Resident self-tissue of proinflammatory cytokines rather than their systemic levels correlates with development of myelofibrosis in Gata1low mice

Figure S1. Gating used to identify endothelial cells (ECs), mesenchymal cells (MSCs), osteoblasts (OBCs), megakaryocytes (MKs) and neutrophils (Neu) in the bone marrow from a wild type mouse. In order to correctly estimate the frequency of all the cell populations, the monocellular cell suspension used for the analyses were prepared by carefully crushing the femurs of the mice (see [34] and [35] for detail).

 Figure S2. (A) Example of the computer assisted image analysis used to quantify the level of fibrosis in bone marrow sections stained by Reticulin (black fibers) from representative wild-type (WT) and Gata1^low mice. The ImageJ program-assisted quantification process includes selection of the color channel (black histogram, Hue), determination of the threshold (Saturation histogram) and quantification of the areas which exceed the threshold (in %) (Brightness histogram). (B) Example of the computer assisted image analysis used to quantify the number of megakaryocytes positive per CXCR1 in in bone marrow sections from wild-type (WT) and Gata1^low mice. In this case, the conditions to quantify the signal in wild-type and Gata1^low mice were set separately. The ImageJ program-assisted quantification process includes selection of the cell type, determination of the threshold per cell (Saturation histogram) and quantification of the number of cells which exceed the threshold (Brightness histogram). Magnification 40x.

Figure S3. Levels of the proinflammatory cytokines TGF-β1 (A), LCN2 (B) and CXCL1 (C) detected in the sera from wild-type, Psel^null and Gata1^lowPsel^null mice (all in the CD1 background). The levels of TGF-β1 in Gata1^low mice were previously reported [33] and are presented as median values from 20 mice as horizontal green line for comparison. Values are presented divided by gender (M, males; F, females) and age (adult and old). P values were calculated with Turkey’s multiple comparison test and are summarized either on the right or in Table S1. The number of mice analyzed in each experimental group is indicated by n.

 Figure S4. Serum levels of pro-inflammatory cytokines (TGF-β1, upper panel; LCN2, middle panel; CXCL1, bottom panel) detected in CD1, C57BL/6 and DBA/2 mice. Gender (M, males; F, females) and age (adult and old) were analyzed as independent variables. The number of mice analyzed in each experimental group is indicated by n. Significant p values calculated with multiple comparison performed separately among groups from each different genotype are indicated in the graph. P values obtained with Turkey’s multiple comparison test among all groups are summarized in supplementary Tables S2-S4.

 Figure S5. Hematoxylin/Eosin (H&E) and reticulin staining and TGF-β1, LCN2 and CXCL1 immunostaining of bone marrow sections from representative wild-type (WT) and Gata1^low littermates showing the fibrosis and increased cytokine bioavailability in the microenvironment of the mutant mice. Results are representative of those observed in 12-month old mice randomly selected from male and female, at least were analyzed 6 mice each staining. The arrows indicate individual megakaryocytes in wild-type mice and megakaryocyte clusters in Gata1^low mice. Magnification 40x. Similar results were previously published in Dunbar et al, 2021[11].

Figure S6. Immunohistochemical staining with (P+S) and without (S) the primary antibodies of consecutive sections from the bone marrow of WT and Gata1^low mice showing the specificity of the staining with the primary antibodies. 

Figure S7. The bone marrow from Gata1^low mice contain great levels of CXCR1. High levels of CXCR1 and CXCR2 in bone marrow sections from Gata1^low mice. A) Representative sections from the bone marrow of wild-type and Gata1^low littermates immunostained with antibodies per CXCR1 and CXCR2. Magnification 40x. B) Computer assisted quantification of the content of the CXCR1 and CXCR2 in 5 randomly selected areas per femur from three-four mice per experimental group. Statistical analysis was done by t-test and statistically significant p values among groups are indicated within the panels. Similar results were reported in Dunbar et al[11].

Figure S8. (A) Median optical slice of a confocal stack microscopy obtained with a 63x objective. The blue arrow indicates a GFP signal present inside nucleus (shown as tridimensional image in movie 1), whereas the red arrow indicates a cell were the GFP signal is localized in the cytoplasm (shown as tridimensional image in movie 2). DAPI and GFP signals are shown separately and combined. B) Hematoxylin-eosin staining (left) and fluorescent microscopy (right) of two representative sections from the bone marrow of one huCD34-GFPH2B mouse. The GFP staining labels either the nuclear areas of cells with small size (1, 2 and 3) or, very rarely, the cytoplasm of large cells (4), possibly macrophages. Slides were counter stained with DAPI. Magnification 40x.