Novel Evidence That Hematopoietic Stem/Progenitor Cells (HSPCs) Are Mobilized During Hemolysis in an Erythrocyte Lysis-Derived, Sphingosine-1-Phosphate (S1P)-Dependent manner—the Crucial Involvement of Complement Cascade (CC) Activation and Attenuation of CXCR4 Retention Signaling

Hemolytic syndromes, such as sickle cell anemia and paroxysmal nocturnal hemoglobinuria, are characterized by an increased number of hematopoietic stem/progenitor cells (HSPCs) circulating in peripheral blood (PB). However, the molecular mechanisms responsible for this effect are unclear. In our previous work we have demonstrated that sphingosine-1-phosphate (S1P) released from lysed erythrocytes and activated platelets is a strong chemottractant for bone marrow (BM)-residing HSPCs (Leukemia 2010;24:976–85).

We hypothesized that S1P released from lysed erythrocytes is a major factor responsible for egress of HSPCs from BM into PB in hemolytic syndromes.

To test this hypothesis, normal mice were injected with phenylhydrazine (PHZ), a compound known to induce hemolysis, and we evaluated the number of Sca-1⁺Kit⁺Lin^– (SKL) HSPCs circulating in PB as well as the number of clonogenic CFU-GM progenitors mobilized into PB. In parallel, we evaluated the blood plasma levels of S1P and stromal derived factor-1 (SDF-1) by sensitive ELISA, the free hemoglobin (Hb) level, as well as complement cascade (CC) activation by measuring the C5b-C9 (membrane attack complex, MAC) level. In some of the experiments, we combined PZH treatment with injection of the CXCR4 antagonist AMD3100. To better assess the role of CC activation, we also performed mobilization in C5-deficient mice, which do not activate the distal part of the CC and thus do not generate C5b-C9/MAC.

We found that hemolysis increases the PB plasma level of S1P but does not affect the SDF-1 level. Furthermore, while PHZ-induced hemolysis mobilizes HSPCs into PB with a peak at 6 h after infusion, this mobilization effect is significantly potentiated by administration of AMD3100, which attenuates CXCR4–SDF-1-mediated retention of HSPCs in the BM microenvironment. Of note, PHZ-induced hemolysis together with AMD3100 mobilized twice as many HSPCs as AMD3100 alone. The degree of mobilization of HSPCs correlated with the free Hb level in plasma and activation of the CC (by an increase in MAC level), and, more importantly, mobilization was not seen in C5-deficient mice.

We confirmed our previous observation that with the steady-state S1P level, the PB has already established a strong chemotactic gradient for BM-residing HSPCs (Leukemia 2010;24:976–85), which are actively retained in BM niches in a CXCR4–SDF-1-dependent manner. Hemolysis alone, even if it elevates the S1P level in PB significantly, requires two important events i) attenuation of CXCR4–SDF-1 axis-mediated retention in BM niches and ii) simultaneous activation of the CC, which is crucial for induction of permeabilization of the BM–PB barrier. Furthermore, our data also support the notion that the S1P but not the SDF-1 level in PB establishes the critical chemotactic gradient for HSPCs and is responsible for egress of these cells. Based on the observation that C5-deficient mice are poor mobilizers, inactivation of the distal part of the CC should be considered as a therapeutic approach, not only in paroxysmal nocturnal hemoglobinuria but also in other hemolytic syndromes.

No relevant conflicts of interest to declare.