Main

In 1940, Kasabach and Merritt (1) described for the first time an infant with a vascular anomaly accompanied by thrombocytopenic purpura. This disease, named Kasabach-Merritt syndrome (KMS), is associated with thrombocytopenia, petechiae, spontaneous bleeding, and enlargement of the hemangioma. Based on reviewing the world literature in 1988, the mortality rate is estimated to be 12% or higher as a result of severe bleeding, sepsis, or invasion in vital organs (2, 3). In comparison with normal platelet counts of ±250,000–500,000/μL in children (0–5 yr), children with KMS typically have counts of 25,000/μL (range 3,000–60,000) with a consumptive coagulopathy-like state (3). The vascular tumors associated with KMS are large, red to purple, warm, and edematous. These tumors can be characterized by an infiltrative growth pattern, spindling of endothelium, microthrombi, and hemosiderin (3). It has also been demonstrated that platelets trap in these vascular anomalies (4, 5). Thus far, no universally successful treatment has been found, although pharmacologic therapy (including steroids, antiplatelet-aggregating agents, and antifibrinolytic agents), radiotherapy, and surgery have been tried (reviewed in Refs. 3 and 6). Platelet transfusions have been given in life-threatening situations, although exacerbations of s.c. bleeding and swelling of the tumors have been reported presumably due to intratumoral trapping and activation of platelets (6, 7). If possible, such transfusions should be avoided.

In 1994, thrombopoietin (Tpo), a growth factor for megakaryocytes, was discovered to be a ligand for the receptor encoded by the c-mpl proto-oncogene (c-Mpl) (8). The c-Mpl receptor is a member of the hematopoietin receptor super family and is involved in megakaryocytopoiesis (9). In addition to Tpo, a truncated form of Tpo that contains the Mpl receptor binding domain was cloned as the megakaryocyte and development factor (MGDF) (10). In clinical trials, both compounds have been shown to elevate platelet counts and hasten platelet recovery in patients (11, 12). For patients with the Kasabach-Merritt syndrome, we hypothesized that stimulation of platelet production may be helpful to prevent life-threatening situations. Stimulation of endogenous platelet production may avoid the adverse side effects of platelet transfusions because there is no partial activation due to storage (13). To test this hypothesis, we stimulated platelet production by Peg-rHuMGDF in a mouse model of KMS (14). In this murine model, mice die from hematologic changes similar to those found in the KMS in human beings including platelet aggregates, areas of hemorrhage in the tumors, and a reduced red cell survival (14).

METHODS

Peg-rHuMGDF was kindly provided by Amgen Inc. (Thousand Oaks, CA) after the institutional board obtained approval for this study. 10 μg/kg of Peg-rHuMGDF, diluted in 0.1% homologous mouse serum diluted in saline, was administered daily by intraperitoneal injection. Treatment was started when mean tumor volume was at least 150 mm3. Controls were treated with daily intraperitoneal injections of saline.

Cell culture.

The murine hemangioendothelioma cell line EOMA and bovine capillary endothelial cells (BCEs) were maintained in Dulbecco's modified Eagle's medium supplemented with 10% bovine calf serum and 1% glutamine-penicillin-streptomycin in 37°C in a 10% CO2 incubator. BCEs grew on precoated plates with 1% gelatin.

In vivo model of the Kasabach-Merritt syndrome.

Six-wk-old male 129J mice, the favorable mice for growing EOMA tumors, were obtained from Jackson Laboratory (Bar Harbor, ME). They were housed in the animal research facilities of Children's Hospital. After trypsinizing and one PBS wash, 1 × 107 Eoma cells were diluted in 1 mL of PBS. Subsequently, 1 × 106 cells (=100 μL) were injected s.c. under sterile conditions in the proximal midline of the back of a mouse. Length and width of tumors were measured to calculate tumor volume with the formula length × (width)2 × 0.52. Mice were weighed once a week. Animal care was reviewed by the animal care committee and maintained in accordance with constitutional guidelines.

Platelet counts.

By tail bleeding, 20-μL samples of blood were sampled in unopette capillary pipettes and diluted in reservoirs of unopette microcollection systems (Becton Dickinson, Franklin Lakes, NJ). Within 1 h, platelets were counted under phase contrast microscopy in a microchamber.

Immunohistochemistry.

Tumor tissues were fixed in Carnoy's fixative and embedded in paraffin according to standard histologic procedures. Tissue sections (4–6 μM) were pretreated with 2 μg/mL proteinase K (Boehringer Mannheim, Mannheim, Germany) at 37°C for 15 min. Staining was performed with a rabbit polyclonal antibody against human von Willebrand factor (DAKO, Carpinteria, CA). Subsequently, a secondary antibody against rabbit conjugated to horseradish peroxidase (DAKO) was used, and finally diaminobenzidine tetrahydrochloride (DAKO) was used as a chromagen. After counter staining with methyl green (Schmid & Co., Stuttgart, Germany), they were mounted in Permount (Fisher Scientific). H&E staining was performed according to standard procedures.

Proliferation assay.

Eoma or BCE cells (2 × 103) in 100 μL were plated in 96-well plates. The next day, Peg-rHuMGDF in a concentration range from 0 to 40 ng/mL in triplicate was added in each well. The effects of Peg-rHuMGDF were studied on basic fibroblast growth factor (bFGF, 1 ng/mL) stimulated or unstimulated BCEs. bFGF is an angiogenic growth factor that stimulates proliferation of endothelial cells. After 72 h, the cells were washed with PBS and fixed with ethanol. Subsequently, the cells were stained with methylene blue in borate buffer for 10 min. The plates were then thoroughly washed and dried. After the addition of 100 μL of 0.1 M HCl per well and a 20-min incubation at 37°C, plates were read at 600 nm. Colorization of the HCl represents the cell number, which is expressed by the plate reader in numbers (15).

Corneal neovascularization assay.

In the stroma of the mouse cornea adjacent to the limbus, pellets were implanted with bFGF as described previously (16). In brief, after anesthetizing the mice, 0.4 × 0.4-mm pockets were made in the cornea. Subsequently, 80 ng bFGF pellets were implanted 1.0–1.2 mm from the limbal vessels. Then, erythromycin was topically applied (E. Fougera, Melville, NY). The vascular response to the bFGF pellets was measured 5 d after implantation by maximal vessel length and number of clock hours of neovascularization. The area of corneal neovascularization was calculated by using a modified formula of a half ellipse, which best approximates the area of neovascularization: Area (mm2) = [πp× clock hours × length (mm) × 0.2 mm].

Statistical analysis.

Results are expressed as mean ± SEM. Instat (Mac) program was used for statistical analysis. Statistically significant differences were calculated by unpaired t test.

RESULTS

Platelet counts.

Once the tumor reached a size of 150–200 mm3 (approximately 4 d after implantation), platelet counts were already depressed from normal levels of 800,000 ± 71,000/μL to 600,000 ± 59,000/μL (p< 0.01). After an additional 11 d, platelet counts in untreated animals had fallen to 300,000 ± 102,240/μL. However, mice treated with Peg-rHuMGDF had significantly elevated platelet counts of 2.4 million ± 0.56/μL (p< 0,001, as shown in Fig. 1). This elevation was sustained at d 18 in treated animals (platelet counts 1.46 million ± 0.18/μL), when most of the untreated control animals had died. Platelet counts of untreated mice that were still alive at d 18 were 151,000 ± 22,000/μL (mean of multiple experiments). The cause of death of these mice was assumed to be the result of severe thrombocytopenia and subsequent anemia, as described by others (14).

Figure 1
figure 1

Platelet counts in eoma-bearing mice. Platelet counts at d 11 of treatment with 10 μg/kg of Peg-rHu-MGDF vs PBS as control.

Full size image

Survival.

The mice were treated for a 30-d period. At the conclusion of treatment, 50% of treated mice were still alive versus none of the untreated animals. This can also be expressed as a minimum of 9 d of prolongation of survival in treated animals versus control (p< 0.001, n= 8). Figure 2 shows a representative experiment.

Figure 2
figure 2

Survival of eoma-bearing mice as expressed in percentage alive (treatment with 10 μg/kg Peg-rHu-MGDF, n= 4, controls n= 5).

Full size image

Tumor growth.

Treatment with Peg-rHuMGDF inhibited tumor growth by 75% (p< 0.001, n= 8). Figure 3 shows a representative experiment. In a nonplatelet-trapping mouse tumor model (Lewis lung carcinoma), Peg-rHuMGDF did not significantly effect tumor growth after 14 d of treatment, despite the elevation of platelet counts by 2–3-fold (data not shown).

Figure 3
figure 3

Tumor growth during treatment with Peg-rHu-MGDF (treated vs controls = 75% inhibition).

Full size image

Immunohistochemistry.

H&E staining of embedded tumor tissue showed prominent endothelial cells and many mitoses of the Eoma cells. The treated tumors were more solid than the untreated, which had more intratumoral spaces filled with blood cells. In addition, intraluminal fresh clots and organized clots were observed in the treatment group, indicative for infarction of the tumors. We were unable to count microvessel density of the tumors as a result of a diffuse staining by von Willebrand factor antibody of both endothelial and Eoma cells.

Cell culture experiments.

Peg-rHuMGDF (20 ng/mL) had no effect on the proliferation of Eoma cells in vitro. In addition, it did not influence the proliferation of BCEs, which were cultured either with or without stimulation of bFGF.

Corneal neovascularization assay.

Peg-rHuMGDF treatment (10 μg/kg) for 5 d did not produce a significant inhibition on bFGF-induced corneal neovascularization (p= 0.08, area of corneal neovascularization = 2.0 mm2 ± 0.2, n= 8 versus 2.2 mm2 ± 0.2, n= 8 for the treatment group and the controls, respectively).

DISCUSSION

Kasabach-Merritt syndrome is a life threatening complication in children with vascular anomalies growing as kaposiform hemangioendothelioma (3, 6). The mortality rate is 12% or higher (2). Patients die of bleeding, sepsis, and/or invasion of vital structures. Platelet transfusions have been given in these children to prevent them from bleeding to death but can worsen the clinical condition (4). It is unknown why transfusions can cause these unwanted effects, but it is assumed that the platelets are immediately trapped by the tumor, become activated, and release their growth factors. These growth factors may cause the adverse effects that have been observed. Storage of platelets, before transfusion, is known to activate platelets to some extent (13).

In a mouse model of the KMS in which mice die from bleeding complications, including anemia (14), treatment with Peg-rHuMGDF increased platelet counts by 7–8-fold, preventing bleeding and lethargy, resulting in significantly prolonged survival. The elevation of platelet counts to levels several times normal is not unusual and is also seen when Peg-rHuMGDF is given to patients with advanced cancer before chemotherapy (11). Interestingly, the ability of Peg-rHuMGDF to enhance platelet production in a mouse model of KMS refutes the assumption that platelet production is already maximally stimulated by endogenous thrombopoietin. Indeed, there is considerable evidence that platelets are self-regulating the activity of endogenously produced thrombopoietin. They express the Mpl-receptor expressed on their cell surface and scavenge free-circulating thrombopoietin (17). It has been found that patients with thrombopenia owing to a decreased production of platelets (bone marrow disease) have very high plasma thrombopoietin levels. In contrast, patients with thrombopenia owing to an increased platelet turnover from autoimmune thrombocytopenia have plasma thrombopoietin concentrations within normal range, indicating that platelets are still able to bind free thrombopoietin even when circulating for only a short time (18). This last possibility may explain why treatment with Peg-rHu-MGDF can increase platelet counts in KMS mice. Based on clinical studies (4), it is expected that the Eoma-bearing mice suffer from an increased platelet turnover rather than a bone marrow failure.

The mechanism of how increased platelet production inhibited tumor growth by 75% remains unknown, but there are several hypotheses. 1) Trapped platelets could inhibit tumor growth by secretion of angiogenesis inhibitors, like platelet factor 4 and thrombospondin, and this secretion may be in excess to the released angiogenesis stimulators by platelets (e.g. vascular endothelial growth factor) (19). Although possible, it seems unlikely that the secretion products of platelets have an inhibitory effect on angiogenesis, as serum (containing the secretion products of platelets) is generally stimulatory for cell lines and primary cultures of endothelial cells (20). 2) Peg-rHuMGDF may inhibit angiogenesis directly or has a direct inhibitory action on the EOMA cells in vivo. We were unable to find an inhibitory activity of Peg-rHuMGDF on cultured EOMA or endothelial cells. As well, Peg-rHuMGDF was not antiangiogenic in an in vivo assay of corneal neovascularization. Further, using a murine tumor (Lewis lung carcinoma), no antiangiogenic inhibition of tumor growth was detected, despite an elevation of platelet counts to over 3 million with Peg-rHuMGDF. 3) The combination of a persistent increase in platelet counts and a platelet-trapping tumor may cause thrombosis of tumor vessels, similar to antivascular therapy (21). The latter possibility was supported by H&E staining of the tumor specimens. Fresh fibrin clots were observed in the treated specimens, whereas clots could not be found in the untreated group. Similar to our finding, it has been described in a case report of a patient with KMS that generating intratumoral thrombosis by aminocaproic acid and cryoprecipitate can be effective (22). Additionally, the observation of a lack of effect in the Lewis lung carcinoma also supports the idea that this antivascular effect is seen only in tumors that trap platelets to such an extent. These data may explain why isolated transfusions of platelets have not shown a benefit in the clinic due to the lack of persistent elevation of platelets necessary for the antivascular effect.

In conclusion, these data indicate in a preclinical model that stimulation of the platelet production is effective against KMS. Two antiangiogenic drugs, angiostatin (23) and TNP-470 (24), have been reported to potently inhibit tumor growth in this mouse model of KMS. However, TNP-470 is currently being tested in the phase-2 clinical trial, and angiostatin is not yet available for patients. Although the stimulator of the platelet production Peg-rHuMGDF used in this study is not available for clinical use anymore, the recombinant form of the natural circulating Tpo is currently being tested in clinical trials. In addition, another stimulator of platelet production, IL-11, is also under clinical investigation in adults. It is expected that these agents will be effective in children (newborns) as well, because Tpo was capable of inducing a dose-dependent proliferation of megakaryocyte precursors of preterm and term babies (25). Based on our findings, we expect that Tpo and IL-11 may be of benefit in life-threatening complications for children suffering from KMS, for whom no effective treatment is known.