Low or Undetectable Levels of Thrombopoietin Receptor (Tpo-R; c-mpl) Expression on Tumor Cell Lines and Primary Tumors Compared with Epo-R, ErbB2, and IGF-1R

INTRODUCTION: Recent discovery of thrombopoietin receptor (Tpo-R; c-mpl) agonists and thrombopoietin mimetics has warranted a better understanding of their effects on solid and liquid tumors. These agents bind to different components of the Tpo-R and therefore signal differently. Previously published data on Tpo-R agonists have shown antiproliferative effects on leukemia and lymphoma cells in vitro (Kalota A, Gewirtz AM. American Association for Cancer Research. 2008. Abstract 2392; Erickson-Miller CL, et al. American Association for Cancer Research. 2008. Abstract 5691). Although the expression of Tpo-R is well documented on cells of the megakaryocyte lineage, there is little quantitative information available on the expression of Tpo-R on tumors.

METHODS & RESULTS: To better define Tpo-R expression, we first performed quantitative reverse-transcriptase PCR (qRT-PCR) on 378 tumor cell lines available from the American Type Culture Collection (ATCC) or the German Collection of Microorganisms and Cell Cultures (DSMZ). Tpo-R was consistently expressed at low levels, with a mean normalized abundance of 1,447 and a mode of 621. Only 3 cell lines expressed Tpo-R mRNA below the limits of reliable quantitation (SAOS-2, SF-539, and WIDR; bone, brain, and colon tumor cell lines, respectively). In comparison, the erythropoietin receptor (Epo-R) was expressed in low-to-moderate levels (mean, 12,587; mode, 7,811) and ErbB2 was expressed at higher levels (mean, 280,190; mode, 40,828), with expression, as expected, much higher among the breast tumor cell lines. IGF-1R was also expressed at higher levels (mean, 78,977; mode, 56,624). Three cell lines had greater than 9,500 normalized abundance: HEL 92.1.7, KG-1 (2 erythroleukemia cell lines), and NCI-H510 (lung cancer cell line). To determine if these trends also occurred in patient tumor samples, microarray data were examined from 118 breast cancer, 29 non-small cell lung cancer (NSCLC), and 151 renal cell carcinoma (RCC) samples collected prior to treatment in GlaxoSmithKline clinical trials. Robust multiarray average (RMA) analysis was used to determine relative mRNA expression levels. Tpo-R mRNA levels were too low for accurate measurement in all breast cancer and RCC samples, but were detectable at low levels in 14 (48%) NSCLC samples (Table 1). In contrast, Epo-R was expressed in 75% breast cancer samples, in all NSCLC samples, and in 87% RCC samples. ErbB2 was expressed in all breast samples, 97% of the NSCLC, and in 81% RCC samples. IGF-1R was expressed in 86% breast cancer samples, all NSCLC samples, and 54% RCC samples. For breast tumors, the levels of Tpo-R message expression rank as follows: Tpo-R<Epo-R<IGF-1R<ErbB2. To determine the relationship between Tpo-R message expression and Tpo-R protein expression, Western blot analyses were performed on several of the tumor cell lines, including 2 with the highest Tpo-R mRNA expression as determined by qRT-PCR (HEL 92.1.7 and NCI-H510) and 2 with undetectable Tpo-R mRNA expression (ML-2 and NCI-H360). Western blots demonstrated that Tpo-R protein was detectable in the lysates of HEL 92.1.7 cells and normal human platelets, which were used as a positive control. However, Tpo-R protein was not detected in NCI-H510, ML-2, or NCI-H360 cells. Thus, even the high levels of Tpo-R mRNA in NCI-H510 cells did not correlate to detectable Tpo-R protein expression.

CONCLUSIONS: In summary, low or undetectable levels of Tpo-R mRNA expression were observed in tumor cell lines and in limited samples of patient tumors, compared with Epo-R, ErbB2, and IGF-1R. In the tumor cells tested, Tpo-R protein was not always detectable, even when Tpo-R mRNA was expressed.

Table 1. Number (%) of primary tumor samples with accurately detectable Tpo-R mRNA expression from RMA analysis of microarray data.