Functional Validation of Genes Implicated in Lymphomagenesis

Tumor Production and Analysis

Generation of Myc- and MycER-induced murine B-lymphomas has been described earlier.^16-18 Syngeneic C57BL6/J mice (NCI, Frederick, MD) were used as hosts for tumor production. Lymphoma cells (5–10 × 10⁶) were injected subcutaneously into the flanks, and tumor growth was monitored over the course of several weeks. For 4-OHT treatment, the hormone powder (H6278, Sigma, St. Louis, MO) was dispersed in corn oil (Sigma) at the concentration of 10 mg/mL. The administration of 4-OHT was performed daily via intraperitoneal injections (1 mg per mouse). For the selection assay, tumors were harvested and dissected approximately 3 weeks after implantation. Neoplastic cells were used for flow cytometric analyses or subsequent passages in vivo.

Microarray Analysis

Total RNAs from MycER tumors were used for cDNA synthesis. Labeled probes were hybridized to the U74v2 gene chip (Affymetrix, Inc. Santa Clara, CA) using the PENN Microarray Facility's standard protocol (http://www.med.upenn.edu/microarr/Data%20Analysis/Affymetrix/methods.htm). Affymetrix MAS 5 probeset signals and presence/absence flags were calculated. The LPE (Local Pooled Error) test for differential expression as implemented in S+ArrayAnalyzer v 1.1 (Insightful Corporation, Seattle, WA) was applied with 1% Bonferonni multiple testing correction to median IQR normalized MAS 5 signal values. The resulting list of 686 genes was imported into GeneSpring v 6.1 (Silicon Genetics, San Carlos, CA), filtered for presence (per Affymetrix MAS5 analysis) in 2 of 2 samples in one or more conditions (Myc^ON or Myc^OFF), and then filtered for fold change.

Generation of Recombinant Retroviruses

The murine IL10Rα-encoding retrovirus has been described earlier.¹⁶ To generate an IL10Rα mutant lacking the intracellular domain, the PCRII-mIL10Rα plasmid¹⁶ was cut with BamH I and Xma I, and the truncated IL10R insert was subcloned into the pSL1180 phagemid vector (Amersham Pharmacia, Piscataway, NJ). In this vector, the polylinker segment between Xma I and Sal I site was replaced with the double-stranded oligonucleotide (CCGGGTTAGAATTACCTCAG) containing a single T-to-G substitution shown in bold. This substitution resulted in a stop-codon (underlined) positioned in-frame with the IL10R coding sequence. The IL10Rstop sequence was cut out with Bgl II and Sal I and inserted into the bicistronic GFP-expressing retroviral vector pMIGRI¹⁹ linearized with Bgl II and Xho I. The full length murine mcl-1 cDNA was purchased from the American Type Culture Collection (Manassas, VA, catalog # MGC-13884, IMAGE ID: 4190996). The cDNA was released from the pCMV-SPORT6 vector using Not I and EcoR V restriction enzymes, cloned into pSL1180, and then recloned, using Bgl II and Hpa I restriction sites, into MIGR1.

Cell Propagation, Transfection, and Infection

Myc3 B-lymphoma cells were cultured on monolayers of gamma-irradiated S17 cells as described previously.¹⁸ BOSC and GP293 packaging cells were cultured in DMEM with 10% FBS. For in vitro infections, either of these cell lines was transfected with retroviral clones using Lipofectamine 2000 (Invitrogen, Carlsbad, CA). Conditioned media were harvested and added to Myc3 cells for 2 to12 h. Three days post infection, the percentages of GFP-positive cells in infected cultures were determined using flow cytometry.

Flow Cytometric Analyses

To detect expression of mIL10Rα and the mIL10Rstop truncated protein, transiently transfected packaging cells were stained on ice for 45 min with a biotinconjugated anti-mouse IL10R antibody (559913, Pharmingen, San Diego, CA), followed by staining with APC-conjugated streptavidin (SA1005, Caltag Laboratories, Burlingame, CA). All samples were analyzed using a FACScalibur flow cytometer (Becton Dickinson, Mountain View, CA) with the CELLQuest software (Becton Dickinson).

Improved Survival of Neoplastic Cells following Myc Inactivation

We have recently established a new mouse model for B-cell lymphoma based on infection of p53-null bone marrow cells with retroviruses encoding either constitutively active Myc¹⁷ or a fusion between Myc and the mutant form of the estrogen receptor (MycER),²⁰ requiring the presence of the synthetic estrogen 4-hydroxytamoxifen (4-OHT) for activity.¹⁶ The use of conditionally active Myc allowed us to address the role of Myc in tumor maintenance (Fig. 1). To this end, cells obtained from primary tumors were injected subcutaneously into flanks of syngeneic mice, and administration of the estrogen was maintained until palpable neoplasms have formed. Animals were then randomized into three groups. Mice in the first group continued to receive 4-OHT and rapidly developed large tumors (Fig. 2, tumors 1 and 2). In the second group, 4-OHT treatment was permanently discontinued (tumors 3 and 4). In the third group, 4-OHT treatment was suspended for two weeks and resumed on day 21 (tumors 5 and 6). We observed that in animals deprived of 4-OHT, tumor growth decreased dramatically, but the neoplasms did not undergo overt regression. Instead tumors remained stable for approximately two weeks. If 4-OHT was re-administered, extremely rapid neoplastic growth ensued (tumors 5 and 6), attesting to the viability of cells with inactivated Myc. Moreover, even without estrogen re-administration, MycER tumors began to slowly regrow after only two weeks of dormancy (Fig. 2 and Ref. 16).

To correlate tumor dormancy with neoplastic cell survival, hematoxylin/eosin and TUNEL stainings of formalin-fixed tumor sections were performed. We observed that while the rate of cell proliferation decreased upon inactivation of Myc (data not shown), so did the rate of cell death. Interestingly, tumors with reactivated Myc exhibited yet higher rates of apoptosis, as judged by the percentage of TUNEL-positive cells (Fig. 2). Conversely, relapsing tumors with inactivated MycER had modest rates of cell proliferation but almost no apoptosis.

Transcriptional Signatures of Myc^ON and Myc^OFF Tumors

In order to correlate Myc-induced changes in cell proliferation and survival to alterations in gene expression, we performed a series of microarray experiments. In these experiments transcriptional signatures of tumors with active and inactive MycER were compared. Differentially expressed genes were identified as described in Materials and Methods. Forty-four genes were upregulated more than fourfold in Myc^ON cells and 68 genes in Myc^OFF cells (Tables 1 and 2).

The functions of many Myc-upregulated and Myc-downregulated genes were consistent with the biological findings. Several genes implicated in cell proliferation maintained high transcription levels only in the presence of functional MycER, for example, cyclins B1 and B2, cdc2A, and DP1 (Table 1). On the other hand, many of the genes upregulated in the absence of Myc were known to contribute to B-cell survival and expansion, in particular, in response to antigen or cytokine stimulation. This list included components of B-cell receptor signaling pathways (receptor type protein tyrosine phosphatase C and B-lymphoid kinase),²¹ NF-κB–activators (lymphotoxin B),²² direct apoptosis inhibitors (Mcl1),²³ and cytokine receptors (interleukin 10 receptor)²⁴ (Table 2 and data not shown). While their expression correlated with neoplastic B-cell survival and eventual relapse, some of them were incompletely characterized or even reported to suppress tumorigenesis. For example, while IL10 has been known to stimulate proliferation of normal B cells in vitro,²⁵ its in vivo effects on both normal and neoplastic B cells are not well understood. In addition, anti-angiogenic, and thus anti-neoplastic, effects of IL10 have been reported.^26-28 Thus, we were interested in developing an unbiased approach capable of revealing both tumor-promoting and tumor-inhibiting properties of IL10R and other Myc target proteins.

An in Vivo Selection Assay for Cells Overexpressing Myc Target Genes

The following approach was devised in order to assign functional significance to genes implicated in B-lymphomagenesis, such as IL10R. Its full-length cDNA was inserted into a retroviral vector MIGR1, co-expressing through the IRES element green fluorescent protein (GFP).¹⁹ Retroviral constructs were transfected into packaging cells, and viral supernatants were harvested 48 h later. Then Myc-overexpressing B-lymphoma cells (Myc3¹⁸) were infected in vitro with either MIGR1/IL10R or just MIGR1 (empty vector). By varying the amount of viral supernatants, the infection rate of approximately 10% was achieved, as judged by flow cytometry for GFP. Partially but stably infected cultures were used for tumor production in syngeneic animals. The resulting tumors were analyzed using flow cytometry and percentages of GFP-positive cells were determined again. By enumerating GFP-positive cells before and after passaging in vivo, both positive and negative effects of the gene of interest on B-lymphomagenesis could be revealed (Fig. 3).

Cells Overexpressing Interleukin 10 Receptor or Its Presumptive Downstream Target Mcl1 Are Selected during B-Lymphomagenesis

The approach described above was applied to determine whether levels of IL10R are rate-limiting during B-lymphomagenesis. IL10Rα full-length cDNA is described in Materials and Methods. As a negative control, we also generated and utilized its truncated version (IL10Rstop) retaining the extracellular and transmembrane domains but lacking the cytoplasmic tail (see Materials and Methods). Transfection of both constructs into the packaging cells resulted in expression of polypeptides of expected lengths (immunoblotting data not shown) located on the cell surface (flow cytometry data in Fig. 4A). The IL10R and IL10Rstop retroviruses were used to infect Myc3 cultures as described in the previous section.

When transduced cells were used for tumor production in vivo, a steady increase in the percentages of IL10R-overexpressing, GFP-positive cells was apparent: after only three passages in vivo such cells accounted for over 90% of total neoplastic cells, up from 10% observed after infection in vitro (Fig. 4B). No increase was observed with GFP- or IL10Rstop-overproducing cells. Additionally, no increase was observed when the experiment was performed in IL10-deficient mice,²⁹ suggesting that the positive selection for IL10R-overexpressing cells depends on the availability of its cognate ligand.

IL-10 is known to signal predominantly through the STAT3 pathway (reviewed in Ref. 30), which activates expression of several anti-apoptotic genes including Bcl-2, Bcl-xl, survivin, and Mcl1.³¹ Consistent with this, Mcl1 was upregulated in Myc^OFF tumors, albeit slightly less than fourfold (data not included in Table 2). We thus asked whether positive selection for Mcl1-overexpressing cells would also be in place. We generated a murine Mcl1/GFP-encoding retrovirus and repeated the basic experiment. As evidenced by data in Figure 4C, after only two passages in vivo, Mcl1-transduced cells accounted for almost 100% of the total cell population. Thus, it is possible that the contribution of the IL10 pathway to B-lymphomagenesis is based on its ability to inhibit apoptosis and improve tumor-cell survival. Consequently, targeting of the IL10 pathway using soluble IL10Rα³² or a dominant-negative STAT3³³ could block expansion of lymphoma cells. Finally, functional validation of other Myc-regulated genes, which is under way in our laboratory, might yield new therapeutic targets.