The Variation in the Traits Ameliorated by Inhibitors of JAK1/2, TGF-β, P-Selectin, and CXCR1/CXCR2 in the Gata1low Model Suggests That Myelofibrosis Should Be Treated by These Drugs in Combination

Abstract

Studies conducted on animal models have identified several therapeutic targets for myelofibrosis, the most severe of the myeloproliferative neoplasms. Unfortunately, many of the drugs which were effective in pre-clinical settings had modest efficacy when tested in the clinic. This discrepancy suggests that treatment for this disease requires combination therapies. To rationalize possible combinations, the efficacy in the Gata1^low model of drugs currently used for these patients (the JAK1/2 inhibitor Ruxolitinib) was compared with that of drugs targeting other abnormalities, such as p27kip1 (Aplidin), TGF-β (SB431542, inhibiting ALK5 downstream to transforming growth factor beta (TGF-β) signaling and TGF-β trap AVID200), P-selectin (RB40.34), and CXCL1 (Reparixin, inhibiting the CXCL1 receptors CXCR1/2). The comparison was carried out by expressing the endpoints, which had either already been published or had been retrospectively obtained for this study, as the fold change of the values in the corresponding vehicles. In this model, only Ruxolitinib was found to decrease spleen size, only Aplidin and SB431542/AVID200 increased platelet counts, and with the exception of AVID200, all the inhibitors reduced fibrosis and microvessel density. The greatest effects were exerted by Reparixin, which also reduced TGF-β content. None of the drugs reduced osteopetrosis. These results suggest that future therapies for myelofibrosis should consider combining JAK1/2 inhibitors with drugs targeting hematopoietic stem cells (p27Kip1) or the pro-inflammatory milieu (TGF-β or CXCL1).

Keywords: Gata1; IL-8; P-selectin; TGF-β; megakaryocytes; myelofibrosis.

Figure 1

Effects of the various treatments on the total number of cells in the femur. (A) Number of cells in the femur in the vehicle groups of all the experiments. Bone marrow cells were recovered either by flushing the femoral cavity or by carefully crushing the whole femur, as indicated. The two sets of values underwent separate statistical analyses. (B) Fold changes induced by the various treatments (each color indicates a different treatment) expressed as means (±SD) and as values in individual mice (each mouse corresponds to one symbol). The color code used to indicate the treatments is summarized by the squares on the right and is used consistently in all the figures. Statistical analyses of the fold changes among groups were carried out using One-way ANOVA; significant differences are indicated by asterisks (* = p < 0.05, ** = p < 0.01, *** = p < 0.001, and **** = p < 0.0001). The number below the X-axis indicates the day of the treatment in which the mice were sacrificed. The original data were published in [17,20,21,23,24,26].

Figure 2

All the treatments effectively reduced fibrosis in the bone marrow of Gata1^low mice; the drugs which sustained the greatest reductions were Reparixin (20 days) and SB431542 (54 days). (A–D) Representative images of the H&E and reticulin staining of bone marrow from mice treated with either vehicle or SB431542 (A); AVID200 for 42 days (B); RB40.34 and Ruxolitinib, alone or in combination, for 54 days (C); or Reparixin for 20 and 37 days (D) as described. Magnification: 20×. (E) Absolute values of bone marrow fibrosis (% of area) of the vehicles in all the experiments and (F) changes in bone marrow fibrosis induced by the various treatments are presented as means (±SD) and as values in individual mice (each mouse corresponds to one symbol). The fold changes in the groups were not statistically different when using one-way ANOVA (* = p < 0.05, ** = p < 0.01, and *** = p < 0.001). The original data were published in [17,20,21,23,24,26].

Figure 3

With the exception of AVID200, all the treatments effectively reduced the vessel density of the bone marrow of the Gata1^low mice. (A) Representative immunohistochemical images of the femoral anti-CD34 antibody from the Gata1^low mice treated with Reparixin. The vessels are indicated by red asterisks. Magnification: 40×. (B) Vessel density in the bone marrow of the mice treated with vehicle or Reparixin for 20 and 37 days, as indicated. The number of vessels is the average of those measured in 5 randomly selected photomicrographs per bone marrow section per mouse (area of each photomicrograph = 1.49 mm²). The results are presented as means (±SD) and as values in individual mice (each mouse corresponds to one symbol); the statistical analyses were carried out by using one-way ANOVA (* = p < 0.05). (C) Vessel density in the bone marrow of mice treated with vehicle in all the experiments and (D) changes in vessel density induced by the various treatments expressed as Delta of the corresponding vehicle values. Results are presented as means (±SD) and as values in individual mice (each mouse corresponds to one symbol). No statistical differences were found in the Delta induced by the various treatments by one-way ANOVA. The original data were published in [17,21,23,24,26].

Figure 4

The treatments tested had limited effects on the areas of the femur occupied by bone in Gata1^low mice. (A) Reconstruction of a femur from a representative Gata1^low mouse treated with vehicle and stained with H&E and depiction of the computer-assisted process used to determine the area of the femur occupied by the medulla and that occupied by bone, respectively. (B) Percentages of area with bone tissue in the femur of wild-type and Gata1^low mice (males, 12 months old). Statistical analyses were carried out by using one-way ANOVA (* = p < 0.05). (C) Percentages of area with bone tissue in the femur from the vehicle group in all the experiments and (D) changes in the area occupied by bone in femurs from mice treated with the various drugs expressed as Delta of the values in the corresponding vehicle group. The results are presented as means (±SD) and as values in individual mice (each mouse corresponds to one symbol). No statistical differences were found among the groups by using one-way ANOVA.

Figure 5

Only AVID200 and Reparixin improved the percentage of mature bone present in the femurs of Gata1^low mice. (A) Representative images of the Mallory trichrome staining of sections of the femurs of wild-type (WT) and Gata1^low mice treated either with vehicle or with SB431542 (54 days), AVID200 (42 days), and Reparixin (20 and 37 days), as indicated. Magnification: 20×. It should be noted that the effects appear more pronounced in the cortical than in the trabecular areas. (B) The quantification of mature bone present in the cortex and trabeculae of the femurs of the WT and Gata1^low mice treated either with the vehicle or with SB431542 (54 days), AVID200 (42 days), or Reparixin (20 and 37 days). Data are represented as means (±SD) and as values in individual mice (each symbol corresponds to a different mouse). The statistical analysis was carried out by using one-way ANOVA (* = p < 0.05, ** = p < 0.01, and *** = p < 0.001). (C) The areas of mature bone present in the cortex and in the trabeculae of the femurs from the vehicle group in all the experiments. (D) Fold changes (such as Delta compared with the corresponding vehicle) in areas of mature bone present in the cortex and in the trabeculae of the femurs from Gata1^low mice treated with the various drugs, as indicated. The results are presented as means (±SD) and as values in individual mice (each mouse corresponds to one symbol). There were no statistically significant differences among the effects exerted by the various treatments by using one-way ANOVA.

Figure 6

The greatest reductions in spleen weight were obtained by treatment with SB431542 and Ruxolitinib, either alone or in combination with RB40.34. (A) Spleen weight in the vehicle of each experiment. (B) Changes induced by the various treatments in spleen weight compared with the corresponding vehicle presented as means (±SD) and as values in individual mice (each mouse corresponds to one symbol). Statistical analyses were performed by using one-way ANOVA (* = p < 0.05, ** = p < 0.01, *** = p < 0.001, and **** = p < 0.0001). The original data were published in [17,20,21,23,24,26].

Figure 7

All the treatments tested induced modest changes in the total number of cells present in the spleen. (A) Spleen cellularity in the vehicle groups of all the experiments. Only in the case of SB431542 was the cellularity of the spleen from the vehicle group significantly greater than that of the other groups. (B) Changes induced by the various treatments in the spleen cellularity compared with the corresponding vehicle. Data are presented as means (±SD) and as values in individual mice (each mouse corresponds to one symbol). The effects induced by the different drugs were not statistically different according to one-way ANOVA (**** = p < 0.0001). The original data were published in [17,21,23,24,26].

Figure 8

All the treatments reduced fibrosis in the spleen of Gata1^low mice; the greatest reductions were achieved with Ruxolitinib alone or in combination with RB40.34. (A–C) Representative unpublished images of the H&E and reticulin staining of spleen sections from mice treated with either the vehicle or SB431542 (54 days) (A), RB40.34 or Ruxolitinib alone and in combination (54 days) (B), or Reparixin (20 and 37 days) (C) as described. Magnification: 40×. (D) Level of fibrosis in the spleen in vehicle group of the various experiments and (E) fold change in the level of fibrosis in the spleen of the treated group compared with that of the corresponding vehicle group. Data are presented as means (±SD) and as values in individual mice (each mouse corresponds to one symbol). Statistical analysis was carried out by using one-way ANOVA; significant differences are indicated by asterisks (* = p < 0.05, ** = p < 0.01, *** = p < 0.001, and **** = p < 0.0001). The original data were published in [17,23,24,26].

Figure 9

The majority of the treatments tested reduced TGF-β1 content in bone marrow from Gata1^low mice. (A–D) Representative images of the immunohistochemistry analyses with the antibody against TGF-β1 of bone marrow sections from mice treated with either the vehicle or SB431542 for 54 days (A), AVID200 for 42 days (B), RB40.34 or Ruxolitinib alone and in combination for 54 days (C), or Reparixin for 20 and 37 days (D), as described. Magnification: 40×. (E) Levels of TGF-β1 in bone marrow of the vehicle group in each experiment. Data are represented as means (±SD) of the percentage of the area positive for staining and as values in individual mice (each mouse corresponds to one symbol) and (F) fold changes in the levels of TGF-β1 with respect to the corresponding vehicle groups (each mouse corresponds to one symbol). The statistical analysis was carried out by using one-way ANOVA; the statistically significant differences are indicated by asterisks (* = p < 0.05, ** = p < 0.01, *** = p < 0.001, and **** = p < 0.0001). The original data were published in [17,21,23,24,26].

Figure 10

Only RB40.34 in combination with Ruxolitinib reduced CXCL1 content in bone marrow from Gata1^low mice. (A–E) Representative images of bone marrow sections from mice treated with either the vehicle or SB431542 for 54 days (A), AVID200 for 42 d (B), RB40.34 or Ruxolitinib or their combination for 54 days (C), or Reparixin for 20 and 37 days (D), stained with the antibody against CXCL1, as described. Magnification: 40×. (E) Levels of CXCL1 in the bone marrow of the vehicle group in each experiment, represented as means (±SD) of the percentage of the area positive for the staining and as values in individual mice (each mouse corresponds to one symbol) and (F) CXCL-1 content in the treated animals presented as fold change with respect to the corresponding vehicle group. The results are presented as means (±SD) and as values in individual mice (each mouse corresponds to one symbol). Statistical analysis among the groups was carried out by using one-way ANOVA; statistically significant differences are indicated by asterisks (* = p < 0.05, ** = p < 0.01, and *** = p < 0.001). Similar data were published in [17,20,21,23,24,26].

Figure 11

Aplidin and inhibitors of the TGF-β signal (SB431542 and AVID200) increased the platelet counts in Gata1^low mice. (A) Platelet counts in the vehicle in each experiment (each mouse corresponds to one symbol). It should be noted that in spite of being statistically different, the platelet numbers observed in the different vehicles were, for the most part, low. (B) The platelet counts were expressed as Delta of the values in the respective vehicle groups and are presented as means (±SD) and as values per individual mouse (each mouse corresponds to one symbol). Statistical analysis of the fold changes among the groups was carried out by using one-way ANOVA; statistically significant differences are indicated by asterisks (* = p < 0.05, ** = p < 0.01, *** = p < 0.001, and **** = p < 0.0001). The original data were published in [17,20,21,23,24,26].

Figure 12

Treatment with the TGF-β inhibitors, Ruxolitinib, or Reparixin increased GATA1 content in the MKs in bone marrow from Gata1^low mice; however, only in the case of the TGF-β inhibitor SB431542, the increase in the GATA1 content was associated with a mature morphology. (A) Representative immunofluorescence analyses with the megakaryocyte marker CD42b and an antibody against GATA1 of bone marrow sections from the femur of a wild-type mouse and from mice treated with either the vehicle or the TGF-β inhibitor AVID200 (42d) or SB431542 (54d), as controls. Magnification: 40× scale bars. (B) Frequency of CD42b-positive cells in the vehicle groups from all the experiments. Data are presented as means (±SD) and as values in individual mice and are the average number of CD42-positve cells detected in 5 randomly selected photomicrographs (1.49 mm²) per bone marrow section per mouse. No statistically significant difference was detected by using one-way ANOVA. (C) Frequency of cD42b-positive cells in bone marrow from mice subjected to the various treatments expressed as Delta values of those in the corresponding vehicle groups. (D) Frequency of CD42b-positive cells positive for GATA1 in the vehicles and treated mice in new experiments with AVID200 and SB431542. Statistical analyses were carried out by using one-way ANOVA (* = p < 0.05). (E) Frequency of CD42b-positive cells positive for GATA1 in the vehicle groups from all the experiments. The results are presented as means (±SD) and as values in individual mice. (F) Frequency of CD42b-positive cells positive for GATA1 in mice undergoing the various treatments expressed as Delta values of that of the corresponding vehicle group. The results are presented as means (±SD) and as values in individual mice (each mouse corresponds to one symbol). Statistical analysis of the fold changes across the groups were carried out by using one-way ANOVA; the statistically significant groups are indicated by asterisks (* = p < 0.05, ** = p < 0.01 and *** = p < 0.001). (G) Frequency of MKs positive for GATA1 containing monolobated or polylobated nuclei among the different experimental groups. Some of the original data were published in [17,21,23,24,26].

Figure 13

Scheme of all treatments investigated in this study. The timeline of each experiment with the age of the mice at the beginning of the experiment and on the day of their sacrifice for histopathological evaluation is indicated. The total number of days in which each drug was administered and the entire duration of each treatment are also indicated. The continuous lines indicate the days when the drugs were administered, while the dotted lines represent the pause between the sequential cycles of drug administration. The original data were published in [17,20,21,23,24,26].