Letters to the Editor 145 human and mouse P-gp efflux capacity for imatinib, it is clear 2 Rumpold H, Wolf AM, Gruenewald K, Gastl G, Gunsilius E, Wolf D. that imatinib is an in vivo substrate for mouse P-gp.7,8 RNAi-mediated knockdown of P-glycoprotein using a transposon- We believe the differences in imatinib resistance seen when based vector system durably restores imatinib sensitivity in imatinib- comparing the K562 data by Rumpold et al. with our data in resistant CML cell lines. Exp Hematol 2005; 33: 767–775. 3 Zong Y, Zhou S, Sorrentino BP. Loss of P-glycoprotein expression in P-gp null mice is best explained by the possibility that CML stem hematopoietic stem cells does not improve responses to imatinib in cells express other ABC transporters that may provide a a murine model of chronic myelogenous leukemia. Leukemia 2005; redundant efflux mechanism for imatinib. The lack of expression 19: 1590–1596. of these compensatory transporters in K562 cells could therefore 4 Ferrao PT, Frost MJ, Siah SP, Ashman LK. Overexpression of explain the sensitizing effect of decreased expression of P-gp in P-glycoprotein in K562 cells does not confer resistance to the K562 cells. For instance, we have shown that K562 cells do not growth inhibitory effects of imatinib (STI571) in vitro. Blood 2003; 102: 4499–4503. express ABCG2 (unpublished data) while primary hematopoietic 5 Lange T, Gunther C, Kohler T, Krahl R, Musiol S, Leiblein S et al. 9 stem cells express significant levels. Therefore, we stand by our High levels of BAX, low levels of MRP-1, and high platelets are conclusion that the use of selective P-gp inhibitors to sensitize independent predictors of response to imatinib in myeloid blast CML stem cells to imatinib is not likely to be clinically useful. crisis of CML. Blood 2003; 101: 2152–2155. 6 Crossman LC, Druker BJ, Deininger MW, Pirmohamed M, Wang L, Y Zong1,2, S Zhou1 and BP Sorrentino1,2 Clark RE. hOCT 1 and resistance to imatinib. Blood 2005; 106: 1Division of Experimental Hematology, Department of 1133–1134. Hematology – Oncology, St Jude Children’s Research 7 Dai H, Marbach P, Lemaire M, Hayes M, Elmquist WF. Distribution Hospital, Memphis, TN, USA and of STI-571 to the brain is limited by P-glycoprotein-mediated efflux. 2The Interdisciplinary Program, College of Graduate Health J Pharmacol Exp Ther 2003; 304: 1085–1092. Sciences, the University of Tennessee Health Science Center, 8 Breedveld P, Pluim D, Cipriani G, Wielinga P, van Tellingen O, Memphis, TN, USA Schinkel AH et al. The effect of Bcrp1 (Abcg2) on the in vivo pharmacokinetics and brain penetration of imatinib mesylate E-mail: [email protected] (Gleevec): implications for the use of breast cancer resistance and P-glycoprotein inhibitors to enable the brain penetration References of imatinib in patients. Cancer Res 2005; 65: 2577–2582. 9 Zhou S, Schuetz JD, Bunting KD, Colapietro AM, Sampath J, Morris 1 Rumpold H, Wolf AM, Wolf D. Letter regarding the article by Zong JJ et al. The ABC transporter Bcrp1/ABCG2 is expressed in a wide et al., ‘P-glycoprotein does not confer imatinib resistance’. variety of stem cells and is a molecular determinant of the side- Leukemia. 2006; 20: 144. population phenotype. Nat Med 2001; 7: 1028–1034. hRap1B-retro: a novel human processed Rap1B blurs the picture?

Leukemia (2006) 20, 145–146. doi:10.1038/sj.leu.2404040; database. Interestingly, three of the cDNAs were cloned from a published online 24 November 2005 Ramos Burkitt’s lymphoma cell line (GIs: 50475494, 50474682, 50470789), one was isolated from a neuroblastoma cell line (GI: In the September issue of Leukemia, Gyan et al.1 report on the 50498405), and one from a placental library (GI: 50492447). expression of a mutant Rap1B cDNA in myelodysplastic These sequences map exactly onto the genomic DNA sequence syndromes (MDS). Specifically, they detect a cDNA with point on 5 (Figure 1, Supplementary Figure 1), revealing mutations, as compared to the wild-type Rap1B gene, that cause a structure of the hRap1B-retro gene with two exons (Figure 1, two amino-acid substitutions (G12R and K42E). Unfortunately, Supplementary Figure 2). As genomic sequences of processed the authors could not perform an analysis of genomic DNA, are expected to lack intronic sequences, we set out to which would directly reveal the relation between the Rap1B analyse this observed peculiarity closely. We found that the gene and the mutant cDNA. Although they point out correctly intron was created via LINE-1-mediated insertion of the SVA that the encoded protein most likely constitutes an activated element into a preferred target site sequence (TTTT/AA)4 present form of Rap1B protein, we hereby present strong evidence that in hRap1B mRNA at 1304 (GI: 58219793). Since this intronic the detected cDNA does not originate from mutation of the region is missing from the sequences of the five deposited in Rap1B gene on . database cDNAs, it is evident that they represent transcribed and An alternative explanation can be offered invoking the spliced sequences rather than contaminating genomic DNA. existence of an expressed processed gene, which we have In conclusion, we think that it is highly likely that the mutant named hRap1B-retro, located on chromosome 5 q13.3 (GI: Rap1B cDNA reported by the authors is actually derived from 51465008). As it is the case for all processed genes,2 this expression of the hRap1B-retro gene in the MDS cells. Whether retrogene was generated by a LINE-1 retrotransposon3 from the this may have contributed to the development of MDS, and mRNA of Rap1B gene. The nucleotide sequence of the full- possibly other malignancies (i.e. Burkitt’s lymphoma), will length coding region of hRap1B-retro differs only at three require a more extensive analysis of the encoded gene product. positions from that of the mother Rap1B gene (Figure 1). Two of For example, it should be determined whether the protein is these changes are identical to the reported mutations (G12R and indeed more GTP-bound. We would like to remark that we also K42E; GGA-CGA and AAG-GAG, respectively), with the found processed genes in the mouse genome (manuscript third mutation being a silent one (G77G; GGA-GGG). The in preparation; Marlena Duchniewicz, Taisa Paluch, Fried latter corresponds to a region in the cDNA that was not analysed Zwartkruis and Tomasz Zemojtel). It should be noted that the by the authors. existence of a functional processed gene belonging to Ras family Direct evidence for expression of hRap1B-retro comes from has been reported previously. Intriguingly, the human ERas analysis of full-length expressed cDNAs deposited in the NCBI encodes an activated (i.e. GTP-bound) version of Ras and was

Leukemia Letters to the Editor 146

Figure 1 The hRap1B-retro gene is localized on chromosome 5 q13.3. Upper part: summary of differences in the sequence alignment between hRap1B-retro and hRap1B cDNAs and the corresponding genomic regions on the 5th and 12th . In addition to the three nucleotide substitutions in the open reading frame (ORF), seven more differences between both genes are present in the untranslated regions (UTRs). The coordinates above the sequences are relative to the start codon ATG in hRap1B cDNA (GI: 58219793). hRap1B-r_Ram1-3: full-length cDNAs of hRap1B-retro from Ramos cell line (GIs: 50475494, 50474682, 50470789); hRap1B-r_Neur: full-length cDNA of hRap1B-retro from neuroblastoma (GI: 50498405); hRap1B-r_Plac: full-length cDNA of hRap1B-retro from placenta (GI: 50492447). 5chr: genomic DNA sequence from chromosome 5 (GI: 51465008: 26064529–26060269), 12chr: genomic DNA sequence from chromosome 12 (GI: 29803948: 31147958– 31197631). Lower part: exonic structure of the hRap1B-retro and hRap1B mother genes mapped onto their cDNAs. Numbering stands for exon numbers.

shown to transform NIH 3T3 cells oncogenically.2 In this light, a 4Department of Physiological Chemistry and the Centre for better expression profiling of the hRap1B-retro gene identified Biomedical Genetics, University Medical Centre Utrecht, by us will help to validate whether it defines a novel drug target. CG Utrecht, The Netherlands E-mail: [email protected]

Acknowledgements

TZ thanks Professor Martin Vingron for the helpful discussions and References support. TZ acknowledges support received from the BioSapiens Network of Excellence, funded by the European Commission 1 Gyan E, Frew M, Bowen D, Beldjord C, Preudhomme C, Lacombe C within its FP6 Programme, under the thematic area ‘Life sciences, et al. Mutation in RAP1 is a rare event in myelodysplastic syndromes. Leukemia 2005; 19: 1678–1680. genomics and biotechnology for health’, contract number LHSG- 2 Takahashi K, Mitsui K, Yamanaka S. Role of ERas in promoting CT-2003-503265. tumour-like properties in mouse embryonic stem cells. Nature 1,2 2,3 2 2003; 423: 541–545. T Zemojtel , T Penzkofer , M Duchniewicz and 3 Penzkofer T, Dandekar T, Zemojtel T. L1Base: from functional FJT Zwartkruis4 1 annotation to prediction of active LINE-1 elements. Nucleic Acids Department of Computational Molecular Biology, Max Res 2005; 33: D498–D500. Planck Institute for Molecular Genetics, Berlin, Germany; 4 Cost GJ, Boeke JD. Targeting of human retrotransposon integration 2 In Silico Miners, Berlin, Germany; is directed by the specificity of the L1 endonuclease for regions of 3Biozentrum, Am Hubland, Wuerzburg, Germany and unusual DNA structure. Biochemistry 1998; 37: 18081–18093.

Supplementary Information accompanies the paper on the Leukemia website (http://www.nature.com/leu)

Reply to ‘hRap1B-retro: a novel human processed Rap1B gene blurs the picture?’ by Zemojtel et al

Leukemia (2006) 20, 146–147. doi:10.1038/sj.leu.2404042; published in the September issue of Leukemia,2 we initially used published online 24 November 2005 previously published oligonucleotides3 to analyze rap1b tran- script in bone marrow samples from 27 patients with myelo- We agree with the comments made by Zemojtel et al.1 and we dysplastic syndromes (MDS) and two secondary acute myeloid thank them very much for the proposed hypothesis. In our work leukemias (sAML). In one patient with RARS and abnormal

Leukemia