC-EPO + 106 MSC null (n = 9), or 2 × 105 MSC null + 106 MSC-IGF (n = 10) suspended in Contigen. Animals receiving no treatment were also kept as controls (n = 6). Blood urea nitrogen (BUN) was assessed in all animals 14 days after left nephrectomy and 48 days after MSC administration by the colorimetric method with a urea nitrogen (BUN) reagent kit (Teco Diagnostics, Anaheim, CA). Plasma samples from every animal in each group were pooled, and assays were performed in triplicate.
Two-dimensional and M-mode echocardiography was performed in the mice with experimentally induced chronic renal failure on day 119 (67 days after implantation), as previously described (1). Heart rates were simultaneously obtained by three-lead electrocardiogram. Percent fractional shortening was measured in M-mode traces by the leading-edge method (33) and derived as previously described (40). Measurements were performed on six randomly selected mice in each of the four groups and averaged from three consecutive beats of three image acquisitions for each animal.
In Vitro Experimentation
RT-PCR analysis of cDNA for IGF-IR in MSC.
MSC (∼5 × 106) were placed in cell lysis buffer, homogenized, and applied to RNA purification columns according to the manufacturer's instructions (RNeasy kit, Qiagen, Mississauga, ON, Canada). After reverse transcription of total RNA (2 μg) in a volume of 50 μl with use of random hexamers, 50 ng of resulting cDNA was subjected to PCR for determination of IGF-IR expression [CCTGAGGCGTGGAGATAGAG (sense primer) and TGTCAGCCCACCCTAAAAAC (antisense primer)]. DNA amplification was performed with a programmable thermal controller set to 95°C for 3 min followed by 40 cycles of 95°C for 30 s, 60°C for 30 s, 72°C for 40 s, and, finally, 72°C for 10 min.
Flow cytometric analysis of MSC immunophenotype and mesenchymal differentiation.
Flow cytometric analysis was used to characterize MSC immunophenotype. Immunostaining was performed by incubation of MSC with the following monoclonal antibodies: allophycocyanin (APC)-labeled rat anti-mouse CD31 (clone MEC13.3, BD Biosciences, San Diego, CA), biotin-conjugated rat anti-mouse CD34 (clone RAM 34, BD Biosciences), PE-labeled rat anti-mouse CD44 (clone IM7, BD Biosciences), APC-labeled rat anti-mouse CD45 (clone 30-F11, BD Biosciences), PE-labeled rat anti-mouse CD73 (clone TY/23, BD Biosciences), FITC-labeled rat anti-mouse CD90 (clone G7, Southern Biotechnology Associates, Birmingham, AL), and biotin-conjugated rat anti-mouse CD105 (clone MJ7/18, eBioscience, San Diego, CA). Biotinylated antibodies were conjugated with PE-streptavidin (BD Biosciences), and isotypic control analyses were performed in parallel. After incubation, cells were washed and acquired using an FACS Calibur flow cytometer (BD Immunocytometry Systems), and data were analyzed with Cellquest Pro software. Induction of MSC null and MSC-IGF differentiation was accomplished by exposure to specific differentiation medium, and differentiation was assessed as previously described (14).
Immunoblot analysis of IGF-IR and MAPK phosphorylation in MSC.
For immunoblot analysis of the IGF-IR, undifferentiated MSC, differentiated osteoblasts, and differentiated adipocytes were lysed in RIPA buffer. Total protein (15 μg) from each sample was separated using precast 8% Tris-glycine SDS-PAGE denaturing gels (Invitrogen, Carlsbad, CA), electroblotted onto 0.45-nm-pore polyvinylidene difluoride membranes (Millipore, Billerica, MA), and probed with a rabbit polyclonal anti-IGF-IRβ C20 antibody (catalog no. sc-713, Santa Cruz Biotechnology, Santa Cruz, CA). ECL anti-rabbit IgG horseradish peroxidase-linked secondary antibody (from donkey; Amersham Biosciences) was subsequently used to probe bound primary antibody. For evaluation of MAPK phosphorylation, 5 × 105 WT MSC were seeded in 10-cm dishes in complete medium and serum starved overnight. On the next day, cells were washed and then incubated in serum-free medium supplemented with 0, 10, or 100 ng/ml recombinant mouse IGF-I (R & D Systems), 10 IU/ml recombinant human EPO (Eprex, Janssen-Ortho, North York, ON, Canada), or IGF-I (10 ng/ml) + EPO (10 IU/ml) for 10 min. Cells were then lysed directly in the culture dish using lysis buffer supplemented with 0.01% SDS, 1 mM Na3VO4, and Complete Mini EDTA-free protease inhibitor cocktail (Roche Diagnostics, Indianapolis, IN). Thirty micrograms of protein from each sample were separated using precast 4–20% Tris-glycine SDS-PAGE denaturing gels (Invitrogen). After transfer to polyvinylidene difluoride membrane and antigen blocking, membranes were probed using a polyclonal rabbit phosphorylated p44/42 MAPK (Thr202/Tyr204) antibody (Cell Signaling Technology, Danvers, MA) for assessment of Erk1 and Erk2 phosphorylation. Secondary antibody probe was performed as described above. Spot densitometry was performed using a MultiImage light cabinet (Alpha Innotech, San Leandro, CA) and analyzed with ChemiImager Alpha Ease software (version 5.5, Alpha Innotech).
Assessment of staurosporin-induced apoptosis by flow cytometry.
MSC apoptosis was induced in vitro and assessed by flow cytometry. Approximately 7 × 105 MSC were seeded in a 10-cm culture dish in complete medium. Cells were placed in serum-reduced conditions overnight (2% FBS) in the presence of 10 ng/ml recombinant mouse IGF-I (in PBS with 0.1% BSA, final concentration 2 × 10−5% BSA), 10 IU/ml recombinant human EPO [in solution containing 0.25% human serum albumin (HSA), final concentration 1.25 × 10−4% BSA], or IGF-I + EPO (with HSA and BSA). In addition, 1.25 × 10−4% HSA was included with IGF-I stimulation and 2 × 10−5% BSA with EPO stimulation as controls. On the next morning, cells were washed with PBS and bathed in fresh medium containing the same stimulus used on the previous night, with the addition of 20 nM staurosporin (Sigma-Aldrich, Oakville, ON, Canada). After 24 h of apoptotic induction, cells were stained using a combination of annexin V and propidium iodide (PI; annexin V FITC kit, Biosource International, Montreal, QC, Canada). Annexin V cell surface staining and PI uptake in MSC were assessed using a FACS Calibur flow cytometer, and data were analyzed with Cellquest Pro software.
Values are means (SD). Student's t-test was performed for statistical comparison of two groups; P < 0.05 was considered significant. ANOVA for statistical comparison of three groups or more was followed by Tukey's post hoc test of significance; P < 0.05 was considered significant.
MSC Express a Functional IGF-IR
RT-PCR performed on cDNA isolated from a polyclonal MSC population using primers for the murine IGF-IR confirmed IGF-IR mRNA expression (Fig. 1A). Using published antibodies (5, 27), we were also able to identify IGF-IR β-subunit protein expression by human and murine MSC (Fig. 1B). Differentiation into adipocytic or osteogenic lineages did not affect receptor expression. In vitro stimulation using recombinant murine IGF-I (rIGF-I) induced a dose-dependent increase in MAPK (Erk1/2) phosphorylation, confirming that this receptor was functional (Fig. 1C).
Phenotypic and Functional Characterization of MSC Engineered to Overexpress IGF-I
WT MSC were genetically engineered using retroviral vectors encoding the murine IGF-I gene or an empty vector control (MSC null). MSC null controls secreted ∼0.4 ng of IGF-I per 106 cells per 24 h, whereas MSC-IGF PC1 secreted 1.8 ng of IGF-I (Fig. 2A). A single clone isolated from this polyclonal population, clone 49 (hereafter referred to as MSC-IGF C49), secreted ∼8.0 ng of IGF-I per 106 cells per 24 h (Fig. 2A) and was therefore used in several experiments. A second population of WT MSC was also transduced, and the newly generated MSC-IGF PC2 was found to secrete ∼5.0 ng of IGF-I per 106 cells per 24 h (Fig. 2B). The MSC-IGF PC2 and MSC null PC2 polyclonal populations were used in all these experiments, unless otherwise specified. Polyclonal MSC-IGF and MSC null were subsequently exposed to differentiation medium, which readily promoted differentiation along adipocytic and osteogenic lineages in both populations (Fig. 2C; MSC null not shown). Polyclonal MSC-IGF and MSC null were then immunophenotyped by flow cytometry for commonly reported markers of MSC and found to be positive for expression of CD34 (90%), CD44 (96%), and CD73 (82%) and negative for CD31, CD45, CD90, and CD105 (Fig. 2D; MSC null not shown). As a safety and quality-control measure, we evaluated whether the implantation of 2 × 106 MSC-IGF C49 in Matrigel had adverse long-term effects on normal mice. It was determined that neither weight gain (Fig. 2E) nor hematocrit (Fig. 2F) was significantly affected up to 30 days after implantation of MSC-IGF. In addition, the subcutaneous implantation of 2 × 106 MSC-IGF C49 did not promote an increase in circulating plasma IGF-I levels 30 days after implantation compared with controls receiving MSC null PC1 and normal mice (Fig. 2G).
Subcutaneous Implantation of MSC-IGF and MSC-EPO in Normal Mice and Assessment of MSC Survival After Apoptotic Induction In Vitro
To test the hypothesis that MSC-IGF would enhance the proerythrocytic effect of MSC-EPO in vivo, polyclonal MSC-IGF were coimplanted with MSC-EPO in normal mice by subcutaneous injection in a bovine collagen matrix. The polyclonal MSC-EPO population secreted ∼530 IU of EPO per 106 cells per 24 h and was shown to lead to a robust EPO-dependent increase in hematocrit upon subcutaneous implantation. Normal mice receiving MSC-EPO + MSC-IGF experienced a similar initial rise in hematocrit compared with animals receiving MSC-EPO coimplanted with MSC null; however, the hematocrits of mice coimplanted with MSC-IGF remained elevated over a significantly greater period of time (P < 0.05 for >42 days), up to a maximum hematocrit of 86% (SD 6.2) (Fig. 3A). We attribute this enhanced elevation in hematocrit to a significant increase in plasma EPO levels (P < 0.05); at 7 wk after implantation, EPO was ∼11.4 mU/ml (SD 5.7) in the plasma of animals receiving MSC-IGF + MSC-EPO compared with only 3.4 mU/ml (SD 1.8) in controls (Fig. 3B).
Histological and flow cytometric analysis of Matrigel plugs and Contigen implants was performed to assess whether angiogenesis was increased by the implantation of MSC-IGF and/or MSC-EPO. Histological analysis of Matrigel implants by immunohistochemistry revealed that the implantation of MSC-IGF C49 does not promote increased vascularization after 14 days compared with MSC-IGF PC1, MSC null PC1, or MSC-IGF C32 (Fig. 3C). In support of this evidence, flow cytometric analysis of enzyme-digested Contigen implants revealed that MSC-IGF do not impact the neovascularization of MSC-EPO implantations: 3.8% (SD 1.3) and 3.9% (SD 1.3) CD31+/CD45− cells for MSC-EPO + MSC null and MSC-EPO + MSC-IGF, respectively (Fig. 3D).
To determine whether IGF-I promotes MSC survival after apoptotic induction in vitro, MSC prestimulated with rIGF-I, recombinant EPO (rEPO), or rIGF-I + rEPO were subjected to staurosporin-induced apoptosis. FITC-labeled annexin V-PI double staining of the MSC revealed that IGF-I preconditioning and stimulation improved the number of nonapoptotic MSC after 24 h of apoptotic induction compared with untreated control MSC: 74.5% (SD 3.0) annexin V−/PI− healthy MSC with IGF-I vs. 65.8% (SD 2.8) in controls (Fig. 3E) and 23.0% (SD 2.7) PI+ late-apoptotic and necrotic cells with IGF-I vs. 30.5% (SD 2.3) in controls (Fig. 3F; P < 0.05). This decrease in apoptosis induced by IGF-I stimulation was not affected by the addition of EPO: 74.5% (SD 3.0) annexin V−/PI− MSC with IGF-I vs. 74.2% (SD 1.3) with IGF-I + EPO (Fig. 3E) and 23.0% (SD 2.7) PI+ MSC with IGF-I vs. 22.1% (SD 1.6) with IGF-I + EPO (Fig. 3F).
Coimplantation of MSC-IGF and MSC-EPO in a Mouse Model of CKD
We tested whether the coimplantation of MSC-EPO with MSC-IGF in a mouse model of CKD would enhance the correction of anemia observed in normal mice. After induction of moderate chronic renal failure in C57Bl/6 mice, the animals were implanted with MSC-EPO + MSC-IGF, MSC-EPO + MSC null, or MSC null + MSC-IGF. Renal failure was evaluated by assessment of BUN 14 days after left nephrectomy, as well as 48 days after implantation (Fig. 4A). Significant increases in BUN in mice from all groups at both time points (P < 0.05 vs. normal mice; Fig. 4A) confirmed that renal function was indeed compromised.
As expected, the anemia resulting from experimental renal insufficiency was corrected in the two groups of mice treated with MSC-EPO (Fig. 4B). In the animals receiving MSC-EPO alone (i.e., coimplanted with MSC null), hematocrits increased gradually to normal physiological levels, attaining a maximum level of 56.5% (SD 4.8) 20 days after implantation (day 72). In subsequent weeks, hematocrits decreased gradually, becoming statistically comparable to untreated controls at around day 100. In contrast, in animals receiving MSC-EPO coimplanted with MSC-IGF, hematocrit increased gradually and continuously to supraphysiological levels, maximizing at 69.8% (SD 18.9) after 15 wk and remaining significantly elevated (P < 0.05) compared with MSC-EPO controls for >9 wk.
Echocardiography was performed in the renal failure mice on day 119 (67 days after implantation) to determine whether the superior correction of anemia in mice implanted with MSC-EPO ± MSC-IGF led to an improvement in cardiac function. We found a significant improvement in cardiac function, based on fractional shortening, in animals receiving MSC-EPO coimplanted with MSC-IGF compared with the three other groups: 42.8% (SD 5.6) for MSC-EPO + MSC-IGF, 36.5% (SD 3.6) for MSC-EPO + MSC null, 36.5% (SD 4.3) for MSC null + MSC-IGF, and 36.5% (SD 6.1) for untreated controls (P < 0.001; Table 1).
Our group previously published a number of studies exploring the use of MSC-EPO for increasing hematocrit in normal mice and in clinically relevant disease models (12–14). The present study is the first to employ a secondary supporting cell type to enhance this elevation in hematocrit. We demonstrated that MSC genetically modified to secrete elevated levels of IGF-I could enhance the therapeutic effect of MSC-EPO when coimplanted in normal mice. MSC-IGF coimplantation significantly increased the plasma EPO levels of animals receiving MSC-EPO, but plasma IGF-I levels were not increased in mice receiving MSC-IGF alone compared with mice receiving MSC null. It is therefore clear that the difference in hematocrit was due to a local interaction between MSC-IGF and MSC-EPO within the implant, rather than a synergistic stimulation of erythroid progenitor development in the bone marrow, as described by Brox et al. (7). Supporting this contention, our study adds to previous work reporting IGF-IR expression by murine MSC (9, 18, 42) and goes on to demonstrate that the IGF-IR is capable of inducing the dose-dependent phosphorylation of downstream effectors and modulating apoptosis in vitro.
In a mouse model of CKD and secondary anemia, as was the case in normal mice, we observed a significant difference in the hematocrits between mice coimplanted with MSC-IGF and those receiving MSC-EPO alone. Specifically, we were able to achieve a correction in hematocrit to physiological levels in the renal failure mice receiving MSC-EPO alone (MSC-EPO + MSC null), whereas those receiving coimplanted MSC-IGF experienced an increase in hematocrit to supraphysiological levels, which were maintained for a significantly greater period of time. Although the difference in plasma EPO levels between the two groups seems to be at odds with the observed hematocrit at the same time point, in considering the physiology of erythropoiesis, this observation is not at all unexpected. An increase in plasma EPO levels will not result in an immediate increase in hematocrit. At least 10–14 days are required between the time erythroid progenitors in the bone marrow are stimulated by increased circulating EPO levels and the time an observable increase in hematocrit will occur (because of the maturation of erythroid progenitors into red blood cells during this period). Therefore, a difference in hematocrit will be observed between the two groups ∼10–14 days after the increase in plasma EPO levels (which is indeed the case in our study). These data clearly suggest an increased efficacy of MSC-EPO therapy for the correction of renal failure-induced anemia when coupled with local IGF-I production. Furthermore, given that reduced fractional shortening is a hallmark of poor cardiac function, the significantly increased fractional shortening in the hearts of animals coimplanted with MSC-EPO and MSC-IGF suggests that cardiac function was improved in these animals compared with all other groups. This is of particular importance given the prevalence of major cardiovascular complications in patients with CKD: heart failure accounts for ∼50% of the deaths of patients with end-stage renal disease (6).
As a safety and quality-control measure, we evaluated whether the implantation of MSC-IGF had adverse long-term effects on mice. Our results led us to conclude that the implantation of MSC-IGF had no significant effect on hematocrit, weight gain, or plasma IGF-I levels after 30 days. We believe this to be a consequence of the elevated levels of endogenous circulating IGF-I in the serum of most mammals, including mice (32), rats (30), pigs (34), and humans (43): the administration of MSC-IGF generates a relatively negligible quantity of systemic IGF-I compared with total plasma IGF-I level and, therefore, cannot be promoting increased red blood cell maturation in the bone marrow (in synergy with EPO). We also demonstrated that MSC-IGF (implanted on their own) do not induce an increase in angiogenesis compared with MSC null, nor do they affect the increase in angiogenesis that is normally observed upon MSC-EPO implantation. The prolonged elevation in hematocrit mediated by MSC-EPO upon MSC-IGF coimplantation is therefore not due to an increase in implant vascularity but, rather (as we show in vitro), is the direct effect of IGF-I on MSC-EPO viability.
It has previously been demonstrated that cotransplanted MSC improve the engraftment of hematopoietic progenitor cells during bone marrow transplantation (3, 11, 29), whereas more recent studies have described the use of MSC engineered to overexpress a variety of intracellular or secreted molecules, including IGF-I, to preserve cellular function and improve cell survival in vivo (22–24, 38, 45). The targeted overexpression of IGF-I in healthy or damaged tissue, typically using viral vectors, has been used to stimulate the regeneration of skin and connective tissue (19, 28), neurons (20, 44), and skeletal and cardiac muscle (10, 21, 37, 41). In addition, there are a number of notable studies describing ex vivo gene modification for cell-based IGF-I delivery (23, 24, 38). A common denominator in all these studies was the observation that a higher fraction of IGF-I-overexpressing than control cells survives the transplantation procedure and that these cells provide paracrine support to local tissues (23, 24). These results are consistent with our data showing that IGF-I reduces MSC apoptosis and likely promotes MSC-EPO survival within subcutaneous implants by a paracrine mechanism, resulting in a prolonged therapeutic effect. We propose that IGF-I secretion by MSC-IGF enhances cellular viability within the harsh anoxic environment of the implant, thereby reducing the large proportion of cells that normally die off after implantation. Indeed, our in vitro apoptosis data demonstrate that the total number of healthy, nonapoptotic MSC is improved from ∼66% to 74% upon stimulation with IGF-I. Although this difference is significant, we recognize that it may seem unlikely that an 8% improvement in protection from apoptosis can account for the enhanced increase in hematocrit. However, the MSC-EPO population employed in the present study secreted EPO at levels that were far higher than in any of our previous studies (>500 IU of EPO per 106 cells per 24 h). As a result, a different (i.e., a much lower) cellular dose was implanted to prevent excessively high hematocrit levels in the mice. Therefore, given the extremely elevated EPO secretion levels of our MSC-EPO population, a seemingly modest 8% increase in cell survival is very likely sufficient to cause a significant difference in EPO secretion from the implant, thereby causing a greater increase in plasma EPO levels in these animals, which in turn resulted in the greater increase in hematocrit.
It has not escaped our attention that other mechanisms may explain these observations, including the possibility that IGF-I stimulation somehow enhances protein secretion in MSC, perhaps by regulating gene function. In addition, a shortcoming of our study is that we did not directly demonstrate that MSC-EPO viability was improved upon coimplantation of MSC-IGF. Nevertheless, we believe that there is a strong body of evidence in the literature (see above) supporting the mechanism that we have proposed. We performed a comprehensive series of experiments confirming that the coimplantation of MSC-EPO with MSC-IGF enhances the elevation in hematocrit normally observed in normal mice and showed that several aspects of MSC biology, taken together, explain how this phenomenon is possible. These include the expression of an IGF-I receptor (at mRNA and protein levels) capable of promoting antiapoptotic signaling, as well as an increase in the survival of IGF-I-stimulated MSC upon induction of apoptosis. We subsequently confirmed the observed phenomenon in a clinically relevant disease model. Although we previously engineered MSC-EPO to overexpress green fluorescent protein (GFP) for in vivo cell tracking, our group recently demonstrated that MSC are capable of antigen presentation (31) and, thus, could possibly present xenoantigens such as GFP. This would increase the likelihood that GFP-expressing MSC are immune rejected compared with unmarked cells. For this reason, we elected to remove the GFP gene from the constructs that were used to genetically modify the MSC in the present study, making it much more difficult to track these cells in vivo. This precluded evaluation of cell survival in the implants. In addition, the extremely high EPO secretion levels of the MSC-EPO population used in the present study necessitated implantation of a very low cellular dose (2.5 × 105 MSC-EPO) to prevent hematocrits from increasing to dangerous levels. It is well established that the number of cells implanted will be reduced dramatically (up to 90%) in the 24–48 h after implantation because of poor engraftment, low nutrient and oxygen availability in the tissues, and immune rejection (25). Therefore, even if we were to have retained the GFP label on these MSC (or used any other fluorescent marker), it would have been very difficult to track such a negligible quantity of cells. Engraftment studies are notoriously challenging to perform, even when several million cells are implanted; with implantation of only 250,000 MSC-EPO, it would have been very difficult to directly evaluate whether MSC survival or proliferation is improved when coimplanted with MSC-IGF. For this reason, we performed a comprehensive series of in vitro experiments to support our in vivo findings and, subsequently, proposed a mechanism that was supported by our data and the literature.
Our study demonstrated that MSC-IGF may be used in conjunction with any other cell type expressing the IGF-IR, whether genetically modified or not, to improve the outcome of cell therapy interventions. This could be accomplished within an implanted matrix, as was the case in our study, or directly within the target tissue. We therefore envision the future use of MSC-IGF, not only for the improvement of MSC-EPO survival in the treatment of CKD-induced anemia, but also for the enhancement of cellular viability in a wide array of cell therapy approaches.
This study was supported by the Fonds de la recherche en santé du Québec (FRSQ) biovigilance program. T. Kucic holds a Canada Graduate Scholarships Master's Award from the Canadian Institutes of Health Research, a Bourse de formation maîtrise from the FRSQ, and a McGill Graduate Studies Fellowship. J. Galipeau is an FRSQ chercheur-boursier sénior.
The authors thank Laurence Lejeune and Marie-Noëlle Boivin for technical assistance with flow cytometry, Kathy-Ann Forner and Moïra François for assistance with animal manipulation and care, and Nicoletta Eliopoulos for dedicated guidance and assistance with the renal anemia model and MSC-EPO platform.
The costs of publication of this article were defrayed in part by the payment of page charges. The article must therefore be hereby marked “advertisement” in accordance with 18 U.S.C. Section 1734 solely to indicate this fact.
- Copyright © 2008 the American Physiological Society