Does the human skeletal muscle harbor the murine equivalents of cardiac precursor cells?

Abstract

The limited plasticity of adult muscle- or bone marrow- derived stem cells intended for cardiac regeneration impedes their conversion into cardiomyocytes. Since murine skeletal muscle was reported to harbor cardiac precursor cells, we assessed whether similar cells exist in man. Skeletal muscle biopsies obtained from 39 patients were sorted by flow cytometry which generated three populations (CD90+/CD34(-), CD34+/CD90(-), CD90(-)/CD34(-)) expressing similar levels of cardiac (Nkx2.5, cTn-T, cTn-I, Cx43) and skeletal muscle (Myf-5, MyoD, myogenin) mRNAs, as assessed by quantitative reverse transcriptase-PCR. However, compared to unpurified myoblasts, CD34+/CD90(-) cells expressed greater amounts of endothelium-specific mRNAs and were, therefore, selected for transplantation experiments. Thirty immunosuppressed rats then underwent coronary artery ligation and, 4 weeks later, were intramyocardially injected with culture medium, myoblasts, or CD34+/CD90(-) cells. After 1 month, left ventricular ejection fraction was significantly higher in the CD34+/CD90(-) group than in the control and myoblast-injected hearts, which was associated with smaller fibrosis and greater angiogenesis. The low engraftment rate suggested a paracrine mechanism supported by the greater release of growth factors by CD34+/CD90(-) cells than by unsorted myoblasts. In conclusion, the human skeletal muscle does not harbor cardiac-specified cells but contains a CD34+ fraction endowed with an angiogenic potential providing superior functional and structural benefits.

Figure 1

Flow cytometric analysis of freshly isolated (a,b) and proliferated (c) human muscle–derived cells. (a) Representative dot plots show that unsorted cells are inhomogeneous in cell size and granularity and express different levels of CD90 and CD34. Representative sorting gates are indicated. (b) Bar diagrams show the expression levels of CD90, CD34, CD45, and CD56 in unsorted cells. The majority of the cells was double-negative for CD90 and CD34, and only 2.19 ± 0.31% were CD90⁺/CD34⁻, whereas 24.05 ± 3.49% were CD34⁺/CD90⁻. Double-positive staining for CD34 and CD56 could virtually not be detected. (c) After proliferation, the percentage of CD56 in CD34⁺/CD90⁻ cells did not increase, whereas myoblasts robustly expressed CD56. Of note, the expression pattern of proliferated CD34⁺/CD90⁻ cells resembled that of myoblasts with regard to CD90, CD144, HLA-I (ABC), and HLA-II (DR). Because of similar expression patterns, the expression of HLA-II (DR) is thus shown in CD34⁺/CD90⁻ cells, and the expression of HLA-I (ABC) is illustrated in myoblasts. Bars and percentages are presented as mean ± SEM. APC, allophycocyanin; HLA, human leukocyte antigen; PE, phycoerythrin.

Figure 2

Representative gels and quantitative reverse transcriptase–PCRs (qRT-PCRs) of cardiac, skeletal muscle, and endothelial mRNAs in the different cell fractions. (a) Both CD90⁺/CD34⁻ and DN (double- negative: CD90⁻/CD34⁻) cells already coexpressed cardiac Nkx2.5 and myogenin after a short culture period (P₀–P₁). Freshly isolated, nonsorted cells expressed myogenin, but not Nkx2.5. Negative and positive controls, derived from skeletal (sm) or cardiac muscle (cm) tissue, were routinely included in each PCR run and are represented on the extreme right of the lanes. (b) Likewise, CD34⁺/CD90⁻ cells expressed cardiac (cTn-T, cTn-I, Cx43) and skeletal muscle (MyoD, myogenin) mRNAs, shown with representative PCR products derived from the three cell populations. (c) Quantitative analysis of mRNA expression revealed similar levels of cardiac (Nkx2.5, cTn-T, cTn-I, Cx43) and skeletal muscle mRNAs (myogenin, Myf-5, MyoD) in CD90⁺/CD34⁻ (white), DN (striated), and CD34⁺/CD90⁻ cells (black), but only low expression levels of endothelial mRNAs (vWF, CD31) in myoblasts (gray; cell fractions and myoblasts derived from n = 3 biopsies each). DN cells did not show cTn-I expression. Of note, CD34⁺/CD90⁻ cells exhibit a relatively low level of skeletal muscle mRNA expression. Bars indicate mean ± SEM values of quantitatively determined molecules of mRNA-derived cDNA per sample.

Figure 3

Comparison of individual baseline (pre) and post-transplantation (post) left ventricular ejection fraction (LVEF) values in the CD34⁺/CD90⁻ cell, control, and myoblast groups. LVEF was significantly ameliorated in the CD34⁺/CD90⁻ group (diamonds) by an average of +6.8% (2.9, 10.7, P = 0.001) whereas it deteriorated in the control group [circles, −7.9% (−11.8, −4.0, P = 0.0003)] and only decreased by an average of −0.1% (−4.0, 3.8) in the myoblast group (triangles).

Figure 4

Quantification of fibrosis by Sirius Red staining (upper panels) and vascularization by immunostaining against rat endothelial cell antigen (lower panels) in the infarct zone of medium-, myoblast-, or CD34⁺/CD90⁻ cell-injected hearts. CD34⁺/CD90⁻ cell-treated hearts showed significantly less fibrosis than the control (^*P = 0.0009) and myoblast group (^**P = 0.05). Scale bar = 400 µm. Both CD34⁺/CD90⁻ cell- and myoblast-treated groups demonstrated a higher vessel density than the control group (CD34⁺/CD90⁻ versus control, ^***P < 0.0001; myoblasts versus control, ^**P = 0.008). Furthermore, vascularization in CD34⁺/CD90⁻ cell-transplanted hearts was greater than in myoblast-treated hearts (^*P = 0.0175). Bar graphs are presented as mean ± SD. The numbers within the bar graphs indicate the number of hearts, each of which was analyzed in 10 random fields out of 18 cryosections derived from six different levels of the cryosectioned hearts. Scale bar = 200 µm.

Figure 5

Phenotype of transplanted cells and assessment of inflammatory and immune reactions in host hearts. After 1 month, only very few human cells could be detected as revealed by lamins-A/C staining (upper panel). Scale bar = 50 µm. Transplanted CD34⁺/CD90⁻ cells expressed fast myosin (scale bar = 100 µm) and CD31 (scale bar = 50 µm). ED1-positive macrophages (middle panel) were detected in similar amounts in hearts of the three groups (scale bars = 200 µm). CD3 expressing lymphocytes (lower panel) were scarce, presumably because effective immunosuppression (scale bars = 50 µm).