The cloning of Tpo, the central regulator of megakaryocyte and platelet production, has introduced a new potential therapy for thrombocytopenia in newborn infants. Platelet transfusion has been the only treatment available to minimize the risk of hemorrhage associated with most forms of neonatal thrombocytopenia, and therefore the possibility that cytokine therapy could contribute to prevention or management of neonatal thrombocytopenia is very appealing.

The humoral regulator of platelet production eluded researchers for more than 30 y. Although its characteristics were predicted in 1958, it was not until 1994 that Tpo was identified by five independent groups as the c-Mpl ligand (1). Two years earlier, the c-mpl gene had been cloned and identified as a cell surface receptor belonging to the cytokine-receptor superfamily, with expression restricted to megakaryocytes, platelets, and CD34+ hematopoietic precursors. With the cloning of the c-Mpl ligand, the elusive Tpo was quickly shown to fulfill its predicted role in stimulating megakaryocyte expansion, nuclear and cytoplasmic maturation, and platelet formation, both in vitro and in animal models. Tpo is a potent megakaryocyte colony-stimulating factor—it acts in synergy with other cytokines, including IL-3, IL-11, steel factor, and erythropoietin, to promote the proliferation of megakaryocytes as well as more primitive multilineage progenitor cells.

The c-DNA for human Tpo encodes a molecule of 332 amino acids. The amino-terminal 153 amino acid sequence (with 46% sequence similarity to erythropoietin) is the domain that binds to the c-Mpl receptor. The carboxyl-terminus, in contrast, has no known homology to any protein; deletion of this portion of the molecule does not affect activity but does reduce bioavailability. Plasma concentrations of Tpo are inversely related to platelet counts in states associated with decreased megakaryocyte mass such as aplastic anemia, whereas patients with destructive thrombocytopenias have only slightly elevated levels (2). Platelets regulate plasma concentrations of Tpo by binding and removing circulating Tpo. Subsequent to platelet transfusion in thrombocytopenic animals, plasma Tpo concentrations fall, and they rise only when thrombocytopenia recurs (3).

Thrombocytopenia is common in the neonatal intensive care unit. In well premature infants, the incidence of platelet counts less than 100 × 109/L is 2–5%. In sick neonates, the incidence is higher, affecting 22–35% of sick premature infants, and is associated with such serious conditions as maternal preeclampsia, intrauterine growth retardation, perinatal asphyxia, sepsis, and intraventricular hemorrhage (4–7). Although the etiologies of early newborn thrombocytopenia are multiple, two major mechanisms are operative: decreased bone marrow production and increased peripheral destruction. There is conflicting evidence regarding which of these mechanisms underlies most cases of neonatal thrombocytopenia. Castle et al. (6) demonstrated that early thrombocytopenia (occurring within the first 48 h of life) was associated with markers of platelet destruction: shortened platelet survival, rising mean platelet volumes, elevated levels of platelet-associated IgG, or evidence of consumptive coagulopathy. In a similar group of patients, Murray et al. (7) reported a decrease in the number of circulating megakaryocyte progenitor cells compared with nonthrombocytopenic newborns, suggesting that impaired platelet production was the principal cause. Both mechanisms probably contribute to the thrombocytopenia present in many sick premature infants.

Neonatal Tpo levels have only recently been reported. Murray et al. (8) showed that Tpo levels in nonthrombocytopenic term and preterm infants were similar to the levels found in healthy adults, about 140 pg/mL (with a wide range) and, at birth, Tpo levels in thrombocytopenic preterm infants were not significantly different from those of the nonthrombocytopenic infants. Among the 13 thrombocytopenic infants studied, 11 had either intrauterine growth retardation or mothers with pregnancy-induced hypertension. In a second study, the same group showed that Tpo levels did rise by postnatal day 5 in thrombocytopenic preterm infants, although serum concentrations were lower than in older children with thrombocytopenia secondary to marrow hypoplasia (9). Sola et al. (10) found that Tpo levels did not correlate either with gestational age in 47 nonthrombocytopenic neonates or with platelet counts in 20 thrombocytopenic infants. A better inverse correlation was found with megakaryocyte mass, as assessed on bone marrow aspirates, except for those infants with intrauterine growth retardation who had decreased megakaryocyte mass and only minimally elevated Tpo levels. In contrast, Colarizi et al. (11) found that mean Tpo concentrations did show an inverse correlation with platelet count, and neonates with sepsis had higher Tpo levels than did neonates with thrombocytopenia from other causes, although the range was very broad. The diverse results of these four small studies suggest that we do not yet understand the nuances of Tpo response in the neonate, although there is evidence that the underlying mechanism of the thrombocytopenia, the megakaryocyte mass, and conditions associated with fetal hypoxemia may all contribute to the response.

None of these studies examined the correlation of Tpo levels with the reticulated platelet count, a marker of platelet production more easily quantified and accessed than megakaryocyte mass. Because they are helpful in determining the etiology of thrombocytopenia and have been evaluated in neonates of various gestations, reticulated platelet counts may be a clinically useful measure for the evaluation of neonatal responses to endogenous or exogenous Tpo (12).

Two forms of recombinant Tpo have been tested in clinical trials: a full-length polypeptide, and a truncated molecule consisting of the amino-terminal receptor-binding domain conjugated to PEG to prolong its circulating half-life. In this issue of Pediatric Research, Sola et al. (13) report the results of administration of the truncated form of Tpo (PEG-rHuMGDF) to newborn rhesus monkeys; similar studies in adult nonhuman primates have been performed (14). As a preamble to studies in human neonates, the investigators evaluated the dose response and toxicity of PEG-rHuMGDF by monitoring changes in platelet counts and hematopoietic progenitor colony assays and, also, the pharmacokinetics in full-term, nonthrombocytopenic monkeys. In the newborn monkeys, the response to pharmacological doses of Tpo was not significantly different from that in adult primates—with good responses at 0.25 μg/kg/day. After administration of Tpo began, there was a lag of 6 d before the platelet count began to rise. No adverse effects were seen in the treated primates. However, it should be noted that PEG-rHuMGDF has been withdrawn from clinical trials because of the development of neutralizing antibodies in a number of subjects (15).

The important question for the neonatal population will be how to use Tpo most effectively: which thrombocytopenic neonates will benefit most from its administration and how to identify them. From studies to date, it appears that some neonates have inappropriately low levels of Tpo relative to the degree of thrombocytopenia. There are clues that conditions associated with fetal hypoxemia may, in part, define this group; other identifiers are yet to be determined. For example, will Tpo be useful in the thrombocytopenia associated with intrauterine viral infection? We do not know whether seriously ill newborns will respond as well to Tpo as did the healthy newborn monkeys in the present study; serious systemic illness or concomitant use of medications that affect platelet production may well modify the response. Finally, as with other growth factors, there is a lag time from the initiation of Tpo therapy to an increase in the platelet count. Therefore, although Tpo may decrease the need for platelet transfusion beyond the first week, it is unlikely to eliminate the requirement for transfusions in the first few days of life to prevent or manage serious bleeding. The preferred course would be introduction of Tpo into the nursery within the context of clinical trials to establish criteria for the most efficacious and safe use of this new and expensive growth factor.