Distinctive Alterations in the Emperipolesis between Neutrophils and Megakaryocytes from the Bone Marrow of Gata1^low Mice Documented By Time-Lapse Observations

The presence in the bone marrow from patients with myelofibrosis, the most severe of the Ph-negative Myeloproliferative Neoplasms, of naked nuclei of megakaryocytes (MK) supposedly originated by para-apoptosis induced by pathological emperipolesis between these cells and the malignant neutrophils (Neu) has been recognized since 1991-2000^1,2. The discovery that the MK from these patients express an IL-8 (CXCL1 in the mouse) signature³, a potent chemo-attractant for Neu⁴, lead us to discovery that treatment of the Gata1^low mouse model with Reparixin, an inhibitor of CXCR1/CXCR2 (the CXCL1 receptors), rescues both fibrosis and pathological MK-Neu emperipolesis in their bone marrow⁵. However, the mechanistic detail of this interaction remains elusive.

By time-lapse confocal microscopy, Huang et al⁶ have identified that emperipolesis between normal murine Neu/MK occurs in two fashions: It can be fast (<10 min) and slow (>60 min) and that both processes leave the two interacting cells alive.

Using this knowledge as foundation, we thought to generate mechanistic insights on the pathological MK/Neu emperipolesis occurring in myelofibrosis by performing time-lapse observations of co-cultures between Neu and MK (1:10 ratio) purified from bone marrow of CD1 and myelofibrotic Gata1^low mice. Gata1^low MK/Neu were cultured with or without Reparixin (10 µM). The RNAseq profilings of CD41+ cells (>90% pure) from the bone marrow of CD1 and Gata1^low mice were also compared.

Time-lapse were performed with 30 sec time frames during 48 hrs of co-culture. The total numbers of MK analyzed were 45 CD1, 94 Gata1^low and 51 Gata1^low with Reparixin for a total of 147 hrs of observations. The observation time was divided into four groups: 0-12; 13-14, 25-36 and 37-48 hrs. We observed 4 types of interactions which were identified with the code: 0=none; 1= low, touch and go (<90sec); 2=long with possible emperipolesis (>90sec); 3=long (>90sec) with MK de-arrangement and death. The morphologic changes of the MK resulting in death occurred in 15min. Interactions type 0 were observed in the 0-12 time frame with a frequency of 25% for CD1 and only 3% for Gata1^low MK, with or without Reparixin. Type 1 and 2 interactions were observed in the 13-14, 25-36 and 37-48 hrs time frame and their frequency was not statistically different among groups. Of note, type 2 interactions were not observed in Gata1^low co-cultures in the 37-48 hrs time frame when instead 33% of the interactions between Gata1^low MK and Neu were type 3. By contrast, type 3 interactions were seldomly observed in co-cultures with CD1 cells or with Gata1^low cells with Reparixin (Chi-Square p<0.0001 with respect to the Gata1^low without Reparixin for both). The fact that both in the CD1 and Gata1^low group, type 2/3 interactions were observed in great numbers after 12 hours of co-culture suggests that they are induced by chemo-attractants released by the MK. The observation that Reparixin reduced type 3 interactions without affecting the type 2 ones indicates that only type 3 interactions are dependent on CXCL1 and are, therefore, not likely to be emperipolesis. Furthermore, RNAseq analyses identified that Gata1^low MKs have an activated CXCR1/CXCR2 signature with respect to control and are enriched for genes implicated in phagocytosis: upregulation of 16 genes promoting (C3, Camk1d, Cfp, Clec7a, Cyba, Fcgr1, Fpr2, Il1b, Lbp, Ptprj, Rab27a, Rab31, Sirpa, Sirpb1b, Sirpb1c, Slc11a1) and of only 8 genes inhibiting (Atg3, Atg7, Cnn2 ,Prtn3, Pten, Siglece , Sirpa, Tlr2) phagocytosis. Of note, purified Gata1^low MK had also a clear Neu signature which is likely the consequence of contamination by Neu emperipolized insight the cells.

In conclusion, the interactions between MK and Neu observed in the co-cultures of both CD1 and Gata1^low mice are faster than those described by Huang et al⁶ and occur after a time lag of cell inactivity, which suggests that they are dependent on the release of chemo-attractants by the cells. In the Gata1^low co-culture, we observe a new type of MK-Neu interaction that occurs after at least 24 hrs of co-culture, is CXCL1 dependent and leads to cell death that we hypothesize is phagocytosis.

Reference: 1.Thiele J et al. J Submicrosc Cytol Pathol 1991;23:93. 2.Schmitt A et al. Blood 2000;96:1342. 3.Emadi S et al. Blood 2005;105:464. 4.Waugh DJJ and Wilson C. Clin Can Res 2008;14:6735. 5.Arciprete F et al. Exp Hematol 2023;121:30. 6.Huang FY et al. Blood Adv 2022;6:2081.

Allegretti:Dompe: Current Employment. Migliaccio:Dompe: Research Funding.