Fig 1.
Basic strategies of sample preparation and FACS analysis.
A. FACS plots of Ficoll-enriched PBMCs vs. RBC lysis-based preparations. To prevent exclusion of large cells, a large mononuclear-cell gate was applied. B. Ratios of enrichment of total live PBMCs, lymphocytes, monocytes, HPCs (as defined in Fig 2) and CECs (as defined in Fig 3) by Ficoll vs. RBC lysis. C-D. Examples of reduction of background (auto-fluorescence) by stringent exclusion of dead cells using Hoechst 33258 (C) and (residual) granulocytes (D). E. Basic procedure used, in addition to standard acquisition (1x106 total events), to enrich for target cells by discarding triple CD34/CD133/KDR negative events. Upper left plot: selection of CD34+ and KDR+ cells; lower left plot: selection of CD34+ and CD133+ cells. Right panels: result of a “Join (Or) gate” in FACS DIVA. KDR+ CD34+ (upper plot) and CD133+ CD34+ (lower plot) populations are better visualized. F. Delineation of a scale of FACS-based fluorescence levels. Negative peak = unstained or isotype control; low (dim) = within 1 log from negative; high (bright) = more than 1 log from negative and ranging from moderately high (1st to 2nd log above negative) to very high (3rd log or higher). “Medium” is sometimes used to define populations in between low and high.
Fig 2.
Identification and characterization of HPCs.
A. Upper panel: HPCs in UCB and AB are initially selected as CD34+/hi cells and subdivided into CD133hi, CD133low and CD133neg. Lower panel: discrimination between CD133 levels is partially missed using a less bright antibody. B. HPCs are refined by gating a single peak in the CD45low region. Left plot: CD45 gating on raw HPCs (CD34+/hi events). Right plot: CD45 gating on CD34+/hi events pre-refined by FSC/SSC. CD45 levels slightly increase with CD133 expression (see also S3 Fig), still remaining within the “low” gate. The % of match of raw or pre-refined HPCs with the CD45low gate is indicated in the plots for each population. C. In a FSC/SSC plot, HPCs appear as a tight cluster in between lymphocytes and monocytes. Upper panel: FSC/SSC gating on raw HPCs (CD34+/hi events). Lower panel: FSC/SSC gating on CD34+/hi events pre-refined by CD45 levels. The % of match of raw or pre-refined HPCs with the FSC/SSC gate is indicated in the plots for each population. Comparison of B (left vs. right plot) and C (upper vs. lower panel) shows that CD45 and FSC/SSC gating converge to the identification of pure HPCs: partial overlapping in HPC cleaning indicates that the two approaches identify the same population, however each approach also provides some independent contribution to HPC purification. The less pure and most affected by cleaning gates appears the CD133neg fraction. D. FACS plots showing the distribution of pure HPCs (CD34 vs. CD133 plot) in UCB vs. AB. E. CD133 and c-kit RT-PCR in CD34+CD133+ cells (n = 22); CD34+CD133- cells (n = 17) and CD34-KDR-CD133- PBMCs (n = 8) isolated from UCB. CD133 mRNA is detected in both CD133+ and CD133- HPCs, although at different levels. c-kit mRNA further confirms the HPC identity of the CD34+CD133- cells. F. Left panel: median frequency of HPCs in UCB (5,6 in 1*103 CD45+ PBMCs) and AB (0,97 in 1*103 CD45+ PBMCs, significantly lower than in UCB, Z = -4,9; p = 1,00E-06). Middle panel: percentage ± SD of the three CD133-based HPC subpopulations in UCB vs. AB. Right panel: RT-PCR confirms higher CD133 expression by total UCB-HPCs than AB-HPCs.
Fig 3.
HPCs: Analysis of c-kit and endothelial markers.
A. FACS analysis for c-kit vs. CD133 confirms the PCR results and refines HPC subdivision. In UCB, c-kit medium-high levels were observed in all the CD133-based subpopulations (with the exception of sporadic events). In AB, an additional c-kitneg-low cluster was observed within the CD133high subpopulation, while the highest c-kit levels were observed on the CD133neg cells (see also S3 Fig). B. Left: distribution of c-kitneg-low cells between the CD133-based subpopulations in AB. Right: quantification of c c-kitneg-low cells in AB vs UCB. Right:. C. KDR, CD144, CD146 are sporadically expressed by HPCs. CD105 is dimly expressed but only a fraction of cells crosses the negative gate (8%). The percentages of positive cells are indicated in the plots. D. Analysis of KDR vs. CD146 expression in HPCs (UCB): the markers do not match but identify independent subpopulations. For better visualization, we used a UCB sample with high frequency of HPCs and relatively high KDR expression on HPCS (see also Panels D and E in S3 Fig). E. Analysis of CD144 vs. CD146 expression in HPCs: the two markers mostly identify independent subpopulations (see also Panel F in S3 Fig).
Fig 4.
Identification, characterization, and quantification of CECs.
A. Initial selection of CECs (UCB) on a CD34 vs. KDR plot (1x106 events) by gating KDR+ cells at different CD34 levels. Lower panel: the acquisition of large sized samples was reached by discarding triple CD34/CD133/KDR negative events resulting in better visualization of CECs. B. Subsequently, true KDR+CD45neg CECs are selected in a CD45 vs. KDR plot. Most CECs belong to the CD34 high region (left plot), since KDR+ CD34low events are almost entirely CD45+ (middle plot). Right plot: the CD45 gate applied efficiently discriminates CECs from sporadic KDR+ HPCs (selected by CD34 and FSC-SSC). C. Comparison of UCB-CECs (left plot) and AB-CECs (right plot). For reference: HPCs are shown in gray. In UCB, most CECs cluster at CD34 levels above HPCs and are KDRmed-hi, while in AB CECs have significantly lower CD34 and KDR levels. D. Frequency of CECs in UCB (14/106 CD45+ PBMCs) vs. AB (1,9/106CD45+ PBMCs). In AB the levels are significantly lower (Z = -3,6; p = 3,00E-04). E. Marker expression by CECs. Upper left: 90% of KDR-selected CECs express CD146. Lower left: 90% of CECs express CD144 but only after signal amplification. Upper right: CD31 is highly positive in CECs, but does not discriminate CECs from HPCs. Lower right: CD133 is expressed by a minority of CECs. F. Upper left: Around 15% of the CECs appear positive for c-kit and c-kit positivity correlates with higher KDR and CD34 levels (lower left). Right panel: RT-PCR confirms the expression of c-kit by CECs at levels comparable with HPCs (n = 3) despite significantly lower surface expression.
Fig 5.
Definition of nucleated events and apoptosis in CECs and HPCs.
A. FACS analysis of live nuclear staining with Hoechst 33342. Upper plot: nucleated events are separated from non-nucleated ones using erythrocytes and mononuclear cells as a reference. Medium plot: lymphocytes and monocytes are 100% nucleated. Lower plot: both CECs (red) and HPCs (green) appear as nucleated cells. CECs show a bimodal distribution with a sub-G1 fraction (probably apoptotic, DNA-fragmented cells). B. Upper plot: representative Annexin V staining of CECs (selected using CD34-FITC, CD146-PE, CD45-VioGreen) and HPCs (selected using CD34-FITC and CD45-VioGreen) from UCB. Bottom left: percentages of apoptotic CECs and HPCs in AB and UCB. Bottom right: apoptosis in c-kit+ vs c-kitneg CECs.
Fig 6.
FACS analysis of markers with homogeneous/stable (A) and heterogeneous/unstable (B-C) expression in early (passage 2–3) and late (passage >6) OEC cultures. A. No signals for CD45 and CD14. Staining for CD105, CD31, CD144, CD146 was homogeneously high and stable. B. CD34, KDR, and c-kit expression was heterogeneous and decreased with ageing of the cultures. The expression of CD133 was sporadic, with the exception of a single outlier, and slightly decreased with time in culture. For each marker, three examples of early and late cultures, and the outlier, are shown. C. Example of positive correlation (early passage OECs) between CD34, KDR, and c-kit. The markers define a triple high cluster, which is lost with ageing of the culture.