Replication stress increases mitochondrial metabolism and mitophagy in FANCD2 deficient fetal liver hematopoietic stem cells

Fanconi Anemia (FA) is an inherited bone marrow (BM) failure disorder commonly diagnosed during school age. However, in murine models, disrupted function of FA genes leads to a much earlier decline in fetal liver hematopoietic stem cell (FL HSC) number that is associated with increased replication stress (RS). Recent reports have shown mitochondrial metabolism and clearance are essential for long-term BM HSC function. Intriguingly, impaired mitophagy has been reported in FA cells. We hypothesized that RS in FL HSC impacts mitochondrial metabolism to investigate fetal FA pathophysiology. Results show that experimentally induced RS in adult murine BM HSCs evoked a significant increase in mitochondrial metabolism and mitophagy. Reflecting the physiological RS during development in FA, increase mitochondria metabolism and mitophagy were observed in FANCD2-deficient FL HSCs, whereas BM HSCs from adult FANCD2-deficient mice exhibited a significant decrease in mitophagy. These data suggest that RS activates mitochondrial metabolism and mitophagy in HSC.


Introduction
FA is an inherited BM failure with hematologic disease onset around school age (1). Yet, murine models and studies on FA patients have revealed that genetic mutation of FA genes affects embryonic HSC function (2-4). FA HSCs exhibit a decrease in number in the fetal liver where HSCs should rapidly expand. Compared to predominantly quiescent adult BM HSCs, FL HSCs are highly proliferative around E13.5 and show a strong RS response. We previously showed that FA FL HSCs exhibited significantly greater RS compared to wild-type (WT) FL HSCs due to delayed recovery from replication fork collapse (5).
Recently, we and others have highlighted mitochondrial metabolism as a critical component for maintenance in adult quiescent HSCs (6,7). Mitochondrial membrane potential (MMP) is lower in HSCs than in progenitor cells but HSCs with large mitochondrial mass exhibit high stem cell potential (8). Due to their proliferation, FL HSCs exhibit higher mitochondrial activity compared to adult HSCs (9). Mitochondria oxidative phosphorylation and TCA cycle-associated gene expression are also higher in FL HSCs compared to adult BM HSCs. Others previously reported that mitochondrial metabolism and mitochondria clearance by mitophagy was reduced in FA cells (10)(11)(12). Moreover, HSC number in FA mice were rescued by treatment with the mitochondrial complex I inhibitor, metformin (13).
To understand how mitochondrial metabolism affects FA pathophysiology, we hypothesized that RS during FL HSC expansion modulates mitochondrial metabolism. Using pharmacologically induced RS, we observed temporary alterations in MMP and increased mitophagy. We then analyzed FL and BM HSCs from Fancd2 knockout (Fancd2 KO) mice. Strikingly, MMP and mitophagy were elevated in FANCD2-deficient FL HSCs. In contrast, mitophagy was downregulated in adult BM HSC. Taken together these observations indicate that the pathophysiological RS in FANCD2-deficient fetal HSCs boosts mitochondrial metabolism and mitophagy.
All data were measured by CytoFlex FACS analyzer (Beckman Coulter). Analysis is performed with FlowJo software.

CFU assay
HSCs were sorted by FACS Aria II or III (BD) and 500 cells are seeded onto 3ml of mouse methylcellulose media (R&D Systems, HSC007) with APH (50-100ng/ml) or imTOR of Rapamycin (Thermo PHZ1235, diluted with DMSO for 10mM and used as 1:1000, final concentration 10mM) or iTgf-b of SD-208 (CAYMAN 16619, diluted with DMSO for 10mM and used as 1:1000, final concentration 10mM). The media was equally divided into three 3.5cm dishes and incubated at 37°C. Colony numbers were counted after 14 days.

Statistics
Statistical analyses were performed with t-test and one-way ANOVA (if the parameter is over 2) by using prism software (GraphPad).

RS temporarily increases HSC mitochondrial activity and mitophagy in vivo
We previously described that FL HSCs experience RS due to their proliferative nature. We also showed that RS responses are increased in Fancd2 KO FL HSCs compared to BM HSCs from adult Fancd2 KO mice ( Figure S3A) (5). Here we tested whether RS alters HSC mitochondrial activity and mitophagy. Administration of 5 fluorouracil (5-FU) to WT mice depletes lineage cells and causes subsequent proliferation and RS in adult BM HSCs (15). It was previously described that 5-FU treatment changes Sca-1 expression so that we excluded Sca-1 marker and measured as CD150 + CD48-LK Hematpoietic Stem and Progenitor Cells (HSPCs) (15). After 6 days of 5-FU administration, HSPCs exhibited a significant increase in MMP (Figures 1A, S1A-C), which returned to baseline by day 12 when cells no longer proliferate. In parallel to changes in MMP, mitochondrial reactive oxygen species (mtROS) in HSPCs were also elevated at day 6 and returned to baseline by day 12 ( Figure 1B).
We next analyzed the mitochondrial morphology in HSPCs. Alterations in mitochondrial morphology and distribution in HSCs have been previously reported in myelodysplastic syndrome (MDS) and leukemia; mitochondria in HSCs from MDS models diffuse and spread out while mitochondria in leukemic stem cells (LSC) exhibit polarity and are concentrated (16,17). Imaging flow cytometer (IFM) was utilized to calculate mitochondrial area, concentration, and polarity in individual HSPCs. Following the 5-FU challenge, the mitochondrial area decreased and diffused, producing a polar redistribution pattern on day 6 that continued through day 12 ( Figure 1C).
Adequate levels of mitophagy and lysosomal activity help maintain HSPCs potential (18)(19)(20). When we evaluated mitophagy and lysosomal activity, we found both increased on day 6, and HSPCs lysosomal activity remained high on day 12 ( Figures 1D, E). Furthermore, the long-term effect of experimental RS on HSPCs was analyzed. HSPCs from mice 1 month after 5-FU injection (1Mo HSC) continued to exhibit a significantly larger mitochondrial area ( Figure 1H), but comparable MMP and mtROS compared to control HSPCs ( Figures 1F, G). Together, these data indicate that in vivo RS drastically alters MMP, mtROS, mitophagy and mitochondrial distribution in adult HSPCs.

RS increases mitochondrial activity and mitophagy in cultured HSC
We next investigated changes in mitochondrial activity in cultured HSCs subjected to aphidicolin (APH) which induces RS by inhibiting DNA polymerase (21). After 7 days of culture APH-treated HSCs exhibited a significant decline in number compared to control HSCs (Figure 2A). APH-treated HSCs exhibited significantly fewer cell divisions compared to control HSCs ( Figure S1D). APH-treated HSCs did not exhibit an increase in apoptosis as shown by Annexin-V staining ( Figure S1E). Hematopoietic progenitor colony formation of HSCs was significantly increased with APH-treatment ( Figure S1F). MMP of APH-treated HSCs transiently increased on day 2 but normalized by day 7 ( Figure 2B). Both on day 2 and on day 7, APH-treated HSCs exhibited unaltered mtROS level ( Figure 2C). Mitophagy activity was also significantly upregulated in APH-treated HSCs at day 2 but comparable at day 7 ( Figure 2D). These data indicate that BM HSCs under experimental RS in an ex vivo culture upregulate MMP and mitophagy.

Fancd2 KO BM HSC exhibits decreased mitophagy
Next, we focused on whether Fancd2 deficiency affected mitochondrial metabolism in quiescent adult BM HSCs, which do not exhibit spontaneous RS (5). While there was variation, Fancd2 KO BM HSCs exhibited comparable MMP to Fancd2 WT BM HSCs and HSPCs ( Figures 3A, S2A, B). mtROS was slightly upregulated in Fancd2 KO BM HSCs but was not altered in HSPCs (Figures 3B, S2C). Mitochondria distribution was diffused and not polar in Fancd2 BM KO HSCs ( Figure 3C). These data indicate mitochondrial activity in FANCD2deficient BM HSCs was comparable to WT cells. We further observed that mitophagy activity significantly decreased in Fancd2 KO BM HSCs ( Figures 3D, S2E), whereas lysosome activity did not change in Fancd2 KO BM HSCs (Figures 3E, S2F). Consistent with existing reports (10), these data illustrate that FANCD2-deficiency significantly decreases mitophagy in quiescent adult BM HSCs.

FANCD2-deficiency increases metabolic activity and mitophagy in FL HSC
Fetal HSCs show increased mitochondrial metabolisms compared to adult BM HSCs (9). Here, we focused on how FANCD2-deficiency changes mitochondrial metabolism in the rapidly proliferating FL HSCs. Unlike Fancd2 KO BM, FL HSCs showed an increased RS response, illustrated by characteristic gains in ssDNA ( Figure 4A) (5). We first measured MMP and found it to be significantly higher in Fancd2 KO FL HSCs and HSPCs compared to WT and Fancd2 heterozygous (HET) (Figures 4B, S3A-G), while mtROS levels were unchanged in Fancd2 KO FL HSCs ( Figure 4C). Further, Fancd2 KO FL HSPCs exhibited a significant increase in mitochondrial area, while the distribution of mitochondria was concentrated and polarized, indicating an overall increase in mitochondrial activity ( Figure 4D). We also measured mitophagy and lysosome acidification in FL HSC. In contrast to adult BM HSCs, mitophagy and lysosome acidification were all upregulated in Fancd2 KO FL HSCs and HSPCs ( Figures 4E, F, S3H, I). mTOR signaling is closely associated with mitochondrial metabolism and lysosome biogenesis in HSCs (22). We thus analyzed changes in phosphorylated 4EBP (p4EBP), a direct substrate of mTOR and OPP incorporation as a measure of global protein synthesis in Fancd2 KO FL HSCs. Both measures of metabolic activity were significantly increased (Figures 4G, S3J). These data indicate that FANCD2 deficiency stimulates mitochondrial activity during fetal development in concert with an increase in mTOR signaling. Additionally, we performed CFU with treatment of mTOR inhibitor (imTOR) of Rapamycin or Tgf-b inhibitor (i Tgf-b) which rescued the RS of Fancd2 KO FL HSPCs (5). Result showed CFU number was significantly decreased in imTOR treated Fancd2-KO HSC compared to untreated Fancd2-KO HSC indicating upregulated metabolism protect from RS ( Figure 4H). Collectively, our data indicate the RS response results increasing metabolism including mTOR pathway, mitochondrial activity and mitophagy and it sustains HSC function in FANCD2-deficient FL HSCs. These results mirror experimental observations following experimental RS in WT BM HSCs ( Figure S3K).

Discussion
In this study, we extended our observations on the unique fetal HSC phenotype in FA. We showed that fetal RS, in the context of FA results in a concurrent increase in mitochondrial metabolism and mitophagy. In contrast, adult quiescent BM HSCs show lower levels of mitochondrial metabolism and decreased mitophagy in Fancd2 KO. These observations provide an in vivo link of RS with mitochondrial metabolism and mitophagy. In 5-FU-treated mice, prominent mitochondrial changes were observed at day 6 after administration, the phase when HSCs are subjected to proliferation (15). Both FL HSCs and BM HSCs respond to RS by increasing mitochondrial metabolism and mitophagy, which suggests an inherent HSC response to RS. However, the change in mtROS and morphology in FANCD2-deficient FL HSCs and 5-FU BM HSCs differed. FL HSCs exhibit higher mtROS levels compared to BM HSC (9). mtROS level closely reflects mitochondrial quality and investigating how fetal and adult HSCs differ in mitochondrial quality control processes would be addressed in the future. We previously described the role of RS in decreasing FA FL HSCs (5). Metformin, an inhibitor of mitochondrial complex I, was reported to ameliorate adult FANCD2-deficient HSC decline in number and potential (13) suggesting a link between FA and fetal mitochondrial metabolism. However, low placental penetration of metformin prevented its use to investigate the effect on FL HSCs (data not shown).
Previous reports showed defective mitophagy as a non-canonical phenotype of FA cells and illustrate the role of several FA proteins in mitophagy (10, 11). Our results revealed a progressive decrease in mitophagy through development. While a mechanism has not been established, the investigation should be done in the future.
The direct correlation between RS, mitochondrial metabolism and mitophagy is difficult to explain. Our data indicate that mTOR signaling and mTOR-related lysosomal biogenesis may sense RS and increase overall metabolic activity including c-Myc (23) and ribosome biogenesis (24), though further investigation is needed. In conclusion, this manuscript provides evidence for the involvement of RS in the metabolic regulation of HSCs by FANCD2 deficiency and highlights the unique characteristics of the FL HSCs compared to adult BM HSCs.

Data availability statement
The original contributions presented in the study are included in the article/Supplementary Material. Further inquiries can be directed to the corresponding authors.

Ethics statement
The animal study was reviewed and approved by TWMU animal experiment committee.