Bone marrow effects of anagrelide therapy in patients with myelofibrosis with myeloid metaplasia


Dr Chin-Yang Li, Division of Hematopathology, Mayo Clinic, 200 First Street SW, Rochester, MN 55905, U.S.A.


In a prospective study investigating the therapeutic role of anagrelide in myelofibrosis with myeloid metaplasia, 20 patients received anagrelide in daily oral doses of 0.5–3 mg. 17 patients were evaluable and received anagrelide for a median of 2 years (range 0.5–4 years). No patient had a clinically appreciable benefit. Bone marrow (BM) examinations at baseline and after 6 and 12 months of treatment were available for 17, 17 and 12 patients, respectively. In all evaluable cases, BM megakaryocyte number increased after 6 months of anagrelide treatment. Also, Ulex europaeus agglutinin-1 staining of megakaryocytes revealed a left-shifted maturation pattern in most patients with a platelet response to anagrelide. However, megakaryocyte staining intensity for transforming (TGF-β) and platelet-derived (PDGF) growth factors was not affected consistently by treatment. No patient had a geqslant R: gt-or-equal, slanted2 grade change in either BM fibrosis or osteosclerosis. These in-vivo data support our previous in-vitro observations that anagrelide interferes with megakaryocyte maturation rather than proliferation. Lack of a positive treatment effect is consistent with the finding that anagrelide did not significantly alter megakaryocyte expression of TGF-β and PDGF.

Myelofibrosis with myeloid metaplasia (MMM) is a clonal stem cell disorder with an associated reactive bone marrow fibrosis (BMF) (Tefferi et al, 1995). MMM may present de novo as chronic idiopathic myelofibrosis or develop during the latter stages of polycythaemia (post-polycythaemic myelofibrosis) or essential thrombocythaemia (post- thrombocythaemic myelofibrosis). Various cytokines released from clonal megakaryocytes, primarily transforming growth factor (TGF-β) and platelet-derived growth factor (PDGF), have been implicated in the pathogenesis of the BMF in MMM (Kimura et al, 1988; Martyre et al, 1991; Martyre, 1995; Reilly, 1997; Yanagida et al, 1997). Both of these growth factors exist in different isoforms: β1, β2 and β3 isoforms of TGF-β (Ohta et al, 1987; Chuang & Li, 1996) and AA, AB and BB isoforms of PDGF (Hammacher et al, 1988; Siegbahn et al, 1990; Osornio-Vargas et al, 1995). We have demonstrated with immunohistochemical methods the strong staining of TGF-β3 (Chuang & Li, 1996) and PDGF-BB (unpublished observations) in megakaryocytes.

Anagrelide is an oral agent that has a specific platelet-lowering effect in humans (Tefferi et al, 1997). The drug is effective in controlling thrombocytosis associated with various clonal haematological diseases. Currently, it is approved in the United States for the treatment of essential thrombocythaemia. In vitro, therapeutic concentrations of anagrelide have been shown to alter the maturation pattern, size and ploidy of developing megakaryocytes without inhibiting proliferation (Mazur et al, 1992; Solberg et al, 1997). Therefore it is reasonable to consider that anagrelide may have a therapeutic benefit in MMM by affecting megakaryopoiesis and the subsequent release of fibrogenic cytokines. Thus, we prospectively studied 20 patients with MMM being treated with anagrelide. Herein, we describe the in-vivo bone marrow changes observed in these patients. In particular, we evaluated (1) megakaryocyte number, morphology and maturation pattern; (2) bone marrow fibrosis and osteosclerosis; and (3) megakaryocyte growth factor expression.



The purpose of the current study was to investigate the effect of anagrelide on both clinical and laboratory variables in patients with MMM (chronic idiopathic myelofibrosis, post-polycythaemic myelofibrosis, post-thrombocythaemic myelofibrosis). After approval by the Institutional Review Board of the Mayo Foundation, 20 consecutive patients were accrued to the study. The morphologic diagnosis in each patient was independently re-reviewed by two of the authors without knowledge of treatment effect on the clinical status of the patients. In addition, eligibility criteria included a platelet count > 100 × 109/l and the absence of concurrent therapy with other treatment agents used in MMM. All patients started receiving treatment with oral anagrelide at 1.0 mg daily. The dose of anagrelide was either increased or decreased to achieve a target platelet count between 100 and 300 × 109/l. In some patients the occurrence of side-effects influenced drug dose regardless of platelet count. The two major clinical objectives were to evaluate drug effect on hepatosplenomegaly and anaemia. The laboratory objectives (elaborated below) were to evaluate drug effect on bone marrow fibrosis, osteosclerosis, megakaryocyte proliferation and maturation, and megakaryocyte growth factor content. According to the protocol requirement, patients were removed from the study (i.e. anagrelide therapy was stopped) if progressive organomegaly, anaemia, or bone marrow fibrosis was demonstrated. Treatment responses were monitored closely and included documentation of changes in anaemia and organomegaly.

Laboratory studies

Haematoxylin–eosin stains of bone marrow biopsy samples were used to evaluate megakaryocyte number and morphology and the extent of osteosclerosis. A reticulin stain was used to assess the degree of BMF. Osteosclerosis and fibrosis were graded according to the criteria stated in 1Table I. Semiquantitative immunohistochemical methods were used to measure megakaryocyte expression of TGF-β3 and PDGF-BB. Immunohistochemical staining was performed with the horseradish peroxidase (HRP)-labelled streptavidin-biotin method using the antibodies for TGF-β3 (Research Diagnostic, Flanders, N.J., U.S.A.) and PDGF-BB (Genzyme, Cambridge, Mass., U.S.A.) (Siegbahn et al, 1990; Chuang & Li, 1996). The staining intensity of TGF-β3 and PDGF-BB were graded microscopically from 0 to 3+. Growth factor expression was quantified by adding the individual grade points of 100 consecutive megakaryocytes.

Table 1. Table I. Criteria for osteosclerosis and fibrosis grading.Thumbnail image of

In patients with MMM, bone marrow aspirates often are inadequate because of the fibrotic process in the marrow. Therefore, to evaluate the in-vivo effect of anagrelide on megakaryocyte development, we needed a method that allowed analysis of bone marrow biopsy samples. For this we used Ulex europaeus agglutinin-1 (UEA-1 [DAKO, Carpinteria, Calif., U.S.A.]) binding according to the unlabelled peroxidase–anti-peroxidase (PAP) technique (Liu & Li, 1996). UEA-1 is a lectin from gorse seed of Ulex europaeus and selectively binds with vascular endothelial cells and megakaryocytes. The staining pattern of megakaryocytes varies with different stages of cytoplasmic maturation. The UEA-1-related granules are initially localized in the Golgi area (low granular megakaryocyte), and then they spread gradually throughout the cytoplasm (diffuse granular stage). Next, in association with the increased number of fine granules, some larger dense granules appear in the Golgi region and spread through the entire cytoplasm (diffuse dense granular stage), and during the process of platelet budding and release, the granules shift toward the periphery (marginal granular stage).

On the basis of the UEA-1 staining pattern, we categorized megakaryocytes into four maturational stages: low granular (LG), diffuse granular (DG), diffuse dense granular (DDG), and marginal granular (MG). Denuded (DMK) and endomitotic megakaryocytes (EndoM) were counted as a separate category.

Appropriate statistical methods were used to evaluate the effect of treatment on baseline clinical and laboratory measures (Wilcoxon signed rank test) and investigate for significant correlations among different variables (Spearman correlation coefficient).


Twenty patients with MMM were studied (13 with chronic idiopathic myelofibrosis, four with post-polycythaemic myelofibrosis, and three with post-thrombocythaemic myelofibrosis). The age and sex distributions are given in 2Table II. Anagrelide therapy was begun at a median of 10 months after the diagnosis of MMM (range 0–72 months). At the time of treatment the majority of patients had advanced disease; four patients were previously splenectomized, and the pretreatment median haemoglobin and leucocyte count were 11.6 g/dl (range 8.1–16.7) and 7.2 × 109/l (range 3.5–71.3), respectively (Dupriez et al, 1996). Of the 20 patients initially enrolled into the study, three received anagrelide therapy for <3 months because of early death from cortical infarct (one patient) or blastic transformation (two patients). These patients were not included in the analysis. The other 17 patients received anagrelide for a median of 2 years (range 6 months to 4 years) and at a median maintenance dose level of 1.0 mg/d (range 0.5–3.0 mg/d). The target platelet count was not always achieved because of the presence of side-effects with higher doses of anagrelide in some patients. No clinical benefit was demonstrated in any of the patients on the basis of treatment effect on anaemia, transfusion requirements, or organomegaly (data not shown).

Table 2. Table II. Effect of anagrelide treatment on platelet count, osteosclerosis, and fibrosis in 20 patients with myelofibrosis with myeloid metaplasia. CIM, chronic idiopathic myelofibrosis; PPM, post-polycythaemic myelofibrosis; PTM, post-thrombocythaemic myelofibrosis.* Daily maintenance dose of anagrelide.† Month after anagrelide treatment.‡ Cause of stopping follow-up or discontinuation of anagrelide treatment.Thumbnail image of

Pretreatment platelet counts and bone marrow aspirates with biopsy samples were obtained from all patients (Table II). Follow-up bone marrow examinations were performed after 6 and 12 months of therapy in 17 and 12 patients, respectively. 13 patients had a platelet response to treatment, and the platelet count was unchanged in one patient and increased in three. None of the patients had a remarkable change (geqslant R: gt-or-equal, slanted2 grade change) in either BMF or osteosclerosis (Table II).

Megakaryocyte expression of TGF-β3 and PDGF-BB was not consistently affected by anagrelide therapy (Table III), and the correlation of growth factor expression with degree of BMF, megakaryocyte number, or platelet count was not statistically significant (data not shown). UEA-1 staining was performed in nine patients who had adequate marrow cellularity and showed a left-shifted maturation pattern in most patients with a platelet response to anagrelide treatment (Table III) (Figs 1 and 2). Regardless of platelet response to anagrelide, bone marrow megakaryocyte number increased in all 12 patients who had bone marrow cellularity adequate enough to estimate megakaryocyte number (Table IV).

Table 3. Table III. UEA-1 staining pattern and staining intensity of TGF-β3 and PDGF-BB at baseline and after 6 months of anagrelide treatment.* DDG, diffuse dense granular; DG, diffuse granular; DMK, denuded; EndoM, endomitotic megakaryocytes; LG, low granular; MG, marginal granular; MK, megakaryocyte; PDGF, platelet-derived growth factor-BB; TGF-β3, transforming growth factor-β3; UEA-1, Ulex europaeus agglutinin-1 binding.* Unfilled boxes indicate that the study was either not done or not evaluable.† Number of megakaryocytes per low-power field (×160). a = baseline values and b = values after 6 months.‡ The TGF-β and PDGF-BB columns represent the total staining scores of 100 consecutive megakaryocytes, with each megakaryocyte scored from 0 to 3 depending on staining intensity. As shown, baseline scores (a) were compared with scores after 6 months of treatment with anagrelide (b) and showed no significant difference (P > 0.05, Wilcoxon signed rank test). In addition, the Spearman correlation method was used to investigate the relationship of the score for intensity with the degree of fibrosis, platelet count, and megakaryocyte count.Thumbnail image of
Figure 1.

. Left shift of megakaryocyte maturation in platelet-decreased group after treatment with anagrelide. (A) Before anagrelide treatment, the major type of megakaryocyte is dense diffuse granular megakaryocyte. (B) 6 months after treatment megakaryocyte maturation has shifted from the dense diffuse granular stage to the diffuse granular stage. (UEA-1 stain; original magnification, ×400.)

Figure 2.

. Comparison of the marginal granular form megakaryocyte before treatment with anagrelide (A) and 6 months after treatment (B). There is a decrease in the marginal granular form of megakaryocyte after treatment. (UEA-1 stain; original magnification, ×400.)

Table 4. Table IV. Change of megakaryocyte number in bone marrow of patients with myelofibrosis with metaplasia treated with anagrelide.Thumbnail image of


Because of the previously demonstrated in-vitro effects of anagrelide on megakaryocytes (Mazur et al, 1992; Solberg et al, 1997), we thought that patients with MMM could potentially benefit from treatment with this drug. To test this hypothesis, we prospectively studied 17 patients with advanced MMM who received anagrelide for a minimum of 6 months and a maximum of 4 years. After a median follow-up of 2 years none of the patients experienced a clinical benefit as indicated by improvement in either anaemia or organomegaly. Similarly, follow-up examinations of the bone marrow after 6 and 12 months of treatment did not reveal a noticeable effect on the degree of BMF or osteosclerosis. This lack of a positive treatment effect was consistent with the immunohistochemical findings that demonstrated the absence of significant alteration in megakaryocyte expression of TGF-β and PDGF. These observations suggest that although thrombocytosis in MMM may be controlled by anagrelide treatment, the underlying disease process and natural history may not be affected. However, because bone marrow fibrosis is a chronic process, the follow-up period of our patients may not be adequate to exclude a delayed effect.

Our in-vivo observations on the effect of anagrelide on megakaryopoiesis are consistent with previously published in-vitro data (Mazur et al, 1992; Solberg et al, 1997). In-vitro, therapeutic concentrations of anagrelide do not inhibit megakaryocyte proliferation. Instead, we and others (Mazur et al, 1992) have demonstrated a profound effect on the maturation of megakaryocytes, as evidenced by a decrease in cell size, ploidy, surface irregularity, and optical density. This left-shifted maturation effect of anagrelide on megakaryocytes was also demonstrated in the present study by the use of special stains on bone marrow biopsy samples obtained after an adequate period of treatment. Although a maturation arrest explains ineffective platelet production, neither the present study nor the previous in-vitro observations provide insight into the mechanism of anagrelide action on the development of megakaryocytes.

The anagrelide-induced increase in the number of megakaryocytes in the bone marrow seen in our patients with MMM has also been reported in normal volunteers receiving anagrelide (Abe Andes et al, 1984). This observation supports the idea that anagrelide, in therapeutic doses, does not inhibit the proliferation of megakaryocytes. Regardless of the precise mechanism of this paradoxical phenomenon (an increase in number of megakaryocytes in the bone marrow associated with a decrease in the number of circulating platelets), the interference of anagrelide with megakaryocyte maturation and subsequent platelet release may modify a possible negative feed-back mechanism associated with effective platelet production. The latter may involve an effect on the production of thrombopoietin by bone marrow stem cells (Wang & Hashmi, 1998) or an indirect effect on thrombopoietin level through the reduction of growth factor clearance associated with a decrease in platelet count (Wang et al, 1997).

Our understanding of the pathogenesis of MMM continues to evolve. Megakaryocyte proliferation and interaction with the bone marrow microenvironment are integral components in the pathogenesis of MMM. It currently is believed that the interaction between the clonal megakaryocytes and the reactive fibroblasts is mediated by various fibrogenic growth factors, including TGF-β, PDGF, and basic fibroblast growth factor (Reilly, 1997) and may involve specific integrin receptors (Schmitz et al, 1998). The role of thrombopoietin and its receptor (c-mpl) in the pathogenesis of MMM is not clearly defined. Mice exposed to persistently high levels of thrombopoietin develop a clinical syndrome identical to MMM (Yan et al, 1995, 1996; Ulich et al, 1996). The observed BMF may be mediated by increases in growth factors, primarily TGF-β associated with the thrombopoietin-induced expansion of megakaryocyte and platelet mass (Yanagida et al, 1997; Frey et al, 1998). Higher than normal blood levels of thrombopoietin (Wang et al, 1997) and decreased megakaryocyte c-mpl expression have been demonstrated in patients with MMM (Moliterno et al, 1998). However, the pathogenetic relevance of these observations is weakened by the presence of similar findings in patients with essential thrombocythaemia and polycythaemia vera.


  1.  Visiting Clinician at the Division of Hematopathology, from the Department of Clinical Pathology, Korea University, Seoul, Korea.