Inhibition of CXCR1/2 reduces the emperipolesis between neutrophils and megakaryocytes in the Gata1low model of myelofibrosis

Abstract

Emperipolesis between neutrophils and megakaryocytes was first identified by transmission electron microscopy. Although rare under steady-state conditions, its frequency greatly increases in myelofibrosis, the most severe of myeloproliferative neoplasms, in which it is believed to contribute to increasing the transforming growth factor (TGF)-β microenvironmental bioavailability responsible for fibrosis. To date, the challenge of performing studies by transmission electron microscopy has hampered the study of factors that drive the pathological emperipolesis observed in myelofibrosis. We established a user-friendly confocal microscopy method that detects emperipolesis by staining with CD42b, specifically expressed on megakaryocytes, coupled with antibodies that recognize the neutrophils (Ly6b or neutrophil elastase antibody). With such an approach, we first confirmed that the bone marrow from patients with myelofibrosis and from Gata1^low mice, a model of myelofibrosis, contains great numbers of neutrophils and megakaryocytes in emperipolesis. Both in patients and Gata1^low mice, the emperipolesed megakaryocytes were surrounded by high numbers of neutrophils, suggesting that neutrophil chemotaxis precedes the actual emperipolesis event. Because neutrophil chemotaxis is driven by CXCL1, the murine equivalent of human interleukin 8 that is expressed at high levels by malignant megakaryocytes, we tested the hypothesis that neutrophil/megakaryocyte emperipolesis could be reduced by reparixin, an inhibitor of CXCR1/CXCR2. Indeed, the treatment greatly reduced both neutrophil chemotaxis and their emperipolesis with the megakaryocytes in treated mice. Because treatment with reparixin was previously reported to reduce both TGF-β content and marrow fibrosis, these results identify neutrophil/megakaryocyte emperipolesis as the cellular interaction that links interleukin 8 to TGF-β abnormalities in the pathobiology of marrow fibrosis.

Figure 1

Validation of confocal microscopy observations with antibodies specific for the megakaryocytes and the neutrophils as a method to detect emperipolesis in the bone marrow. (A) A representative image acquired by confocal microscopy of bone marrow sections from a Gata1^low mouse stained with CD42b (red, megakaryocyte) and Ly6b (green, neutrophil). Nuclei are counterstained with Hoechst (blue). The asterisks indicate two representative megakaryocytes (in red), one of which contains Ly6b-positive cells (blue, indicated by the arrow). This image is from an area of the bone marrow parenchyma close to a vessel, as indicated by the presence of red blood cells (the autofluorescent green Hoechst-negative cells). (B) Three-dimensional reconstruction of the emperipolesed megakaryocyte shown in Figure 1A showing that the Ly6b signal is inside the cytoplasm of the CD42b-positive cells (Supplementary Video E1). Original magnification, 60 ×.

Figure 2

The bone marrow from patients with MF contains significantly greater numbers of megakaryocytes (MKs) (CD42b-positive [CD42b^pos] cells, red) and neutrophils (neutrophil elastase−positive [Neu Elast.^pos] cells, green), which are located around or emperipolesed with MKs, than the bone marrow from patients without MF or with ET. (A) Confocal microscopy analyses of bone marrow sections from one patient without MF, one patient with ET, and one patient with MF stained with CD42b (red), neutrophil elastase (green), and Hoechst (blue). The panels on the top and bottom present representative images showing the frequency of neutrophils around (as a surrogate marker of chemotaxis) and inside (as evidence of emperipolesis) the MKs, respectively. Original magnification, 60 × . (B) Frequency of MKs (CD42b-positive cells) and neutrophils (neutrophil elastase−positive cells) in bone marrow sections from three patients without MF, three patients with ET, and three patients with MF. (C) Frequency of neutrophils (neutrophil elastase−positive cells) surrounding or emperipolesed with MKs (CD42b-positive cells) of the different groups. (D) The number of nuclear lobules (as detected by Hoechst staining) contained in the cytoplasm of MKs emperipolesed by neutrophils present in bone marrow sections from the patients. MKs were divided into cells containing 1−2, 3, or >4 lobules. Data are presented as mean (±SD) and as values for individual patients (each dot represents a different patient). In Figure 2B−D, data are presented as mean (±SD) and as values per individual patient (each dot represents a different patient). p values were calculated by Tukey’s multiple comparisons test.

Figure 3

The bone marrow from Gata1^low mice contains significant greater numbers of megakaryocytes (MKs) (CD42b-positive [CD42b^pos] cells, red) and neutrophils (Ly6b-positive [Ly6b^pos] cells, green), which are located around or emperipolesed with MKs, than the bone marrow from WT littermates. (A) Confocal microscopy analyses of bone marrow sections from one Gata1^low mouse (8 months old) and one age-matched WT littermate, as indicated, stained with CD42b (red), Ly6b (green), and Hoechst (blue). The panels for Gata1^low mice show a group of MKs surrounded by neutrophils (as a surrogate marker of chemotaxis) and one MK in emperipolesis with neutrophils (asterisk in the right panel). Original magnification, 60 × . (B) Frequency of MKs (CD42b-positive cells) and of neutrophils (Ly6b-positive cells) in bone marrow sections from three WT and three Gata1^low mice. (C) Frequency of neutrophils (Ly6b-positive cells) surrounding or emperipolesed with MKs (CD42b-positive cells) in the two groups. (D) The number of nuclear lobules (as detected by Hoechst staining) contained in the cytoplasm of MKs in emperipolesis, with the neutrophils present in bone marrow sections from the two groups. MKs were divided into cells containing 1−2, 3, or >4 lobules. Data are presented as mean (±SD) and as values for individual patients (each dot represents a different patient). In Figure 3B−D, data are presented as mean (±SD) and as values per individual patient (each dot represents a different patient). p values were calculated by one-way ANOVA in Figure 3B and C and by Tukey’s multiple comparisons test in Figure 3D.

Figure 4

Treatments with reparixin do not affect the frequency of megakaryocytes (MKs) present in the bone marrow but significantly decreases that of neutrophils (Ly6b-positive [Ly6b^pos] cells, green). It also decreases the frequency of neutrophils around or emperipolesed with MKs in the bone marrow of Gata1^low mice. (A) Confocal microscopy analyses of bone marrow sections from one Gata1^low mouse (9 months old) treated either with a vehicle or with reparixin, as indicated, stained with CD42b (red), Ly6b (green), and Hoechst (blue). The rectangle in the panels on the left are shown at greater magnification on the right. Representative MKs and neutrophils are indicated by the red and green arrows, respectively. Original magnification, 60 × . (B) Frequency of MKs (CD42b-positive [CD42b^pos] cells) and neutrophils (Ly6b-positive cells) in bone marrow sections from three WT and three Gata1^low mice. (C) Frequency of neutrophils (Ly6b-positive cells) surrounding or emperipolesed with MKs (CD42b-positive cells) in the two groups. (D) The number of nuclear lobules (as detected by Hoechst staining) contained in the cytoplasm of MKs in emperipolesis with the neutrophils present in bone marrow sections from the two groups. MKs were divided into cells containing 1−2, 3, or >4 lobules. Data are presented as mean (±SD) and as values for individual patients (each dot represents a different patient). In Figure 4B−D, data are presented as mean (±SD) and as values per individual patient (each dot represents a different patient). p values were calculated by one-way ANOVA in Figure 4B and C and by Tukey’s multiple comparisons test in Figure 4D.