Expression of MSC-derived S100A8 increased in cases of AML
In our previous study, we found a significant increase in the expression of BM MSC-derived S100A8 in a cohort of rapidly advancing murine models of AML (Fig. 1A–B), indicating that MSC-derived S100A8 plays a role in the regulation of leukemic disease progression. Here, we validated the expression of S100A8 in MSCs derived from BM with AML status. Volunteer-derived BM mononuclear cells were obtained and in vitro cultured to produce human BM MSCs (Fig. 1C). MSC expression of the marker (Fig. 1D–K)and differentiation functions were validated (Fig. 1L–M). After 48 hours of cocultivation with leukemic cells, MSCs showed an increase in S100A8 expression (Fig. 1N–O). Furthermore, MSC-derived S100A8 expression in MLL-AF9 AML mice increased similarly to that in normal mice (Fig. 1P). Furthermore, our analysis of human BM sections showed very little expression of S100A8 in healthy human MSCs, whereas MSCs derived from patients with AML expressed a lot of S100A8 (Fig. 1Q). The data presented above support the aberrant expression of S100A8 in the presence of AML, implying that S100A8 may be involved in the progression of AML.
MSC-derived S100A8 is essential for the growth of human AML cells
To further validate the role of MSC-derived S100A8 in AML development, we performed an experiment in which S100A8 was overexpressed in human BM MSCs (Fig. 2A–C) and then cocultured with the human AML cell line Kasumi-1. Overexpression of S100A8 in MSCs increased the growth of primary human AML cells (Fig. 2D) and facilitated their progression through the cell cycle (Fig. 2E). Furthermore, overexpression of S100A8 did not significantly alter the rate of apoptosis (Fig. 2F), but it increased the resistance of Kasumi-1 cells to etoposide (Fig. 2G). Then, leukemia cells from AML patients were used, and CD34+ leukemia cells were enriched using magnetic-activated cell sorting (MACS). CD34+ leukemia cells were cocultured with S100A8-overexpressing and control MSCs for 48 h. Then, leukemia cells were transplanted into immunodeficient mice, along with the MSC suspension (Fig. 2H). After 30 days, the WBC count in S100A8-overexpressing mice was higher than in the control group, as were the levels in the BM and SP (Fig. 2I–K). Finally, these findings indicate that S100A8 plays a role in the development and drug resistance of human AML cells.
MSC-derived S100A8 drives murine AML progression and maintains LSC function
To fully understand the role of S100A8 in leukemogenesis, MSC-specific S100A8 knockout (Prx1-Cre+/−;S100A8flox/flox) mice and control mice was created (Prx1-Cre−/−;S100A8flox/flox) by crossing S100A8flox/flox mice with Prx1-Cre mice (Figure S1), then MLL-AF9-induced murine AML model was constructed with S100A8 knockout and control mice as recipients (Fig. 3A–B). S100A8 deletion significantly improved the survival rate of AML mice (Fig. 3C). Not surprisingly, S100A8 knockout recipients had lower spleen (SP) weights than controls (Fig. 3D). Furthermore, we found that S100A8 knockout recipients had a significantly lower ratio of YFP+ leukemic cells in the PB, SP, and BM, as well as lower malignant cell and total WBC counts in the PB (Fig. 3E–H). Furthermore, the percentages and numbers of LSCs (YFP+c-Kit+Gr-1− cells) in PB, SP, and BM were significantly reduced after S100A8 deletion (Fig. 3I–M), as were L-GMPs (IL-7R/Lin−YFP+c-KithiCD34+CD16/32hi) in SP and BM (Fig. 3N–P). Finally, we examined the characteristics of LSCs in a murine AML model, and we discovered that S100A8 knockout recipients had more G0 phase cells and higher apoptosis rates than control mice (Fig. 3Q–R). Collectively, we can reasonably conclude that S100A8 plays an important role in promoting the maintenance of LSCs.
MSC-derived S100A8 regulates AML development via the TLR4/PI3K/Akt pathway
RAGE and TLR4 are the most common S100A8 receptors, and to investigate the mechanism by which S100A8 acts on leukemic cells, we added a RAGEE inhibitor (FPS-ZM1) or a TLR44 inhibitor (Resatorvid) to MSC and Kasumi-1 coculture systems, and the addition of TLR4 inhibitors resulted in the most significant decrease in leukemic cell proliferation (Fig. 4A–B), indicating that MSC-derived S100A8 acts primarily by binding leukemic cell TLR4. To delve deeper into the molecular mechanism, we conducted transcriptomic analyses. It revealed notable differences in gene expression in Kasumi-1 cells between the group cocultured with overexpressed S100A8 and the control group (Fig. 4C), as well as significant alterations in the PI3K/Akt signaling pathway (Fig. 4D–E). WB experiments revealed that phosphorylated PI3K and Akt protein levels were significantly higher in the S100A8 overexpression group but lower in the S100A8-deficient group, and PI3K inhibitor (LY294002) significantly reduced proliferation of leukemic cells (Fig. 4H). The most recent article reported that PI3K/Akt regulates oxidative stress to promote drug resistance in leukemia progression, and changes in the HIF-1 pathway were also revealed in our sequencing results (Fig. 4D). We then examined ROS levels in LSCs. Flow cytometry analysis revealed a lower intracellular level of ROS in LSCs from S100A8 knockout mice compared to the control group (Fig. 4I). TMRE staining was used to determine mitochondrial membrane potential. Flow cytometry analysis revealed a decrease in TMRE fluorescence in LSCs from S100A8 knockout mice compared to the control group (Fig. 4J). Furthermore, DNA damage was found to be lower in BM cells from S100A8 null AML mice compared to control cells (Fig. 4K). Overall, these findings showed that S100A8 promoted leukemia progression through TLR4/PI3K/Akt signaling.
MSC-derived S100A8 deletion is dispensable for normal hematopoiesis but inhibits HSC expansion under short-term stress
Impairment of normal hematopoiesis and leukemia progression are 2 closely related processes during leukemia development and are regulated by the BM niche. We discovered that removing S100A8 had no discernible effect on the frequencies and numbers of HSCs, hematopoietic progenitors (HPCs) (except for multipotent progenitors, MPPs) (Figure S2A–E), and mature cells were comparable in the BM and PB of control and knockout mice (Figure S2F–I). Thus, MSC-derived S100A8 was not required for homeostatic hematopoiesis in mice. However, in the MLL-AF9 mouse model, we discovered that S100A8 knockout recipients had more HPCs and mature myeloid cells than control group mice (Figure S2J–R), implying that S100A8 may regulate hematopoiesis under stress conditions.
Interestingly, we found that S100A8 was significantly upregulated in MSCs after 5-fluorouracil (5-FU) and irradiation (IR) exposure (Figure S3). As expected, the frequencies and numbers of LSKs, long-term HSCs (LT-HSCs), and short-term HSCs (ST-HSCs) were nearly all lower in S100A8-/- BM at day 14 after 5-FU treatment and IR exposure than in controls (Fig. 5A–D). Notably, 5-FU treatment and IR exposure resulted in lower percentages and numbers of granulocyte/macrophage progenitors (GMPs), while decreased megakaryocyte/erythroid progenitors (MEPs) was observed in S100A8-/- mice after IR exposure, whereas common myeloid progenitors (CMPs) showed no significant difference between the two groups after 5-FU treatment and IR exposure (Fig. 5E–F). Consistent with these findings, S100A8-/- mice had significantly fewer mature myeloid cells following 5-FU and IR treatment in both BM and PB (Fig. 5G–J). Overall, our findings indicate that S100A8 ablation promotes short-term HSC expansion under hematopoietic stresses.
Deletion of MSC-derived S100A8 increases HSC quiescence under hematopoietic stress
Given the observation that S100A8 causes short-term expansion of HSCs during hematopoietic stress, we hypothesized that S100A8 may modulate HSC quiescence. As expected, HSCs from S100A8 knockout mice had a higher percentage in the G0 phase but a decreasing trend in the G1 and S/G2/M phases after 5-FU treatment and IR exposure (Fig. 6A–C). Furthermore, an in vivo BrdU incorporation assay revealed that deletion of S100A8 reduced HSC proliferation of HSCs in mice following 5-FU and IR challenge, while apoptosis rates were comparable between S100A8 knockout and control HSCs (Fig. 6D–H). These findings suggest that S100A8 promotes hyperactivation of HSCs in response to stress stimuli. To see how S100A8 affected HSCs functionality, we used a colony formation assay. The results showed that the number of colonies formed in the group with S100A8 deletion was significantly higher after treatment with 5-FU and IR (Fig. 6I–J). These findings suggest that S100A8 promotes the entry of HSCs into the proliferation cycle as a means of short-term hematopoietic recovery under stressful conditions, albeit at the expense of impairing the general function of HSCs.
To see if S100A8 affects long-term HSC maintenance, we gave S100A8 knockout and control mice four injections of 5-FU (50mg/kg) once per week. The hematopoietic phenotypes of S100A8 knockout and control mice were then assessed at 4, 8, 12, and 16 weeks (Fig. 6K). The results revealed a great increase in white blood cell (WBC) count following the deletion of S100A8 after 4 weeks (Fig. 6L). Surprisingly, the S100A8 knockout mice had lower percentages of mature PB and BM myeloid cells than control mice (Fig. 6M–N), but higher percentages of GMPs (Fig. 6O). However, it is worth noting that the proportion and quantity of LT-HSCs in the S100A8 knockout group were lower than in the control group (Fig. 6P). This finding suggests that S100A8 reduces HSCs and myeloid differentiation bias in a prolonged stressful environment, causing HSCs to exhibit a senescent phenotype.