Trisomy 21-Associated Abnormalities in IGF Signalling and the Fetal Microenvironment Both Contribute to Disruption of Fetal Hematopoiesis in Down Syndrome

In Down syndrome (DS), trisomy 21 (T21) causes perturbation of fetal liver (FL) hematopoiesis leading to expansion of megakaryocyte-erythroid (MK-E) progenitors. On this background, FL hematopoietic stem/progenitor cells (HSPC) may acquire GATA1 gene mutations which usually present as a neonatal preleukemic condition, Transient Myeloproliferative Disorder (TMD). TMD develops only in fetal life and most cases spontaneously regress after birth, suggesting unique features of the T21 fetal microenvironment may be important in driving both abnormal DS FL hematopoiesis and mutant GATA1 clones in TMD. Previous work showed activation of insulin-like growth factor (IGF) signalling in TMD and DS-AMKL and fetal-specific IGF dependence of murine MK development implicating T21 in altered fetal IGF signalling.

To investigate the role of IGF signalling in fetal MK-E development in DS FL without GATA1 mutations, we performed gene expression profiling (GEP) of DS FL (n=5) and normal FL CD34+ cells (n=3). GSEA showed significant enrichment of IGF targets. We therefore measured expression of the principal IGF receptor, IGF1R, on DS FL HSPC (n=6). Surface IGF1R was uniformly expressed on all DS FL HSPC subpopulations. However, there was no difference in IGF1R surface or gene expression between DS and normal FL HSPC (n=8) suggesting that increased IGF1R expression on TMD and DS-AMKL cells is an early event and may reflect their fetal rather than T21 origin. To determine if the enrichment of IGF targets in DS FL HSPC might reflect increased IGF production in the FL microenvironment, we compared IGF expression in DS FL mesenchymal stromal cells (MSC)(n=4) and hepatocytes with normal FL MSC (n=4). Both DS and normal FL MSC expressed large amounts of IGF2, but little or no IGF1 with no difference between DS and normal MSC. Similarly, IGF2 was expressed by IHC at similar levels in both DS and normal fetal hepatocytes and MSC whereas IGF1 was barely detectable. Normal and DS fetal bone marrow (BM) MSC also expressed high levels of IGF2, but not IGF1 in contrast to adult BM MSC. Thus, IGF2 is the main IGF produced in the human fetal hematopoietic environment. Consistent with a specific role in DS FL hematopoiesis, IGF2 caused marked proliferation of DS FL clonogenic MK and MK-E cells (n=3) and stimulated MK proliferation in liquid culture while in normal FL, IGF2 stimulated EPO-independent BFU-E and Pre-BFU-E but had no effect on MK cells. Since differential effects of IGF2 on DS/normal FL HSPC might be due to altered expression of the Insulin Receptor (InsR), which binds IGF2 with low affinity, we compared InsR surface and gene expression in DS and normal FL HSPC but found no significant difference. IGF2 also binds to IGF2R, a negative regulator of IGF2, however there was no difference in IGF2R expression between DS and normal FL HSPC. Co-culture of normal FL CD34+ cells with DS or normal FL MSC (n=3) promoted MK and erythroid cell proliferation and differentiation. Since the effect of DS FL MSC on normal FL CD34+ cells was significantly greater (~4-fold) than that of normal FL MSC, the data above suggest this is unlikely to be due to IGF2 alone. To investigate whether T21 alters the MSC secretome, we performed GEP of DS FL MSC (n=3) and compared this with normal FL MSC (n=5). Both normal FL and DS FL expressed known hematopoietic cytokine genes, including SCF, CXCL12, G-CSF, M-CSF, Flt3 and IL-6, as well as IGF2 but not IGF1, EPO or TPO. Of the top 80 differentially expressed genes (Fold Change>2), ~40% encode secreted proteins, including several which affect IGF availability.

Conclusion: T21 perturbs FL hematopoiesis through both HSPC-intrinsic mechanisms, including changes in IGF responsiveness, and through alterations to the FL microenvironment.