PDK1 (3-phosphoinositide-dependent protein kinase 1) plays an important regulatory role in B cells, T cells and platelets. We previously also showed that PDK1 play a vital role in HSC function probably via regulating ROS levels. However, less is known about how PDK1 acts in the fetal liver (FL) hematopoiesis during embryonic development, as the FL is the primary fetal hematopoietic organ and the main site of HSC expansion and differentiation. Here, we deleted the PDK1 gene in hematopoietic cells by crossing Vav-Cre transgenic mice with PDK1f/f mice. Using a transplantation assay, we found that HSCs from the E15.5FL of Vav-Cre;PDK1f/f embryos were severely impaired when compared with HSCs from PDK1f/f or PDK1f/+ FLs. When E15.5 WT FL cells were transplanted at 0.01 embryo equivalent (ee) per recipient, 9 out of 11 recipients (from 5 independent experiments) were repopulated with an average long-term chimerism of 60.9 ± 9.4% while none of the recipients were repopulated in the Vav-Cre;PDK1f/f group, and with markedly reduced chimerism (0.02 ± 0.01%).

Additionally, we showed that there were more FL HSCs in an apoptotic state and active cell cycle in PDK1 -deficient embryos than in control embryos. Apoptotic analysis of sorted HSCs revealed an approximate two-fold increase in the percentage of Annexin V+7-AAD- (A+7-) and Annexin V+(A+) in Vav-Cre;PDK1f/f HSCs, when compared with wild-type controls. Ki67 staining for DNA content of sorted HSCs revealed an increase in the percentage of HSCs in the S/G2/M phase in Vav-Cre;PDK1f/f FLs in comparison to wild type FLs. Interestingly, the enhanced apoptosis and proliferative ability of the Vav-Cre;PDK1f/f HSCs did not lead to a significant difference in the overall frequency and number of HSCs in the FL.

By comparing the expression profiles of FL-derived HSCs in Vav-Cre;PDK1f/f embryos to the control HSCs, we found that the BH3-only protein PUMA was highly expressed in the Vav-Cre;PDK1f/f group. PUMA is known to participate in the release of mitochondrial apoptogenic proteins such as cytochrome c to activate apoptosis process. Indeed, our results showed higher expression of Apaf1 and caspase-3 in the Vav-Cre;PDK1f/f group than in the control group. Furthermore, we also demonstrated that the expression of FoxO3a was higher in PDK1 -deficient HSCs. These data suggests that the Akt-FoxO3a-PUMA axis may participate in regulating HSC apoptosis at the E15.5 FL stage.

Interestingly, FoxO1 expression was found to be lower in PDK1 -deficient HSCs. Previous studies by others have shown that the down-regulation of FoxO1 was caused by phosphorylation of Akt, which led to the FoxO1 nuclear exclusion and degradation by the proteasome. Indeed, we observed an increase in the levels of Akt phosphorylation at Ser473 in PDK1 -deficient embryos probably due to the compensation for the loss of Akt-Thr308 phosphorylation. Additionally, FoxO1-mediated cell cycle arrest is linked with cyclin D1 and cyclin D2 suppression in mammal. Here, we found that the Cyclin D was over-expressed in the PDK1-deficient group indicating that Akt-FoxO1-CCND may regulate the HSC cell cycle. Taken together, our findings support a critical role for PDK1 in maintaining FL hematopoiesis via regulating HSC apoptosis and cell cycle possibly via the Akt-FOXO signaling pathways.


No relevant conflicts of interest to declare.

Author notes


Asterisk with author names denotes non-ASH members.