sdf1 Expression reveals a source of perivascular-derived mesenchymal stem cells in zebrafish

Abstract

There is accumulating evidence that mesenchymal stem cells (MSCs) have their origin as perivascular cells (PVCs) in vivo, but precisely identifying them has been a challenge, as they have no single definitive marker and are rare. We have developed a fluorescent transgenic vertebrate model in which PVC can be visualized in vivo based upon sdf1 expression in the zebrafish. Prospective isolation and culture of sdf1(DsRed) PVC demonstrated properties consistent with MSC including prototypical cell surface marker expression; mesodermal differentiation into adipogenic, osteogenic, and chondrogenic lineages; and the ability to support hematopoietic cells. Global proteomic studies performed by two-dimensional liquid chromatography and tandem mass spectrometry revealed a high degree of similarity to human MSC (hMSC) and discovery of novel markers (CD99, CD151, and MYOF) that were previously unknown to be expressed by hMSC. Dynamic in vivo imaging during fin regeneration showed that PVC may arise from undifferentiated mesenchyme providing evidence of a PVC-MSC relationship. This is the first model, established in zebrafish, in which MSC can be visualized in vivo and will allow us to better understand their function in a native environment.

Keywords: Angiogenesis; Mesenchymal stem cells; Multipotential differentiation; Pericytes; zebrafish.

Figure 1

The sdf1a:DsRed transgene illuminates perivascular cells. (A) Caudal fin of 10-week old sdf1a:DsRed zebrafish at 5×. Scale bar represents 200 microns. Rightmost panel shows live fluorescent microscopy of the tail-fin in a 10-week old fli1:EGFP crossed to sdf1a:DsRed shows a population of sdf1a^DsRed cells as perivascular in anatomical location. Scale bar represents 500 microns. (B) Confocal microscopy of a caudal tail fin from a fli1:EGFP × sdf1a:DsRed transgenic zebrafish shows enhanced discrimination of sdf1^DsRed cells. Scale bars represent 50 microns. (C) Live confocal microscopy of perivascular cells followed by 3D surface rendering of a z-stack using the Zeiss Zen software. Scale bar represents 20 microns. (D) Immunohistochemistry of fli1:EGFP × sdf1a:DsRed caudal tail fin cross-section. Co-staining was performed with anti-EPG-FITC and anti-DsRed primary antibodies. Anti-DsRed detection was performed using a secondary antibody, donkey anti-rabbit-Cy3. Scale bar represents 50 microns.

Figure 2

Perivascular sdf1a^DsRed cells express markers of perivascular cells and can be cultured-expanded. (A) The caudal tail fins of adult fli1:EGFP × sdf1a:DsRed animals (n = 8 – 12) were amputated and treated with collagenase to produce a single cell suspension. Shown is a representative scatter plot of green fli1^EGFP endothelial cells and sdf1a^DsRed perivascular cells. (B) The fli1^EGFP endothelial cells and sdf1a^DsRed perivascular cells were sorted by flow cytometry and RNA isolated for qRT-PCR. N = 3 experiments, * and ** represent p < 0.05 and p < 0.01 respectively from a Student’s t-test. ND, not detected. (C) Phase-contrast microscopy of culture-expanded perivascular sdf1a^DsRed cells 6 weeks after primary isolation. Scale bar represent 100 microns. (D) RNA was extracted from cultured sdf1a^DsRed cells and RT-PCR performed for MSC and hematopoietic markers. Shown are the zebrafish gene names as well as the human cluster of differentiation (CD) designations. Whole kidney marrow (WKM) cells were used as a positive control. (E) RT-PCR showing FGF-family gene receptor expression and cumulative sdf1a^DsRed cell counts cultured in increasing amounts of bFGF.

Figure 3

Differentiation of cultured sdf1a^DsRed cells into mesodermal lineages. (A’–C’) Cells were cultured in adipogenic, chondrogenic, or osteogenic promoting media for 4 weeks. Relative gene expression was determined by qRT-PCT and ΔΔC_T calculation relative to undifferentiated cells as shown. N = 3 experiments, * and ** represent p < 0.05 and p < 0.01 respectively from a Student’s t-test. (A”–C”) Show examples of PVC line H differentiated and stained Oil-Red-O, toluidine blue, or Alizarin red to indicate adipogenic, chondrogenic, and osteogenic cells respectively. (A”’ – C”’) Show examples of PVC line Y differentiated and undifferentiated followed by staining with Oil-Red-O, toluidine blue, or Alizarin red to indicate adipogenic, chondrogenic, and osteogenic cells respectively. Scale bars represents 100 microns in adipogenic and osteogenic images and 500 microns in chondrogenic images.

Figure 4

Cultured sdf1a^DsRed cells support engraftable hematopoietic cells. (A) Genes known to be important for the maintenance of hematopoietic cells were amplified by RT-PCR. Shown are the results from two of three sdf1a^DsRed cell lines. (B) 50,000 WKM from bactin2:GFP fish were co-cultured on a monolayer of sdf1a^DsRed cells (lines H, X, Y) for 16 days followed by enumeration using a flow cytometry and counting beads (n = 15/group). * indicates p-value < 0.001 from a Student’s t-test. (C) Two days after 20 Gy irradiation, wild-type zebrafish were transplanted with 50,000 bactin2:GFP WKM from freshly isolated donors or previously co-cultured with sdf1a^DsRed PVC from Line H. Engraftment analysis was performed 7 days after transplant by flow cytometry for GFP (n = 5 – 7 animals/group). Student’s t-test showed no significant difference between the groups (p = 0.65). (D) WKM from bactin2:GFP transgenics was co-cultured with sdf1a^DsRed PVC cells from Line H for 16 days versus cells on plastic alone. The entire contents of a well were used for transplant as above and engraftment determined at seven days post transplant (n = 10 – 12 animals/group). * indicates p-value < 0.001 from a Student’s t-test.

Figure 5

Proteomics of cultured sdf1a^DsRed cells and hMSC. Whole cell lysates were prepared and subjected to 2-dimensional liquid chromatography as described in the supplemental methods. Tandem mass spectrometry was then used to identify peptides on a capillary LC-nanoESI-Orbitrap XL mass spectrometer. Peptide matching to full-length proteins was performed using Sequest (version 27, rev. 12). (A) Absolute numbers of spectra, peptides, and proteins identified. (B) Results of IPA® analysis showing distribution of proteins by cellular compartment. (C) IPA® analysis showing top hits in the “Physiological System and Development” category. (D) Venn diagram showing common proteins identified between cultured sdf1a^DsRed cells and hMSC, their cellular compartment, and IPA analysis showing top hits in the “Physiological System and Development” category. (E) Flow cytometry histograms of hMSC for three novel proteins identified in both cultured sdf1a^DsRed cells and hMSC. Grey histograms show the isotope control antibody.

Figure 6

Co-culture of sdf1a^DsRed cells with HUVECs shows vascular branching. 15,000 HUVECs were labeled with CellTracker™ Orange and 10,000 cultured sdf1a^DsRed cells were labeled with CellTracker™ Green. Cells were placed atop 100 µL of matrigel in a 96-well plate in the presence of EGM-2 media for 10 – 16 hours prior to imaging. (A) HUVECs and two sdf1a^DsRed cell lines cultured alone. (B) Co-culture of HUVECs with sdf1a^DsRed cells and imaged under the appropriate filters. Scale bars represent 1000 microns.

Figure 7

The contribution of sdf1a^DsRed cells to fin regeneration. The caudal tail of sdf1a^DsRed × fli1:EGFP double transgenic was amputated on day 0 (amputation plane indicated by the white line). The same animal was anaesthetized and imaged using the appropriate filters at the time points given. Photos are representative of the same regeneration experiment performed in ten animals. Scale bars represent 100 microns.