Plasma thrombopoietin levels in thrombocytopenic states: implication for a regulatory role of bone marrow megakaryocytes


Dr HansWadenvik Department of Internal Medicine, Sahlgrenska University Hospital/Sahlgrenska, S-413 45 Gothenburg, Sweden.


To evaluate the diagnostic value of thrombopoietin (TPO, c-mpl ligand) measurements, and clarify the regulatory mechanisms of TPO in normal and in thrombocytopenic conditions, the plasma TPO concentration was determined in normal individuals (n = 20), umbilical cord blood (n = 40), chronic idiopathic thrombocytopenic purpura (ITP; n = 16), in severe aplastic anaemia (SAA; n = 3), chemotherapy-induced bone marrow hypoplasia (n = 10), myelodysplastic syndrome (MDS; n = 11), and sequentially during peripheral blood progenitor cell transplantation (n = 7). A commercially available ELISA and EDTA-plasma samples were used for the analysis. The plasma TPO concentration in the normals and umbilical cord blood were 52 ± 12 pg/ml and 66 ± 12 pg/ml, respectively. The corresponding values in patients with SAA and chemotherapy-induced bone marrow hypoplasia were 1514 ± 336 pg/ml and 1950 ± 1684 pg/ml, respectively, and the TPO concentration, measured sequentially after myeloablative chemotherapy and peripheral blood progenitor cell transplantation, was inversely related to the platelet count. In contrast, the plasma TPO recorded in patients with ITP (64 ± 20 pg/ml) and MDS (68 ± 23 pg/ml) were only slightly higher than normal levels. In conclusion, TPO levels were significantly elevated in patients in which bone marrow megakaryocytes and platelets in circulation were markedly reduced, whereas TPO levels were normal in ITP patients, and only slightly increased in the MDS patients. These latter patients displayed a preserved number of megakaryocytes in bone marrow biopsies. Our data support the suggestion that megakaryocyte mass affects the plasma TPO concentration. In thrombocytopenic patients a substantially increased plasma TPO implies deficient megakaryocyte numbers. However, TPO measurements do not distinguish between ITP and thrombocytopenia due to dysmegakaryopoiesis, as seen in MDS patients.

Thrombopoietin (TPO), the ligand for the c-mpl proto-oncogene, also known as the megakaryocyte growth and development factor (MGDF), has been purified and cloned by several groups ( Kaushansky, 1995). It is expressed mainly in liver and kidney ( de Sauvage et al, 1994 ; Lok et al, 1994 ), and regulates the growth and maturation of megakaryocytes ( de Sauvage et al, 1994 ; Lok et al, 1994 , Kaushansky, 1995). When administered to normal mice or non-human primates, recombinant TPO induces a significant increase in bone marrow megakaryocyte colonies and the circulating platelet number ( Kaushansky et al, 1994 ; Farese et al, 1995 ). Also, the clinical effectiveness of pegylated and recombinant TPO has been demonstrated in patients treated with chemotherapy because of advanced cancer ( Basser et al, 1996 ; Fanucchi et al, 1997 ). However, the homeostatic factors that maintain a constant platelet mass under normal circumstances have not been fully defined. Studies of experimental thrombocytopenia induced in animals by chemotherapy, radiation or antiplatelet antibodies indicate an inverse relationship between the circulating platelet counts and TPO levels, suggesting that TPO is regulated by a feedback mechanism dependent on the platelet mass. It has been proposed that circulating TPO is selectively cleared by platelets through receptor-mediated endocytosis and destruction ( Kuter & Rosenberg, 1994, 1995; Kuter et al, 1994 ; Wendling et al, 1994 ; Ulich et al, 1995 ). Further support for this hypothesis was obtained by Stoffel et al (1996 ). These investigators demonstrated that the TPO concentration during thrombocytopenia is not regulated at the TPO mRNA level, but at a post-transcriptional level and through absorption and metabolism by platelets. Other findings indicate that megakaryocyte mass may be another important regulator of plasma TPO. Nagata et al (1997 ) found a relationship between serum TPO and megakaryocytes in bone marrow and spleen during acute thrombocytopenia in mice. Shivdasani et al (1997 ) found that mega-karyocytes acted as a sink for plasma thrombopoietin in knockout mice lacking the erythroid transcription factor NF-E2; these mice exhibit maturation arrest of megakaryocytes, megakaryocytosis and profound thrombocytopenia.

To study whether this model of plasma TPO regulation is operative in thrombocytopenic patients and to evaluate the value of TPO measurements in the differential diagnosis of thrombocytopenic states, we studied groups of patients with thrombocytopenia due to: (i) enhanced platelet destruction, (ii) megakaryocyte deficiency, and (iii) bone marrow dysplasia.



Plasma samples were obtained from 20 healthy volunteers (23–60 years old; 11 males and nine females), 16 patients with chronic idiopathic thrombocytopenic purpura (ITP; 17–76 years old; nine males and seven females), three patients with severe aplastic anaemia (SAA; 54–79 years old; one male and two females), 10 patients with chemotherapy-induced thrombocytopenia (42–82 years old; five males and five females), 11 patients with thrombocytopenia due to a myelodysplastic syndrome (MDS; 52–80 years old; two males and nine females), and sequentially from seven patients undergoing myeloablative chemotherapy with peripheral blood progenitor cell support (44–59 years old; four males and three females; multiple myeloma, two patients; Hodgkin's disease, one patient; non-Hodgkin's disease, one patient; acute leukaemia, two patients; and breast cancer, one patient). In the patients treated with chemotherapy the samples were collected 10–14 d after start of treatment. Also, umbilical cord blood plasma samples were obtained from 40 healthy new-born babies.

TPO assay

Plasma was separated from EDTA-anticoagulated whole blood by centrifugation at 1000 g for 10 min and stored at −20°C until assay. The plasma TPO concentration was determined by a commercially available ELISA kit (QuantikineTM Human TPO Immunoassay, R&D Systems, Minneapolis, Minn., U.S.A.). Briefly, 200 μl of recombinant human TPO standard, plasma samples or blank were added in duplicates to the wells of a microtitre plate precoated with an anti-TPO monoclonal antibody. The plate was incubated for 3 h at 4°C. After washing, 200 μl of horseradish peroxidase conjugated anti-TPO antibody was added and incubated for 1 h at 4°C. The colour was developed by using tetramethylbenzidine as substrate. The reaction was stopped by adding 50 μl of acid solution to each well, and the absorbance was recorded at 450 nm. The sample TPO concentration was calculated from the corresponding standard curve. According to the manufacturer, this assay recognizes recombinant and natural TPO, and has a detection limit of less than 15 pg/ml.

Bone marrow morphology

Bone marrow biopsies were obtained from the posterior iliac crest, fixed in 4% neutral formaldehyde and embedded in paraffin. Paraffined sections were immunostained with an indirect immunoperoxidase technique ( Naish, 1989) using a rabbit antisera against glycocalicin, a cleavage product of platelet glycoprotein Ib. The number of megakaryocyte profiles was determined in the immunostained sections, by an experienced bone marrow pathologist, and with a previously described method ( Wadenvik et al, 1991 ). The rabbit anti-glycocalicin sera was a generous gift from Professor Nils-Olav Solum, Research Institute of Internal Medicine, Rikshospitalet, Oslo, Norway.


Standard statistical methods were used for calculation of mean values and standard deviation. Unless otherwise stated, the mean value ± SD is reported. The differences between mean values were evaluated using the two-tailed Student's t-test for unpaired data, and a P value < 0.05 was considered statistically significant.


The TPO results and the platelet count for the normal individuals and the different patient groups are summarized in Fig 1. The mean plasma TPO values in healthy volunteers and in umbilical cord blood were 52 ± 12 pg/ml (range 30–76 pg/ml) and 66 ± 12 pg/ml (range 46–115 pg/ml), respectively. In contrast, patients with SAA and chemotherapy-induced bone marrow hypoplasia displayed markedly increased TPO values (1514 ± 336 pg/ml and 1950 ± 1684 pg/ml), and differed significantly from the normals (P < 0.0001). Similarly, in the patients treated with myeloablative chemotherapy and peripheral blood progenitor cell rescue, the plasma TPO varied inversely with the peripheral platelet count (Fig 2).

Figure 1.

Fig 1. Plasma TPO and platelet count (mean ± SD) in normal individuals, umbilical cord blood and different patient groups. ITP: chronic idiopathic thrombocytopenic purpura; MDS: myelodysplastic syndrome; SAA: severe aplastic anaemia.

Figure 2.

Fig 2. Platelet count and plasma TPO (mean ± SE) in seven patients treated with myeloablative chemotherapy and peripheral blood progenitor cell rescue. The chemotherapy was administered on day −6 and the progenitor cell on day 0.

Patients with chronic ITP and with a MDS displayed only slightly increased plasma TPO values compared to the normals. The mean platelet count and plasma TPO in patients with ITP and MDS were 67 ± 54 × 109/l and 64 ± 20 pg/ml (range 34–106 pg/ml) versus 26 ± 20 × 109/l and 68 ± 23 pg/ml (range 38–113 pg/ml), respectively. The slight increase in plasma TPO seen in the MDS patients was statistically significant (P = 0.023), compared to the normals. There was no significant difference in mean TPO value between normals and ITP-patients (P= 0.052). All patients with chronic ITP had a normal or increased number of megakaryocytes in the bone marrow biopsies (data not shown). Also, all but one patient with MDS exhibited a normal or increased number of megakaryocytes in immunostained sections of the bone marrow biopsies ( 1 Table I), and the megakaryocytes were frequently small with dysplastic features.

Table 1. Table I. Clinical characteristics and bone marrow findings in 11 patients with myelodysplastic syndrome. nc: not classifiable according to the FAB classification; RA: refractory anaemia; RAEB: refractory anaemia with excess of blasts; RAEB-t: refractory anaemia with excess of blasts, in transformation; MPD: unclassified myeloproliferative disorder; nd: not determined.Thumbnail image of


Our results show that TPO levels were significantly elevated in patients with amegakaryocytic thrombocytopenia, in which bone marrow megakaryocytes and platelets in circulation were severely reduced. In contrast, TPO levels were normal in ITP patients, and only slightly increased in the MDS patients. These patients (ITP and MDS) displayed a well-preserved number of megakaryocytes in the bone marrow biopsies. These data strongly support the suggestion that TPO production is constitutive and plasma TPO levels are regulated by circulating platelets and bone marrow megakaryocytes. In thrombocytopenic patients a substantially increased plasma TPO implies deficient megakaryocyte numbers. Furthermore, in the patients with thrombocytopenia but a normal or increased number of bone marrow megakaryocytes, i.e. the patients with chronic ITP and MDS, the plasma TPO was only slightly increased, and overlapped with healthy volunteers. This latter finding indicates that TPO measurements cannot distinguish between chronic ITP and thrombocytopenia due to the dysmegakaryopoiesis seen in MDS patients. Furthermore, acute myeloid leukaemia cells have been reported to express the c-mpl protein (TPO-receptor) ( Vignon et al, 1993 ). Thus, it is possible that TPO was partly removed from the plasma by binding to the receptors on cells other than megakaryocytes, in patients with MDS and excess of blasts.

For the measurement of TPO levels, both bioassays and immunoassays have been developed. Although the presence of biologically active TPO can be measured with bioassays, they are not sensitive enough to detect low to normal TPO levels; in addition, bioassays are tedious and not specific. Moreover, when used for diagnostic purposes, they are not fully reliable since the presence of toxic substances in blood may affect cell growth ( Folman et al, 1997 ). On the other hand, although a few ELISA-based TPO assays are sensitive enough to measure TPO concentrations in healthy individuals ( Folman et al, 1997 ; Marsh et al, 1996 ; Tahara et al, 1996 ), the newly commercialized QuantikineTM TPO immunoassay has several advantages. The assay is readily available and appears to be one of the most sensitive TPO assay currently available, with a detection limit of < 15 pg/ml. Also, it has been shown that serum TPO levels are on average 3.4 times higher than the corresponding plasma TPO ( Folman et al, 1997 ), and it was suggested that the plasma-serum difference might be due to TPO or TPO degradation products released from platelets during blood clotting. Thus, it appears that free and circulating TPO is best measured on plasma samples.

By using other ELISAs, TPO has been measured in plasma and serum from normal individuals and patient populations ( Tahara et al, 1996 ; Marsh et al, 1996 ; Usuki et al, 1996 ; Emmons et al, 1996 ). It is generally agreed that the average level of TPO in normal human plasma is < 200 pg/ml ( Tahara et al, 1996 ; Marsh et al, 1996 ; Emmons et al, 1996 ). In patients with thrombocytopenia the TPO levels are usually elevated ( Chang et al, 1996 ; Tahara et al, 1996 ; Marsh et al, 1996 ; Emmons et al, 1996 ; Kosugi et al, 1996 ; Kunishima et al, 1996 ; Ichikawa et al, 1996 ; Meng et al, 1996 ; Kojima et al, 1997 ). However, there are some exceptions. Patients with ITP have been demonstrated to have TPO levels lower than would be predicted by their platelet counts ( Chang et al, 1996 ; Emmons et al, 1996 ; Kosugi et al, 1996 ; Kunishima et al, 1996 ; Ichikawa et al, 1996 ). TPO levels in patients with chronic liver disease and severe liver failure are also reported to be unexpectedly low. Since the liver is thought to be the primary site for TPO production, these low levels have been speculated to be due to disregulated production of TPO ( Shimodaira et al, 1996 ; Martin et al, 1997 ). Also, increased TPO levels have been observed in patients with elevated platelet counts due to myeloproliferative disorders, i.e. essential thrombocythaemia and polycythaemia vera ( Tahara et al, 1996 ; Cerutti et al, 1997 ; Horikawa et al, 1997 ). This increase in TPO was explained by the recent observation of reduced number and binding affinity of receptors for TPO, i.e. c-mpl protein, on ET platelets ( Li et al, 1996 ).

Little information is available regarding TPO levels in MDS patients ( Usuki et al, 1996 ; Kunishima et al, 1996 ). Usuki et al (1996 ) reported two patients with MDS, both displaying normal TPO levels, whereas Kunishima et al (1996 ) observed a significantly elevated TPO level in five MDS patients. However, in these studies TPO was measured on serum samples and very little data was given about the patients' clinical characteristics and bone marrow findings.

In conclusion, we found a heterogeneity with respect to TPO levels in different thrombocytopenic states, as well as their relationship to the platelet count. However, our data corroborate previous findings of other investigators in that TPO levels are typically high in conditions with insufficient platelet production secondary to a megakaryocytic hypoplasia. The marginally increased TPO levels in thrombocytopenic patients with normal or increased number of megakaryocytes further substantiate that circulating platelets alone do not control plasma TPO. The megakaryocyte mass seems to be another major regulator of plasma TPO concentration. Indeed, TPO measurements could not differentiate thrombocytopenia due to maturation arrest of megakaryocytes, from the increased platelet destruction seen in ITP.


This study was supported by grants from the Swedish Medical Research Council (project K98-19X-11630-031A), the Swedish Medical Society, the Göteborg Medical Society, Assar Gabrielssons Foundation, Stiftelsen Jubileumklinikens Forskningsfond mot Cancer, and FoU Västra Götaland. The authors thank Ms Iréne Andersson for expert technical assistance.