Thrombopoietin Levels in Patients with Cirrhosis before and after Orthotopic Liver Transplantation

Thrombocytopenia in patients with cirrhosis has historically been attributed to hypersplenism. Radiolabeled platelet studies have clearly demonstrated splenic sequestration in cirrhosis and have shown that up to 90% of platelets can localize to the spleen [1-3]. However, portal decompression procedures have failed to consistently improve thrombocytopenia [4-7] and thrombocytopenia can persist after splenectomy [8]. Therefore, other factors must be involved. Thrombopoietin is a potent stimulator of megakaryocyte growth and platelet production [9]. Synthesis of thrombopoietin probably occurs in the liver [10]. In this study, we measured plasma thrombopoietin levels in cirrhotic patients who had thrombocytopenia and evaluated changes in these levels throughout orthotopic liver transplantation. Levels of messenger RNA (mRNA) in thrombopoietin in liver samples taken from patients with cirrhosis and controls were compared. Our hypothesis was that production of thrombopoietin is altered in patients with cirrhosis and that impaired transcription of thrombopoietin mRNA may contribute to this alteration. Methods Two separate convenience samples taken from patients with cirrhosis (17 patients who had and 27 patients who did not have orthotopic liver transplantation) were recruited from the gastroenterology services at the University of California, San Francisco, Medical Center. All 44 patients had persistent or stable thrombocytopenia for more than 2 weeks (platelet count < 120 000 cells/microL). Patients with acute intercurrent illnesses were excluded. Patients were selected for transplantation on the basis of standard criteria (severity of disease and organ availability). These patients received an ABO-matched cadaver liver and standard post-transplantation immunosuppressive drugs, including corticosteroids, azathioprine, and cyclosporine. Approval was obtained from the Committee on Human Research at the University of California, San Francisco: all patients provided informed consent. The enrollment period was from June 1995 through April 1996. The 27 patients with cirrhosis who did not have transplantation provided one random plasma sample. The 17 patients who had orthotopic liver transplantations had plasma samples taken before transplantation; within 4 to 12 hours after transplantation; and every Monday, Wednesday, and Friday thereafter until discharge or until the platelet count was normal. This schedule was adopted from previous studies of patients who had bone marrow transplantation [11]. Platelet counts were obtained daily for all 17 patients until discharge, as per standard protocol for orthotopic liver transplantation. The samples were frozen at 80C before being measured. Two enzyme-linked immunosorbent assays (ELISAs) were used during the study. The initial assay used a chimeric molecule that consisted of c-mpl extracellular domain fused to the Fc portion of human immunoglobulin for capture and a rabbit polyclonal antibody to thrombopoietin for detection. The second was a murine antithrombopoietin monoclonal antibody sandwich-type assay. The detection limits of the assays were 160 and 40 pg/mL, respectively. The second ELISA was available for the final 13 patients with cirrhosis and 10 patients who had liver transplantation. Cirrhotic liver samples were obtained from explanted livers during surgery (n = 9). Samples of livers from controls were obtained during partial hepatectomy for localized metastases of colon cancer (n = 5) or primary liver cancer (n = 2; 1 patient had leiomyosarcoma, and 1 had hepatoma) and during orthotopic liver transplantation for hereditary hyperoxaluria (n = 1). Liver tissue was obtained from areas of normal-appearing parenchyma. All patients had normal platelet counts, prothrombin times, and aminotransferase levels. Samples were snap-frozen in liquid nitrogen and stored at 80C. Total RNA was isolated from cryopreserved liver tissue by using the ULTRASPEC RNA Isolation System (Biotecx Lab, Houston, Texas) according to the manufacturer's instructions. Reverse-transcription polymerase chain reaction (PCR) was done using 100 ng of total RNA, a thrombopoietin-specific reverse primer, and the Promega Access RT-PCR system (Promega, Madison, Wisconsin). The Perkin Elmer 7700 Sequence Detector System (Perkin Elmer, Foster City, Caliornia) was used for DNA amplification as described elsewhere [12]. The thrombopoietin-specific forward primer was 5GCCAAGATTCCTGGTCTGCTGAAC3, and the reverse primer was 5GCTGATGTCGGCAGTGTCTGAGAA3. A fluorogenic probe specific for thrombopoietin (5FAM-TGCCTGGACCAAATCCCGGATACCTGAA3) was added to the PCR amplification. Repetitive cycles resulted in increasing release of fluorescence. System software used fluorescence intensity to determine mRNA levels in thrombopoietin and in glyceraldehyde-3-phosphate dehydrogenase (a housekeeping gene). Funding provided by the National Institutes of Health and Genentech. Inc., had no influence on design of the study, conduct of the study, or reporting of the findings. Results Selected patient characteristics are shown in the (Table 1). All patients had moderate-to-severe liver dysfunction. Thrombocytopenia was profound in most patients; 21 patients had platelet counts less than 50 000/L, and 6 patients had counts less than 30 000/L. All 17 patients had successful transplantation without serious complications. Liver biopsies were done approximately 8 days (range, 6 to 16 days) after transplantation. Results showed no evidence of acute rejection (11 patients), mild rejection (4 patients), or moderate rejection (2 patients). Preservation injury was mild to moderate in all 17 patients, and all 17 had improved liver function after transplantation (data not shown). Table 1. Patient Characteristics Baseline levels of thrombopoietin were undetectable in 39 of 43 patients with cirrhosis. The mean detectable level of thrombopoietin was 130 pg/mL in 5 patients, and all levels were less than 200 pg/mL. In patients who had orthotopic liver transplantation, the mean interval to first detection of thrombopoietin was about 3 days and peak levels occurred about 8 days after transplantation. Peak concentrations of thrombopoietin varied considerably (46 to 2541 pg/mL), with a mean peak value of 492 pg/mL (median, 339 pg/mL). Platelet counts increased during or immediately after peak concentrations of thrombopoietin. Thrombocytopenia completely resolved in all 17 patients within 23 days after transplantation (10 patients before discharge, and 7 patients after discharge). As platelet counts reached normal levels, most patients (8 of 10) developed low or undetectable levels of thrombopoietin. A representative graph of thrombopoietin levels and platelet counts in 1 patient who had orthotopic liver transplantation is shown in the (Figure 1). Figure 1. Serial thrombopoietin levels and platelet counts before and after orthotopic liver transplantation. The ratio of thrombopoietin mRNA transcript levels to glyceraldehyde-3-phosphate dehydrogenase mRNA transcript levels was 0.264 in cirrhotic liver samples and 0.357 in control liver samples. Results are expressed as ratios to control for the number of liver cells. Thrombopoietin mRNA transcript levels were 26% lower in cirrhotic liver samples than in control liver samples (unpaired t-test, P = 0.0103 [95% CI, 0.025 to 0.161]). Discussion The mechanism (or mechanisms) responsible for homeostasis of thrombopoietin has not been fully elucidated. One model proposes that regulation occurs solely through binding of thrombopoietin to receptors on platelets [13, 14]. Receptors have the ability to bind, internalize, and degrade thrombopoietin. When platelet counts are normal (>140 000/microL), most thrombopoietin is bound to receptors and plasma levels are low. When platelet counts are low (<140 000 cells/microL), few receptors are available for binding with thrombopoietin and plasma levels increase. Elevated levels of thrombopoietin then stimulate megakaryopoiesis and platelet production. This model establishes an inverse relation between thrombopoietin levels and platelet counts. We previously [11] reported undetectable levels of thrombopoietin (<160 pg/mL) in plasma in 88 of 89 healthy persons and markedly increased plasma thrombopoietin levels in 58 of 61 patients with cancer and thrombocytopenia (mean thrombopoietin level, 1095 pg/mL); these findings support an inverse relation. On the basis of this model, we expected elevated levels of thrombopoietin in cirrhotic patients who had thrombocytopenia. On the contrary, we found low or undetectable levels, which suggests that thrombopoietin levels and platelet counts are not inversely related in patients with cirrhosis. Several explanations may account for this. One hypothesis is that synthesis of thrombopoietin is impaired in cirrhotic livers. Defective production of thrombopoietin results in lower serum levels, decreased megakaryopoiesis, and lower platelet counts. Another explanation is that hypersplenism causes thrombocytopenia and production of thrombopoietin is normal. As a result, thrombopoietin is bound to platelets and both are sequestered in the spleen. The persistent thrombocytopenia found in patients with cirrhosis after splenectomy or portal decompression procedures argues against this explanation. A third possibility is that thrombopoietin levels and platelet counts are low because of rapid destruction of platelets. Low levels of thrombopoietin have been described in persons with idiopathic thrombocytopenic purpura [15]. The low levels probably result from binding thrombopoietin to platelets and subsequent rapid destruction of platelets (and thrombopoietin) in the spleen. Rapid destruction of platelets is unlikely in patients with cirrhosis because nuclear medicine studies of such patients [1] have demonstrated normal or only slightly decreased platelet survival. We did measure thrombopoietin levels in five patients who had cirrhosis and normal platelet c