Transgene mRNA expression and Cre-loxP–mediated DNA recombination. (A) Expression of human JAK2-V617F mRNA and endogenous Jak2 in tissues of 2 VavCre;FF1 and 3 MxCre;FF1 mice. BM, bone marrow; SPL, spleen; THY, thymus; LIV, liver; KID, kidney; INT, intestine; LUN, lung; HEA, heart; BRA, brain; TES, testis; OVA, ovary. Real-time PCR was performed with primers specific for the activated configuration of the human JAK2 transgene (top panel) or specific for mouse Jak2 (bottom panel). The numbers represent relative expression values calculated by the ΔΔCT method after normalization to the mRNA of mouse Gusb and arbitrarily choosing one bone marrow sample from a VavCre;FF1 mouse as the calibrator. Separate calculations were carried out for human and mouse Jak2. (B) Expression of human JAK2-V617F (left panel), mouse Jak2 (middle panel), and ratio between human JAK2-V617F and mouse Jak2 (right panel) in total bone marrow. Expression was measured in control mice (n = 7), VavCre;FF1 (n = 7, age 20-30 weeks), MxCre;FF1 (n = 8, 15-20 weeks after 3 × pIpC injection), or in mice transplanted with retrovirally transduced bone marrow expressing JAK2-V617F (pMSCV-hV617F; n = 6, 20 weeks after transplantation; values corrected for the percentage of transduced cells based on green fluorescent protein (GFP) expression are shown, GFP-positive cells = 35% ± 24%). Expression in MxCre;FF1 was significantly higher than in VavCre;FF1 samples (Mann-Whitney test, P = .0012). The highest levels of JAK2-V617F were observed with retroviral transduction (P = .0016 vs MxCre;FF1). A slight decrease in expression of mouse Jak2 was noted in MxCre;FF1 compared with controls (middle panel, P = .04), whereas no significant difference was found between VavCre;FF1 and MxCre;FF1. The ratios between human JAK2-V617F and mouse Jak2 in bone marrow were calculated from the absolute expression values of human JAK2 and mouse Jak2 that were determined by comparison with standard curves set up from purified plasmids containing human JAK2 or mouse Jak2. Significant differences between VavCre;FF1 and MxCre;FF1 (P = .0003) and between MxCre;FF1 and pMSCV-hV617F (P = .0007) were noted. All significant differences (P < .05) are marked by asterisks. (C) Assessment of Cre-mediated recombination by Southern blot analysis and copy number determination by real-time PCR. DNA from bone marrow and spleen cells was digested with XbaI and the Southern blots of DNA fragments separated by electrophoresis were visualized with a ³²P-labeled human JAK2 cDNA probe (thick solid line). The scheme below shows the position of XbaI restriction sites and of the probe. The expected fragments sizes are 800 bp for the native configuration and 3′500 bp for the flipped configuration. Vav, bone marrow or spleen from VavCre;FF1 mice; Mx, bone marrow or spleen from MxCre;FF1 mice. The percentage of flipped transgenes (% flipped) was determined by quantification of the flipped and native bands on a phosphorimager. Transgene (Tg) copies were determined by real-time PCR, and the numbers of active copies were calculated by multiplying with the percentage of flipped alleles. (D) Correlation of human/mouse JAK2 mRNA ratio with the number of flipped alleles (left) and correlation of hemoglobin with the human/mouse mRNA ratio (right). MxCre;FF1 mice received 1 ×, 3 ×, or 6 × injections of pIpC and were analyzed at 12 weeks. The number of flipped alleles was determined by real-time PCR, which yielded similar results as Southern blot analysis. VavCre;FF1 mice (age 12 weeks) were included for comparison.