Biology and physiology of thrombopoietin

INTRODUCTION — Thrombopoietin (TPO) is the physiologically relevant regulator of platelet production. Although the concept of a platelet growth factor analogous to erythropoietin had been proposed in the 1950s, it was not until 1994 that the existence of this hematopoietic growth factor was demonstrated and the protein purified [1-5]. Although historically called "thrombopoietin" [6], its discoverers also called it by several other names, including megapoietin [3], megakaryocyte growth and development factor (MGDF) [4], and c-Mpl ligand [2]. The last name is often used instead of TPO because the receptor for TPO, called c-Mpl, was discovered prior to the identification of TPO [7] and was instrumental in helping to purify the ligand (ie, the c-Mpl ligand) that bound to it.

This topic will review the biology and physiology of TPO. The potential clinical applications of TPO, ranging from the management of thrombocytopenic states to improving yields from platelet apheresis, are discussed separately [8]. (See "Clinical applications of thrombopoietic growth factors".)

STRUCTURE OF THROMBOPOIETIN — TPO is produced primarily in liver parenchymal cells with much smaller amounts being made in the kidney and bone marrow [9,10]. It is synthesized as a 353 amino acid precursor protein with a molecular weight of 36 kDa [2,4,11]. Following the removal of the 21 amino acid signal peptide, the remaining 332 amino acids undergo glycosylation to produce a 95 kDa glycoprotein (figure 1). The glycoprotein is then released into the circulation with no apparent intracellular storage in the liver.

TPO is an unusual hematopoietic growth factor in a number of ways:

●It is much larger than most other regulators of blood cell production such as G-CSF (granulocyte colony-stimulating factor) and erythropoietin.

●It has an unusual structure in that the first 153 amino acids of the mature protein are 23 percent homologous with human erythropoietin (figure 1) [12] and probably 50 percent similar if conservative amino acid substitutions are considered. This region also contains four cysteine residues just like those in erythropoietin and is highly conserved among different species. Despite these similarities, TPO does not bind the erythropoietin receptor and erythropoietin does not bind the TPO receptor.

●Amino acids 154 to 332 comprise a novel sequence that contains six-N-linked glycosylation sites and is less well conserved among different species. Structure-function studies have demonstrated that the first 153 amino acids of the c-Mpl ligand are all that is required for its thrombopoietic effect in vitro [2,4]; however, this truncated molecule has a markedly decreased circulatory half-life compared with the 20 to 40 hour half-life of the native protein [13]. Presumably, the glycosylated second half of the molecule confers stability and prolongs the circulatory half-life. Similar carbohydrate sequences regulate the stability of erythropoietin [14].

The crystal structure of TPO has been determined and reveals an antiparallel four-helix bundle fold with two different binding sites for the TPO receptor. One has a binding constant for the TPO receptor of 3.3 x 10^-9 M and the other 1.1 x 10^-6 M [15].

THROMBOPOIETIN GENE — There is a single copy of the gene for TPO (THPO) on human chromosome 3q27-28 [11,12,16]. The gene spans approximately 7 kb with seven exons, the first two of which are noncoding. The third exon contains part of the 5'-untranslated sequence and part of the signal peptide. The erythropoietin-like region is coded for by exons 4 to 7 and all of the carbohydrate domain is encoded by exon 7.

Comparison with the gene for erythropoietin shows conservation of the boundaries of the coding exons except for the addition of the carbohydrate domain sequence in the final exon of the THPO gene. In addition to the functional mRNA encoded (TPO-1), two other nonfunctional mRNA sequences (TPO-2 and TPO-3) are present due to alternative splicing [11,12].

Studies in mice have challenged assumptions about whether hepatic TPO production is constitutive or induced. In mouse model studies, the Ashwell-Morell receptor (AMR) on hepatocytes can bind platelets that have lost sialic acids on their surface. This binding activates a signaling pathway that increases hepatic TPO mRNA and TPO production [17]. This pathway might partially explain thrombocytosis seen in JAK2 mutated myeloproliferative neoplasms; whether it plays any role in normal human physiology is unclear. Subsequent studies called this analysis into question by showing that it is hepatic Kupffer cells, not hepatocytes, that clear senescent platelets from the circulation [18].

THROMBOPOIETIN RECEPTOR — TPO acts through the TPO receptor, called c-Mpl. Prior to the identification of TPO, a murine myeloproliferative leukemia virus was identified that contained a retroviral oncogene, v-Mpl (ie, myeloproliferative leukemia); this oncogene encoded the cytoplasmic portion of a membrane protein which had all the characteristics of a new hematopoietic growth factor receptor [19].

When the full-length homolog (c-Mpl) was cloned [7], its mRNA was found to be present at significant levels primarily in platelets and megakaryocytes and in a small percentage of CD34+ cells (which are early hematopoietic precursors) [20]. The identification of c-Mpl as the putative TPO receptor was further strengthened by the demonstration that the formation of megakaryocyte (Meg-CFC), but not myeloid (GM-CFC) or erythroid (E-BFU), precursors in bone marrow culture was decreased when synthesis of the c-Mpl protein was inhibited by the addition of c-Mpl antisense constructs [21].

It is now known that the TPO receptor (c-Mpl) is present on platelets and megakaryocytes and, at a lesser density, on most other hematopoietic precursor cells. Studies with human platelets indicate the presence of approximately 56 ± 17 receptors per human platelet with an affinity of 163 ± 31 pmol/L [22].

Upon binding to TPO, the receptor undergoes dimerization that results in a number of signal transduction events that prevent apoptosis, improve cell viability, promote growth, and possibly increase differentiation in megakaryocytes [23]. In addition, binding to the receptor provides the major mechanism by which TPO is removed from the circulation by platelets and possibly megakaryocytes [3,24-27]. After TPO binds to platelets, the receptor-ligand complex undergoes internalization and the bound thrombopoietin is degraded [22,28]. The receptor is not re-expressed on the platelet surface [29].

PHYSIOLOGY OF THROMBOPOIETIN — Much has been learned about the normal physiology of TPO, its effects upon bone marrow precursor cells, megakaryocytes, and platelets, as well as regulation of its concentration [30,31].

Effects on platelet production — TPO is the only physiologically relevant regulator of platelet production; it acts to "amplify" the basal production rate of megakaryocytes and platelets. When TPO or its receptor have been "knocked-out" by homologous recombination in mice, the megakaryocyte and platelet mass are reduced to approximately 10 percent of normal, but the animals are healthy and do not spontaneously bleed (figure 2) [32-34]. The neutrophil and erythrocyte counts are normal. In animals in which only one copy of the TPO gene (THPO) has been deleted, the platelet count is reduced to approximately 65 percent of normal. Such TPO-deficient mice can increase their platelet count if treated with other thrombopoietic growth factors such as interleukin (IL)-6, IL-11, and stem cell factor (c-kit ligand) [35]. However, neither endogenous IL-6 nor endogenous IL-11 appears to be responsible for the residual platelet production in these animals [36].

Binding of TPO to its receptor prevents apoptosis of megakaryocytes [37] and increases their number, size, and ploidy (see "Megakaryocyte biology and platelet production"). The rate of cellular maturation is probably also increased. These events are mediated via signal transduction pathways involving JAK, STAT, and other intracellular mediators (figure 3). Furthermore, the addition of TPO to CD34+ cells in culture results in the majority of cells becoming megakaryocytes and then shedding platelets [38]. This last step, the shedding of platelets from megakaryocytes, does not require, and actually may be inhibited by, the presence of thrombopoietin [39].

When TPO or its receptor have been "knocked-out" by homologous recombination in mice, the megakaryocyte and platelet mass are reduced to approximately 10 percent of normal, but the animals are healthy and do not spontaneously bleed (figure 2). The factors responsible for the residual platelet production in this setting are not well understood.

A predictable response occurs after the daily administration of a recombinant form of TPO to baboons (figure 4A-B) [40,41]. During the first four days of administration, bone marrow megakaryocyte ploidy rises to a maximum but there is no change in the platelet count. On day five, the platelet count begins to rise and does so at a dose-dependent rate. With continued administration of TPO, a dose-dependent plateau platelet count is attained on days 8 to 12. There is a log-linear relationship between the TPO dose and the plateau platelet count with the maximum response being a sixfold increase in the rate of platelet production. Upon stopping the TPO, the platelet count returns to its baseline over 10 days without a rebound thrombocytopenia.

A similar time course and platelet response have been demonstrated after recombinant TPO administration to humans with no apparent toxicity [42-44] (see "Clinical applications of thrombopoietic growth factors"). There is no effect on red or white blood cells.

Effects on platelet function — In addition to increasing the number of megakaryocytes and platelets, TPO can also affect the function of platelets. When TPO binds to its platelet receptor, it induces phosphorylation of the c-Mpl receptor and a number of other molecules in several different signal transduction pathways [45-47]. Although TPO does not cause platelet activation directly, it reduces the threshold for activation by other platelet agonists such as ADP and collagen by 50 percent. It is unclear if this is a clinically relevant effect posing a risk for thrombosis. (See "Clinical applications of thrombopoietic growth factors".)

Effects on bone marrow precursor cells — Although TPO affects late cellular maturation events only in megakaryocytes, it and its receptor (c-Mpl) appear to play an important role in the regulation of the growth of early precursors of other lineages and even of the pluripotential stem cell [48-51]. (See "Megakaryocyte biology and platelet production", section on 'Thrombopoietin (role in megakaryocyte maturation)'.)

Most evidence suggests that binding of TPO to c-Mpl enhances the survival and proliferation of already committed precursor cells, rather than a true stimulation of differentiation from stem cells. Homozygous knock-in mice, in which the gene for c-Mpl was replaced by a gene for a chimeric molecule consisting of the extracellular domain of c-Mpl and the cytoplasmic domain of the G-CSF receptor, have a normal platelet count [52]. Since the intracellular domain of the G-CSF receptor can functionally replace the signaling activity of c-Mpl, this result suggests that thrombopoietin primarily affects cells already committed to the megakaryocyte lineage.

When administered to animals, TPO increases bone marrow and peripheral blood megakaryocyte precursor cells (Meg-CFC), bone marrow megakaryocyte number and ploidy, and the platelet count. In addition, both erythroid and multipotential precursor cells are increased in the bone marrow and peripheral blood but without affecting the erythrocyte or neutrophil count.

In animals made deficient in TPO or its receptor (c-Mpl), Meg-CFC are reduced by 90 to 95 percent as expected, and myeloid and erythroid precursor cells are reduced by 60 to 80 percent (figure 5) [32,35]. The normal neutrophil and erythrocyte counts in these animals are presumably maintained by intact feedback mechanisms mediated by G-CSF and erythropoietin.

Regulation of circulating levels — Hepatic TPO production is usually constitutive; circulating levels are determined by the circulating platelet mass (figure 6). In contrast with the production of red blood cells, which is regulated by a system that senses hypoxia and alters the rate of transcription of the erythropoietin gene, there is no such "sensor" of the platelet mass [3,24-26,53]. (See "Megakaryocyte biology and platelet production", section on 'TPO role in maintaining platelet count' and "Regulation of erythropoiesis", section on 'Hypoxia and EPO expression'.)

TPO mRNA is produced at the same rate in healthy and thrombocytopenic individuals and is reduced in patients with liver disease [54]. TPO levels can be increased by the cytokine interleukin 6 (IL-6), as illustrated in the examples below. (See 'Inflammatory disorders and ovarian cancer' below.)

There is no evidence for post-transcriptional regulation [55].

Once TPO is produced, circulating TPO levels are regulated by the volume of the total platelet mass. Platelets and megakaryocytes contain high affinity TPO (c-Mpl) receptors that bind and clear TPO from the circulation, directly determining the circulating TPO concentration [22]. Evidence for this includes:

●When platelet production is decreased, as in thrombocytopenia due to megakaryocytic hypoplasia, clearance of TPO is reduced and levels rise [22,25].

●Plasma TPO concentrations fall after platelet transfusion [56].

This sort of feedback system is not unusual in hematology. Both M-CSF and G-CSF are normally regulated primarily by the amount of circulating monocytes and neutrophils, respectively [57]. It seems as if only erythropoietin has a true sensor of the circulating blood cell mass that in turn alters production of this hematopoietic growth factor.

DISORDERS OF THROMBOPOIETIN AND ITS RECEPTOR — Understanding of the physiology of TPO has led to consideration of a number of possible clinical disorders ranging from excess TPO production to decreased expression of the TPO receptor [53]. Some of these postulated disorders have now been identified in humans or animals.

Familial thrombocythemia — Familial thrombocythemia, also called hereditary or inherited thrombocythemia or thrombocytosis, is a rare disorder that is transmitted in an autosomal dominant pattern. This disorder results from activating mutations in the genes for TPO or c-Mpl or from mutations in the genes for other proteins [58,59]. Mutations in the TPO or c-MPL genes do not appear to be involved in the much more common sporadic cases of essential thrombocythemia. (See "Clinical manifestations, pathogenesis, and diagnosis of essential thrombocythemia", section on 'Familial essential thrombocythemia'.)

Liver failure — Since the liver is the primary site of TPO production and the TPO gene is usually not inducible, TPO deficiency may be responsible (perhaps along with splenic sequestration) for the thrombocytopenia seen in patients with liver failure. Partial resection of the liver in animals results in a proportional decrease in the platelet count [60] and plasma TPO concentrations appear to be inappropriately low in patients with cirrhosis [61]. Upon liver transplantation into patients with liver failure, TPO levels rise and platelet counts return to normal levels [62]. This observation has led to the suggestion that TPO treatment may be an effective therapy in patients with liver failure. (See "Clinical applications of thrombopoietic growth factors".)

Thrombocytopenia due to anti-thrombopoietin antibodies — In a clinical trial, administration of one recombinant form of TPO, PEG-rHuMGDF, caused the appearance of antibodies to the recombinant molecule in some subjects. The antibody cross-reacted with endogenous TPO and neutralized its activity. Since TPO is produced in a constitutive fashion, these patients developed TPO deficiency and severe amegakaryocytic thrombocytopenia [63,64]. One other individual with amegakaryocytic thrombocytopenia was diagnosed with an anti-TPO antibody and no prior exposure to recombinant TPO. This patient had severe thrombocytopenia and improved following treatment with cyclosporine A [65]. Subsequently, many patients with apparent immune thrombocytopenia (ITP) have been screened; of 205 ITP patients screened, only one has been found to have an anti-TPO antibody as the genesis for the thrombocytopenia [66]. Of 961 ITP patients on clinical trials with romiplostim, none were found to have neutralizing antibodies to TPO [67]. (See "Hematologic manifestations of systemic lupus erythematosus", section on 'Thrombocytopenia'.)

Thrombocytopenia due to anti-c-Mpl antibodies — Measurement of anti-c-Mpl antibodies is not standardized or clinically available. Several studies have suggested that anti-c-Mpl antibodies may be associated with thrombocytopenia due to inhibition of megakaryocyte growth. In one study, 8 out of 69 (11.6 percent) of patients with systemic lupus erythematosus and 7 out of 84 (8.3 percent) of patients with ITP had anti-c-Mpl antibodies in their serum; among 84 healthy controls, no anti-c-Mpl antibodies were present [68]. Antibodies from some of these patients inhibited megakaryocyte growth in culture. Another study reported that 54 out of 187 (29 percent) ITP patients (but none of 59 healthy controls) had antibodies against c-Mpl; the antibody-positive patients appeared to have fewer megakaryocytes and platelets and less responsive disease [69]. This area requires more study and validation of assay sensitivity and specificity before clinical significance is established.

Amegakaryocytic thrombocytopenia and absent TPO receptor — Congenital amegakaryocytic thrombocytopenia (CAMT) is a rare disorder that presents with severe thrombocytopenia and absence of megakaryocytes in the bone marrow. Some of these patients are homozygous for mutations that produce an inactive TPO receptor (c-Mpl) leading to minimal platelet production similar to that seen in c-Mpl knockout mice [70]. Some of these patients ultimately develop bone marrow aplasia, confirming the multipotential effect of TPO activity [71,72].

Thrombocytosis and chromosomal defects — A number of hematopoietic disorders associated with thrombocythemia or abnormal megakaryocyte formation have been associated with defects involving chromosome 3q [73], and some myeloid leukemias associated with thrombocytosis have a characteristic rearrangement of chromosome 3q21 and 3q26 [16]. Since the TPO gene is located on chromosome 3q27-28 [11,12,16], it has been suggested that the TPO gene might be mediating these effects. However, closer analysis of these chromosome regions in these patients has not demonstrated involvement of the TPO gene [16,74] and blood TPO levels have been normal. These findings suggest that other genes close to the TPO gene may be responsible for other aspects of megakaryocyte differentiation and growth. Patients with the 5q- syndrome often have pathogenic variants affecting the ribosomal protein gene RPS14 (not in the gene for TPO or its receptor Mpl), with thrombocytosis attributed to haploinsufficiency of the microRNA genes miR-145 and miR-146a [75,76]. (See "Cytogenetics, molecular genetics, and pathophysiology of myelodysplastic syndromes/neoplasms (MDS)".)

Myeloproliferative disorders — Mutations in the TPO receptor (MPL exon 10) are found in <5 percent of patients with either essential thrombocythemia or primary myelofibrosis, but not in patients with polycythemia vera. Overall survival of those with primary myelofibrosis with the MPL mutation is no different from patients carrying the JAK2 V617F mutation (8 and 11 years, respectively) but shorter than patients with the CALR mutation (21 years) [77]. Overall survival of patients with essential thrombocythemia and the MPL mutation was no different from patients carrying either the JAK2 V617F or CALR mutations. The relative thrombotic risks associated with MPL mutation versus the JAK2 V617F or CALR mutations is unknown due to the low numbers of patients with the MPL mutation. (See "Overview of the myeloproliferative neoplasms", section on 'MPL mutations'.)

Inflammatory disorders and ovarian cancer — Inflammatory disorders can cause an increase in the levels of the cytokine interleukin-6 (IL-6). IL-6 has minimal effect on normal platelet production (in a mouse model, the levels in wild-type and IL-6 knockout animals were 1.16 x 10¹² and 0.90 x 10¹² cells/L, respectively); however, IL-6 may affect TPO production in pathologic settings and may increase platelet production [78]. When inflammation was induced in these mice, hepatic TPO mRNA increased and plasma levels of IL-6 and TPO rose in the wild-type but not the IL-6-deficient animals; however, there was no increase in platelet count [78].

A study involving administration of IL-6 to six individuals with advanced cancer showed increased TPO levels in response to five days of continuous IL-6 infusion [79]. Administration of IL-6 to mice induced thrombocytosis and was linked to increased hepatic TPO mRNA and plasma TPO levels; antibody neutralization of TPO eliminated the thrombocytosis [79].

In a separate study involving patients with ovarian cancer, thrombocytosis was associated with increased levels of IL-6 and TPO; neutralization of the IL-6 eliminated the thrombocytosis [80]. In mice bearing these human ovarian tumors, the rise in platelet count was mediated by IL-6-induced hepatic TPO production.

ASSAYS FOR THROMBOPOIETIN — Measurement of the circulating concentration of TPO may be clinically useful [81-83]; commercial ELISA assays are available. Several conclusions may be made:

●TPO concentrations are increased 10 to 20-fold over normal in bone marrow failure states such as aplastic anemia or following myeloablative chemotherapy [84].

●TPO levels are only slightly elevated above normal in immune thrombocytopenia (ITP) at platelet counts comparable to those seen in states of bone marrow failure [84]. The relatively normal plasma TPO concentration in ITP probably reflects enhanced TPO clearance by the increased mass of bone marrow megakaryocytes and by the increased flux of platelets through the circulation.

●TPO levels are normal or slightly elevated in essential thrombocythemia [85].

●TPO levels may allow the distinction between states of increased platelet production (normal TPO concentrations) versus decreased platelet production (elevated TPO concentrations) (figure 7) [81,84].

●TPO levels may predict response to TPO receptor agonists in patients with ITP.

SUMMARY

●TPO production – Thrombopoietin (TPO), a 95 kDa glycoprotein (figure 1) primarily produced in the liver, is the physiologically relevant regulator of platelet production, amplifying the basal production rate of megakaryocytes and platelets. (See 'Physiology of thrombopoietin' above.)

●TPO function – TPO acts through its receptor (c-Mpl), present on platelets and megakaryocytes and, at a lesser density, on most other hematopoietic precursor cells. Binding of TPO to its receptor prevents apoptosis of megakaryocytes and increases their number, size, and ploidy. (See 'Thrombopoietin receptor' above and "Megakaryocyte biology and platelet production".)

●TPO regulation – Hepatic TPO production is mostly constitutive; circulating levels are determined by the circulating platelet mass (figure 6). TPO production may be reduced in liver failure and increased by interleukin-6 (IL-6) in inflammatory states and cancer. (See 'Regulation of circulating levels' above and 'Liver failure' above and 'Inflammatory disorders and ovarian cancer' above.)

●Clinical disorders of TPO or c-Mpl – Disorders involving TPO or c-Mpl include the following. (See 'Disorders of thrombopoietin and its receptor' above.)

•Familial thrombocythemia is a rare autosomal dominant disorder most commonly resulting from activating mutations in the genes for TPO or c-Mpl. (See "Clinical manifestations, pathogenesis, and diagnosis of essential thrombocythemia", section on 'Familial essential thrombocythemia'.)

•Antibodies to TPO or c-Mpl may rarely cause thrombocytopenia, often associated with reduced numbers of bone marrow megakaryocytes. (See 'Thrombocytopenia due to anti-thrombopoietin antibodies' above and 'Thrombocytopenia due to anti-c-Mpl antibodies' above.)

•Congenital amegakaryocytic thrombocytopenia is a rare genetic disorder with severe thrombocytopenia and absence of bone marrow megakaryocytes. Some patients are homozygous for inactivating mutations of c-Mpl. (See "Aplastic anemia: Pathogenesis, clinical manifestations, and diagnosis", section on 'Abnormal thrombopoietin or its receptor'.)

•Hepatic TPO deficiency is a major cause of thrombocytopenia in patients with liver failure. (See "Hemostatic abnormalities in patients with liver disease", section on 'Thrombocytopenia and platelet dysfunction'.)