Abstract
We have previously shown that uniquely during the first trimester mesenchymal stem cells (MSC) co-circulate in fetal blood (FB) with hemopoietic stem cells (HSC) in a high MSC:HSC ratio (1:1) with a frequency of ~4 per 103 nucleated cells. The numbers of circulating MSC decline dramatically by the end of the first trimester as MSC and HSC home to the fetal liver (FL) and fetal bone marrow (FBM) as definitive hemopoiesis becomes established in these sites. The function of these circulating first trimester MSC, not hitherto described in mice, in the ontogeny of human hemopoiesis is unknown. However, their high frequency and close association with HSC suggests that one role is likely to be the production of growth factors to maintain HSC viability and self-renewal. To investigate this we assessed the ability of human fetal MSC from first trimester FB to (i) support HSC as assessed in LTC-IC assays of cord blood (CB)-derived MACS-enriched CD34+ cells; and (ii) produce cytokines acting on early HSC as measured by quantitative RT-PCR. Fetal MSC were isolated from first trimester (6+0 to 14+4 weeks post-conception) FB (n=10), FL (n=5) and FBM (n=6) by adherence to plastic and characterised by morphology, immunophenotype and multilineage differentiation in vitro. Early passage FB, FL and FBM-derived MSC were used as stromal layers in limited dilution LTC-IC studies and analysed with L-Calc software. MSC were CD45−, CD34−, CD31−, CD14−, CD105+,SH3+, vimentin+, collagen I-, SMA±. MSC from all three sites supported LTC-IC. The frequency of LTC-IC was three-fold higher in CB grown on FB-derived MSC in comparison with FL and FBM MSC. To investigate the mechanism of increased numbers of LTC-IC supported by FB-derived MSC, we developed quantitative RT-PCR (Taqman probe™/SYBR Green) to compare the levels of early acting growth factors produced by MSC derived from FB, FL and FBM. Results were standardised against the numbers of copies of ABL in the sample (mean ± SEM). There were no significant differences in the amounts of LIF, BMP-4, SCF, IL-6 or IL-11 between MSC from different sites or gestational ages. Tpo was virtually undetectable and was not detectable by ELISA in supernatants from any of the MSC samples. However, FB MSC produced significantly greater amounts of TGFβ (633 ± 28 units/ABL copies) compared to FL (386 ± 32 units/ABL copies; p=0.004) and FBM (496 ± 34 units/ABL copies; p=0.01) and compared to adult BM MSC (316 units/ABL copies). In addition, VEGF expression by the same FB MSC was significantly lower (482 ± 53) than in FL (760 ± 41; p=0.01) and FBM (994 ± 41; p=0.03). FB MSC also expressed relatively low amounts of Flt3L (217 ± 20) compared to FL (245 ± 59; p=1.0) and FBM (350 ± 131; p=0.9); interestingly expression of Flt3L by FBM MSC increased with gestation (888 units/ABL copies by 14 weeks gestation; p=0.0091) to levels comparable to adult BM MSC (693 units/ABL copies). Together these data indicate that there are ontogeny-related differences in the function of human fetal MSC, both with respect to their ability to support LTC-IC and to the profile of growth factors which they produce. We hypothesise that circulating FB MSC, by producing high levels of TGFβ together with low levels of Flt3L, VEGF and Tpo in a developmentally-regulated manner, maintain co-circulating HSC in a quiescent state during homing from the AGM to FL and FBM.
Author notes
Corresponding author