Hematopoietic stem cells (HSCs) reside in a quiescent state, which is thought to preserve their genomic stability during aging. HSCs are forced to exit this so-called dormant state and enter into cycle in response to stress stimuli such as infections or severe bleeding. This situation may provoke high levels of proliferative stress in HSCs and a subsequent decline in stem cell function. We recently found that de novo DNA damage can be precipitated in HSCs in vivo by enforcing cell cycle progression using agonists that mimic physiologic stress, such as interferons, G-CSF, TPO or serial bleeding, (Walter et al., Blood, 122, 21:799). The Fanconi anemia (FA) DNA repair pathway is an important route via which this replication damage is resolved in HSCs in vivo. In FA deficient mice, DNA damage repair was impaired, provoking HSC depletion and severe aplastic anemia (p<0.01) upon serial treatment with the synthetic double-stranded RNA mimetic polyI:polyC (pI:pC). Here, we sought to identify the mechanistic basis of the stress-induced DNA damage acquisition and concomitant HSC attrition in vivo.

Activated HSCs exhibited elevated mitochondrial membrane potential, indicative of increased energy production via oxidative phosphorylation (>2-fold increase, p<0.01). Next, to determine whether there was an associated increase in intracellular reactive oxygen species (ROS) production, we made use of genetically encoded fluorescent biosensors to detect the status of specific redox couples within different HSC compartments in vivo. Activated HSCs demonstrated increased levels of oxidized mitochondrial glutathione (2.3-fold increase, p<0.01) and cytoplasmic hydrogen peroxide (1.6-fold increase, p<0.05) compared to dormant HSC controls. These enhanced ROS levels directly correlated with elevated 8-Oxo-dG lesions on the DNA of HSCs that had been activated into cycle in vivo(>1.3-fold increase, p<0.05). Finally, retroviral over-expression of ROS-detoxifying enzymes completely rescued gH2AX foci formation in cycling HSCs, demonstrating a direct functional link between stress-induced DNA damage and altered redox biology.

We next performed live cell video imaging on individual WT and Fanca-/- LT-HSCs in vitro in order to track cell fate decisions upon exit from quiescence. In the first division upon exit from quiescence, Fanca-/- HSCs were frequently observed to undergo abnormal mitoses while this was not evident in WT HSCs. At this time point, we observed elevated DNA damage in Fanca-/- HSCs as measured by gH2AX, 53BP1 and RAD51 foci, as well as increased ROS-induced 8-Oxo-dG lesions (>5-fold increase, p<0.01). HSCs from Fanca-/- mice demonstrated a significantly higher rate of replication-dependent cell death following the first division (24% vs. 6%, p<0.05%) suggesting that apoptosis is the major route via which HSCs are lost in response to stress-induced DNA damage.

Taken together, these data strongly implicate stress-induced exit from dormancy as a cause of physiologic DNA damage in HSCs in vivo. Under stress conditions, the increased energy demand of cycling stem cells leads to elevated levels of ROS in mitochondria and cytoplasm, which is a direct source of DNA damage. If unresolved by the FA-dependent DNA damage response, this DNA damage accumulates in the cell and provokes apoptotic cell death. This recapitulates the highly penetrant bone marrow failure syndrome in FA patients and suggests that their HSCs are lost due to an aberrant response to HSC activation, most likely as a consequence of infection or other physiologic stress. These data provide a novel link between stress hematopoiesis, ROS, DNA damage and HSC loss and may have important clinical implications in the study of age-related hematopoietic defects in both FA and non-FA patients. Moreover, these data provide the first evidence that FA knockout mouse models can be utilized to accurately recapitulate the etiology of bone marrow failure through the progressive application of stress-induced alterations in HSC function that mimic usual physiologic stressors such as infection.


No relevant conflicts of interest to declare.

Author notes


Asterisk with author names denotes non-ASH members.