The iMyc^Eμ-1 cell line, which demonstrated the typical cytological features of mouse LBL (Fig. 1A top), was derived from a primary IgM⁺LBL (Fig. 1A bottom) that exhibited moderate plasmacytic differentiation potential in situ (not shown). FACS analysis using a panel of antibodies to B cell surface markers (Fig. 1B) showed that iMyc^Eμ-1 cells were positive for CD40, CD48, CD54, CD80/86 (B7-1/2) and CD138 (syndecan 1). Expression of class I and II MHC antigens and CD45 (B220) was detectable at low or very low levels, respectively, but CD95 (Fas) was absent. Treatment of iMyc^Eμ-1 cells with antibody to CD40 led to the induction of Fas and upregulation of CD45 and CD54, activation markers CD80/86, and CD138, indicating that CD40 signaling was functional (Additional File 1). Southern blotting of genomic DNA from the LBL from which the cell line was derived demonstrated V(D)J rearrangement at the Ig heavy-chain and κ light chain loci (Fig. 1C top). The expression of surface IgM^highIgD^low by the derivative cell line was consistent with this and indicated that the rearrangement was productive (Fig. 1C bottom). SKY analysis of metaphase chromosomes from iMyc^Eμ-1 cells (Fig. 2) uncovered four chromosomal translocations that took the form of a reciprocal T(9;11) exchange and three non-reciprocal exchanges: T(13;16), T(14;13) and T(17;6). Although it is unclear whether these translocations occurred during tumor development or establishment of the cell line 5, repeat karyotyping showed that the present iMyc^Eμ-1 line is cytogenetically stable.

The Mouse Lymphochip, a microarray of hematopoietic mouse cDNA clones, provides a powerful tool to evaluate the similarity of primary mouse B cells, B-cell tumors, and tumor-derived cell lines at the level of global gene expression 6. To compare the gene expression profile of LBL and iMyc^Eμ-1 cells, RNA was obtained from primary B cells, fresh-frozen tumors and iMyc^Eμ-1 cells. The RNA was labeled with Cy5-dUTP, and hybridized to the cDNA microarray spotted on a glass slide. An RNA control pool labeled with Cy3-dUTP was co-hybridized to the same array and used as a common denominator by which all samples were compared to one another. Further information on microarray make-up, analysis and data interpretation is available at: http://lymphochip.nih.gov/ShafferPCfactors/.

Three independent RNA samples of iMyc^Eμ-1 cells and ten primary LBL were analyzed together with a collection of normal, resting mouse B cells and T cells, mouse embryonic fibroblasts (MEF), and peritoneal plasmacytomas that arose in pristane-treated BALB/c mice. A total of 414 well-characterized array elements that clustered across these samples based on gene expression patterns (Fig. 3) demonstrated a clear distinction of lymphoid and non-lymphoid cell types (B and T cells versus MEF), lymphocyte lineages (B versus T cells), and transformation- and development-associated differences within the B-cell lineage (normal B cells versus LBL and PCT). The gene expression profiles of LBL and iMyc^Eμ-1 cells, which clustered in one tight group (Fig. 3 top, blue rectangle), exhibited a remarkable homogeneity. Compared to the plasmacytomas, LBL and iMyc^Eμ-1 cells maintained many hallmark genes in the B cell signature (e.g., those encoding CD19, CD79 and μ heavy-chain) but under-expressed numerous genes in the plasma cell signature (e.g., Sec61, Ssr4 and DNAjc3) and matrix signature (e.g., those encoding vinculin, gelsolin and integrin B1) 78. A more detailed analysis of the plasma cell signature revealed that in contrast to Xbp1 and its target genes, the iMyc^Eμ-1 cells expressed Prdm1 (Blimp1). This suggested that the cells underwent neoplastic transformation at the early stage of plasmacytic differentiation (Additional File 2).

The expression levels of Myc, as measured by the arrays, was clearly elevated in LBL and iMyc^Eμ-1, intermediate in "premalignant" B cells from tumor-free iMyc^Eμ mice, and absent, as expected, in resting lymphocytes from normal mice (Fig. 4A). Because the inserted Myc cDNA in iMyc^Eμ mice also encodes a C-terminal His₆ tag, it is possible to distinguish message and protein encoded by Myc^His and normal Myc. Allele-specific RT-PCR analysis of Myc^His and Myc mRNA demonstrated that, in common with LBL, iMyc^Eμ-1 cells expressed predominantly the transgene (Fig. 4B top, lanes 2–3). This pattern of suppression of the normal Myc gene 9 is also a feature of human B-cell lymphomas containing constitutively deregulated MYC 10. Western blotting with an anti-Myc antibody detecting both Myc^His and normal Myc proteins (Fig. 4B bottom) showed that LBL and iMyc^Eμ-1 cells over-expressed Myc at comparable levels (lanes 2–3) relative to B splenocytes from non-transgenic littermates (lane 1). To compare the levels of Myc in LBL and iMyc^Eμ-1 cells more precisely, we performed qPCR using Aktb mRNA levels as internal standard. The iMyc^Eμ-1 cells expressed nearly twice as much Myc as the LBL (Fig. 4C). The levels of Myc also correlated with the expression of genes from the proliferation cluster when the gene expression from the proliferation signature, as defined in Figure 3, was averaged for each cell type. Proliferation gene expression was low in unstimulated cells (MEF and B/T cell samples), intermediate in pre-malignant B cells from iMyc^Eμ mice, and upregulated in LBL and iMyc^Eμ-1 (Fig. 4D).

To further examine the contribution of the Myc^His transgene to the gene expression profile of LBL and iMyc^Eμ-1, we performed a statistical analysis (Student's T test) of genes differentially expressed between normal B cells versus LBL and iMyc^Eμ-1 cells. A total of 122 array elements from Figure 3 were significantly differential in their expression when B cells were compared to iMyc^Eμ-1 cells and LBL (p < 0.015, 1.5-fold minimal difference in average expression). The vast majority (97%) of these elements, many of them previously identified as proliferation-associated Myc targets http://www.myc-cancer-gene.org, behaved similarly in both the cell line and LBL (Fig. 4E). Twenty-four known Myc targets were up-regulated in the LBL and iMyc^Eμ-1 cells (Ahcy, Apex1, Cbfb, Cdk4, Ctps, Eef2, Gapd, Hdgf, Hnrpa1, Hnrpd, Idh1, Myc, Ncl, Nme1, Npm1, Pa2g4, Pcna, Pim1, Pkm2, Ppia, Sfrs2Slc7a5, Tfdp1), two were down-regulated (Btg1, Igk), and five were undetermined as to the effect of Myc on their expression (Aldh2, Hint1, Mcl1, Rheb, Slc1a4). These findings were in accordance with the nature of LBL and iMyc^Eμ-1 as Myc-driven B-cell tumors and firmly established the similarity of LBL and iMyc^Eμ-1at the level of a single gene (Fig. 4A), a gene expression signature (Fig. 4D), and globally (Fig. 3).

To further compare the gene expression profiles of LBL and iMyc^Eμ-1, and validate the cDNA microarray results with an independent method, we used cDNA macroarrays on nylon membranes to assess the expression of selected "pathway" genes in LBL, iMyc^Eμ-1 and normal B cells. Included in the analysis were RNA samples of iMyc^Eμ-1 and LBL previously analyzed on the microarray. Freshly prepared RNA from normal, MACS purified, B220⁺ splenocytes were used as control. RNA samples were labeled with ³²P-dUTP and individually hybridized to the macroarrays. Individual expression profiles were determined and compared with each other. Reproducible two-fold or higher changes in hybridization signal intensity were used as threshold for gene expression changes. Eight different macroarrays, each containing 96 genes involved in cell cycle regulation, apoptosis, cancer, signal transduction, stress and toxicity responses, and the NFκB and MAPK pathways, were used. The primary data set is depicted in Additional File 3.

The changes on the macroarrays were remarkably consistent in LBL and iMyc^Eμ-1 cells compared to B cells. This is illustrated in Figure 5A, using the apoptosis array as the example. Among a total of 768 genes present on eight different macroarrays, 121 (16%) genes were concordantly up- or down-regulated in LBL and iMyc^Eμ-1 cells relative to B cells. The NFκB array showed the highest number of changes (n = 22) among the eight different macroarrays, followed by the MAPK (n = 19), cell cycle (n = 18) and apoptosis arrays (n = 17). The stress and toxicity array contained the lowest number of changes (n = 9). Altogether, down-regulated genes (83/121, 69%) outnumbered up-regulated genes (38/121, 31%) by a factor of 2.2. The presence of some genes on two or more arrays afforded an opportunity for additional quality control. Genes of that sort exhibited the same trend on different arrays, either up or down relative to normal B cells, thus adding confidence in the results. This is illustrated in Additional File 4 using one representative array each of iMyc^Eμ-1 and normal B cells.

Among the differentially regulated genes that exhibited concordant changes in LBL and iMyc^Eμ-1 cells relative to normal B cells on both the gene micro- and macroarrays were 16 genes that were selected for further confirmation using semi-quantitative RT-PCR (Additional File 5). These genes were of particular interest to us because of possible follow-up studies on signaling pathways in iMyc^Eμ-1 cells. Eleven of the 16 genes were down regulated (Bcl2a1, Birc2, Cflar, Cdkn1b, Grb2, Irf1, Jun, Map2k1, Rb1, Ripk1, Traf5) and five genes were up regulated (Ccna2, Ccnb, Myc, Nfkb1, Odc). Figure 5B presents average quantitative changes on the macroarrays when iMyc^Eμ-1 cells were compared with normal B cells. These changes were readily confirmed by RT-PCR in all cases (Fig. 5C) except Myc, which was not included because it was confirmed in previous work (Fig. 4).

Transgenic iMyc^Eμ mice develop a high incidence of B cell and plasma cell tumors with LBL being the predominant phenotype (8). Tumor samples obtained at autopsy were fixed in formalin for later histopathology or frozen for later preparation of protein, DNA and RNA. Histological criteria used for diagnosing mouse LBL are detailed elsewhere 11. Highly enriched splenic B cells were prepared from C57BL/6 mice using CD45R (B220) microbeads and MACS separation columns (Miltenyi Biotec, Auburn, CA). All mice were maintained under Animal Study Protocol LG-028. The iMyc^Eμ-1 cell line was derived from a LBL and maintained at 37°C and 5% carbon dioxide in RPMI 1640 medium supplemented with 10% fetal calf serum, 200 mM L-glutamine, 50 μM 2-mercaptoethanol and penicillin/streptomycin (Gibco-BRL, Rockville, MD).

For cytological analysis, cytofuge specimens were stained according to May-Grünwald-Giemsa and inspected by microscopy. For detection of chromosomal aberrations, cells were analyzed by spectral karyotyping (SKY) as previously described 12. For flow cytometry, single-cell suspensions were stained and analyzed on a FACSort^® using the CELLQuest™ software (BD Pharmingen, San Diego, CA). Rat anti-mouse CD16/CD32 was used to block FcγII and FcγIII receptors. Antibodies to mouse CD45 (catalog number 553076), CD80 (553766), Fas (CD90, 554255), CD86 (553689), CD40 (553787), I-A^b (MHC class II, 553551), H-2K^b (MHC class I, 553569), CD48 (557483), CD54 (553250), CD138 (553712), IgD (553438) and IgM (53519) were purchased from BD Biosciences. For the evaluation of surface marker changes upon ligation of CD40, cells were incubated with rat anti-mouse CD40 (553787) using 3.5 μg antiboy per 5 × 10⁵ cells. For Southern blot hybridization of clonotypic V(D)J rearrangements, genomic DNA (20 μg) was digested with BamHI and EcoRI, fractionated on a 0.7 % agarose gel, transferred to a nylon membrane, and crosslinked under UV light. Following pre-hybridization (Hybrisol I, Intergen) at 42°C, the membrane was hybridized to a 1.5-kb HindIII/EcoRI fragment of Igh spanning J_H2 and Eμ or to a 1.1-kb Cκ probe, which was generated by PCR using a primer pair obtained from Dr. Michael Kuehl (NCI): 5'-GAT GCT GCA CCA ACT GTA TCC A-3' and 5'-GGG GTG ATC AGC TCT CAG CTT-3'. Probes were labeled with [³²P]-CTP using a random priming kit.

For semi-quantitative determination of Myc and Myc^His mRNA, total RNA was isolated using TRIzol (Sigma, St. Louis, MO, USA). Double stranded cDNA was synthesized from 1 μg of total RNA, using the AMV Reverse Transcriptase kit (Roche, Indianapolis, IN). A common 5' primer for both Myc^His and Myc (5'-TCT CCA CTC ACC AGC ACA AC-3') was combined with a specific 3' primer for Myc^His (5'-CCT CGA GTT AGG TCA GTT TA-3') and Myc (5'-ATG GTG ATG GTG ATG ATG AC-3') to distinguish the two messages. Thermal cycling conditions were as follows: 95°C for 5 min followed by 20 cycles of amplification at 57°C, 72°C and 95°C, each for 1 min. PCR amplification of Aktb cDNA was performed as control using the following primer pair: 5'-GCA TTG TTA CCA ACT GGG AC-3' and 5'-AGG CAG CTC ATA GCT CTT CT-3'. PCR products were analyzed by electrophoresis in 1% agarose gel and visualized by staining with ethidium bromide.

For quantitative Taqman RT-PCR of Myc (Myc plus Myc^His), total RNA was isolated from cells using TRIzol Reagent (Invitrogen). Serial dilutions of input RNA (100 ng - 1.56 ng) were analyzed in triplicates using the ABI PRISM 7900HT sequence detector system, primers, probes, and the Taqman One-Step RT-PCR Master Mix Reagents kit, all purchased from Applied Biosystems. The reaction mixture was held at 48°C for 30 min for reverse transcription of RNA into cDNA. This was followed by incubation at 95°C for 10 min to activate the Taq polymerase. PCR amplification of cDNA was performed for 40 cycles using the following cycling conditions: denaturing for 15 s at 95°C and annealing and extending for 1 min at 60°C. All samples were tested in triplicates, and average values were used for quantification. Analysis was performed using SDS v2.1 software (Applied Biosystems) according to the manufacturer's instruction. Aktb was used as internal reference gene. The comparative CT method (ΔΔCT) was used for quantification of gene expression.

cDNA made from total RNA (50 μg) from each tumor, primary cell sample, or iMyc^Eμ-1 cells was labeled with cyanine 5-conjugated dUTP (Cy5). cDNA made from pooled mouse cell line RNA (50 μg) was labeled with cyanine 3-conjugated dUTP (Cy3) and used as reference. Microarray hybridizations were performed on Mouse Lymphochip microarrays 7. After washing, the slides were scanned using an Axon GenePix 4.0 scanner (Axon Instruments Inc., Union City, CA). After normalization, those elements that failed to meet confidence criteria based on signal intensity and spot quality were excluded from analysis. In addition, data were discarded for any gene for which measurements were missing on >30% of the arrays or were not sequence-verified. The Cy5:Cy3 intensity ratios of the remaining spots were log₂ transformed. To compare normal samples, hierarchical cluster analysis was performed using the Gene Cluster and Treeview programs 8.

The relative mRNA expression of genes involved in regulation of apoptosis, cell cycle progression, NFkB signaling, and cellular stress and toxicity responses was analyzed with GEArray (SuperArray Inc., Bethesda, MD) according to the manufacturer's protocol. Cells were treated for 24 hrs with 0.4 mM and 1 mM CDDO-Im, respectively, followed by preparation of total RNA using TriReagent (Sigma). Five μg from each sample were reverse transcribed into ³²P-labeled cDNA using MMLV reverse transcriptase (Promega, Madison, WI) and ³²P-dCTP (NEN, Boston, MA). The resulting cDNA probes were hybridized to gene-specific cDNA fragments spotted in quadruplicates on the GEArray membranes. After stringent washing of the arrays, the signal of the hybridized spots was measured with a STORM PhosphorImager (Molecular Dynamics, Sunnyvale, CA) and normalized to the signal of the housekeeping gene Gapd. Array results on six CDDO-Im inducible genes were validated using semi-quantitative RT-PCR.

For semi-quantitative determination of mRNA levels, total RNA was isolated and double stranded cDNA was synthesized as described above for Myc. Information on PCR primers and thermal cycling conditions is available in Additional File 6. PCR products were analyzed by electrophoresis in 1% agarose gel and visualized by staining with ethidium bromide.