BCL2 Accelerates Inflammation-induced BALB/c Plasmacytomas and Promotes Novel Tumors with Coexisting T(12;15) and T(6;15) Translocations

Peritoneal plasmacytomagenesis (PCTG) in mice provides a model of neoplastic B-cell transformation that is dependent on the BALB/c (C) genetic background, a source of interleukin 6 (IL-6), and deregulated Myc (reviewed in Ref. 1). Whereas strain C is susceptible to peritoneal plasmacytomas (PCTs), most other common inbred strains of mice, including C57BL/6 (B6), are resistant (2). The susceptibility of strain C is a complex genetic trait that involves multiple loci, including a hypomorphic (weak efficiency) allele of the p16^INK4a gene, which exhibits a compound defect in the coding sequence (3) and the promoter (4). However, C mice do not develop PCT spontaneously. Instead, the tumors must be induced by i.p. administration of proinflammatory agents, such as pristane (5). Pristane provokes the formation of a chronic granulomatous tissue that serves as the site of PCT development (6) and a rich source of IL-6 (7) in situ. Pristane-induced PCTs are abrogated in C mice homozygous for an IL-6 null allele (8). Conversely, PCTs develop spontaneously in IL-6 transgenic C mice (9), illustrating the critical role of IL-6 in PCT development. Virtually all pristane-induced PCTs harbor chromosomal translocations (10) that activate Myc (11), demonstrating the requirement for Myc deregulation in PCTG. The translocations join the immunoglobulin heavy-chain locus Igh at 12F2 or one of the immunoglobulin light-chain loci, Igκ at 6C1 or Igλ at 16A3, with the Myc-Pvt1 locus at 15D1, which results in T(12;15)(Igh-Myc), T(6;15)(Igκ-Pvt1), or T(15;16)(Pvt1-Igλ). The most common translocation in pristane-induced PCT is the T(12;15).

Peritoneal PCTG in inbred C mice is an inefficient process characterized by incomplete penetrance (tumor incidence, ≤60%) and long latency (220 days, on average). In addition, it is necessary to maintain the mice in a conventional [nonspecific pathogen-free (SPF)] colony, which is thought to promote PCT by exposing the mice to environmental antigen. Antigen signaling through the B-cell receptor can provide a strong survival signal for mature B cells (reviewed in Refs. 12 and 13) and may be important for malignant B-cell and plasma cell transformation, particularly in the early stages (reviewed in Ref. 12). C mice raised in a germ-free environment are resistant to PCT (14), whereas C mice raised in SPF conditions exhibit a tumor incidence of ≤5% (15). Recent studies on Myc have further supported the contention that insufficient survival signals and consequent widespread apoptosis of PCT precursors may be the underlying reason for the abrogated and inefficient PCT induction in germ-free and SPF-maintained C mice, respectively. Although normally, Myc promotes cell growth and proliferation in the presence of growth factors (16), deregulated Myc, in the absence of growth factors, can prevent withdrawal of cells from the cell cycle or force quiescent cells into active cell cycle (17). However, when mitogens, cytokines, and other growth and survival factors are missing or limiting, deregulated Myc can also be a potent trigger of programmed cell death (apoptosis), a safeguard mechanism for eliminating unwanted cells with constitutively active oncogenes (18, 19). Thus, PCT precursors containing activated Myc may be removed by Myc-induced apoptosis whenever positive survival signals, such as IL-6 and antigen stimulation, fall below a critical threshold in vivo.

Studies on T(12;15) translocations in mice support the hypothesis that Myc-dependent apoptosis limits pristane-induced PCT development. The evidence is primarily based on identifying the reciprocal Igh-Myc translocation breakpoint junctions by PCR, which has been useful to evaluate the occurrence, trafficking, and diversification of translocation-harboring cell clones in mice (reviewed in Ref. 20). The abundance of Igh-Myc-carrying cell clones at 7 or 30 days after pristane administration (21) and the propensity of C mice to generate such cell clones throughout tumor development (22) indicated that the majority of T(12;15)-harboring cells do not evolve into PCT, possibly because they are eliminated by Myc-induced apoptosis. Genetic studies showed that although PCT-resistant B6, DBA/2N, and C3H/HeJ mice are able to generate Igh-Myc-containing cell clones on treatment with pristane, the frequency is low, and clonal expansion is minimal relative to PCT-susceptible C mice (23). This suggested that PCT-resistant strains of mice are more effective than C mice in clearing T(12;15)-carrying B cells and plasma cells from tissues.

The enforced expression of a death repressor in B cells (reviewed in Ref. 24) may increase the apoptotic threshold for PCT precursors, creating an improved PCT model with increased tumor incidence, shorter tumor latency, and performance under SPF conditions. During normal B-cell development, BCL2 blocks many death signals, including those induced by Myc (18, 19), cooperates with Myc in vitro to promote mitogen-independent survival (25), and accelerates Myc-driven B lymphoma development in vivo (26). Suppression of BCL2 expression, an important Myc-dependent pathway of apoptosis, is circumvented frequently during neoplastic B-cell transformation in mice (27). Previous work with two independently developed BCL2 transgenes (28, 29) showed that BCL2 selects B cells with a rearranged Myc locus for additional tumor development (30, 31). Although the Myc rearrangements were not identified, it is likely that they were T(12;15) translocations. Transfer of a BCL2 transgene to PCT-susceptible C mice might accelerate development of PCT with T(12;15) translocations.

To investigate this, we transferred the EμSV-Bcl-2-22 transgene (human BCL2 cDNA driven by the intronic heavy-chain enhancer Eμ) from PCT-resistant B6 mice onto PCT-susceptible C mice. We chose the EμSV-Bcl-2-22 transgene because BCL2 expression was confined to B cells and accompanied in some mice by spontaneous PCT (incidence, 7%) with rearrangements of Myc (four of six cases) detected by Southern blotting (32). Here, we show that in C.BCL2 congenics, PCT development is twice as fast (mean tumor onset, 113 days) as in conventionally maintained C mice (mean tumor onset, 220 days) and is independent of environmental antigen stimulation (SPF maintenance). We have, thus, generated a robust mouse model of neoplastic plasma cell development that may be valuable for developing and evaluating approaches to inhibit BCL2 in human plasma cell neoplasms where it is overexpressed [e.g., Waldenström’s macroglobulinemia and multiple myeloma (33, 34, 35)].

B6 mice carrying the EμSV-Bcl-2-22 transgene (B6.BCL2 mice) were a kind gift from Drs. Andreas Strasser, Alan Harris, and Jerry Adams (Walter and Eliza Hall Institute, Melbourne, Australia; Ref. 31). The mice were derived at the Walter and Eliza Hall Institute from mixed background (C57BL/6Wehi × SJL/JWehi) transgenics and imported to the Karolinska Institute (Stockholm, Sweden), where the transgene was transferred onto C mice by introgressive backcrossing. The EμSV-Bcl-2-22 consists of a human BCL2 cDNA inserted into an expression vector containing the intronic heavy-chain enhancer Eμ and regulatory elements (promoter, splice, and polyadenylation signals) from the SV40 virus. Among three independently developed EμSV-Bcl-2 transgenic lines, line 22 was particularly promising because BCL2 expression was confined to B cells. Moreover, the mice exhibited a propensity to PCT (incidence, 7%) with rearrangements of Myc, which were detected by Southern blotting in four of six tumors. Myc rearrangements of this sort are highly suggestive of T(12;15) translocations, because the variant Myc-activating T(6;15) and T(15;16) translocations cannot be detected by conventional Southern analysis (the breakpoints are ≥80 kb 3′ of Myc), and alternative mechanisms of Myc rearrangements, such as proviral insertion (36) and Eμ transpositions (37), are rare. Although C.BCL2 mice exhibited traits (e.g., leukocytosis, expansion of the B-cell compartment, excessive immunoglobulin production) that caused severe autoimmune disease on the B6 background (32), the mice seemed to be less prone to this disease than their B6 counterparts.

C.BCL2 N₆ and B6.BCL2 N₁₈ mice, 4–6 weeks of age and both hemizygous for EμSV-bcl-2-22, were treated with three i.p. injections of 0.3 ml of pristane spaced 1 month apart. Untreated C.BCL2 and B6.BCL2 mice and inbred C and B6 mice treated with pristane or left untreated were used as controls. Incipient PCT and lymphomas were detected by monitoring the mice for the development of ascites and lymphosplenomegaly. Each mouse developing ascites underwent abdominal paracentesis by the insertion of a sterile 25-gauge needle and collection of a drop of peritioneal fluid in the plastic adaptor of the needle. Two smears were made and stained with May-Grünwald-Giemsa. Tumor diagnosis was established by finding ≥100 characteristic, large, hyperchromatic tumor cells on a slide. The tumors were classified as PCT or lymphoma according to the cytological parameters defined in a previous study (38). The distinction of PCT from lymphoma was reliable and reproducible between two investigators (S. S. and A. L. K.) despite some cytomorphological variation among tumors. The tumor induction study was terminated on day 300 after pristane administration. All surviving mice were autopsied to confirm that they had not developed tumors.

Chromosome spreads were prepared by lysing tumor cells in hypotonic 75 mm KCl (30 min, ambient temperature), followed by drop-fixation with methanol/acetic acid (3:1). To establish tumor karyotypes, 15–25 well-spread trypsin/Giemsa-banded metaphase plates were evaluated. To confirm the presence of PCT-associated Myc translocations and detect hidden numerical and/or structural aberrations of chromosomes 12, 6, and 15, whole chromosome fluorescence in situ hybridization painting was performed using commercially available probes (Cambio, Cambridge, United Kingdom). FITC and rhodamine were used to label chromosomes 12 (green) and 15 (red), respectively. Both fluorochromes were combined to label chromosomes 6 (yellow). Chromosomes were counterstained with 4′,6-diamidino-2-phenylindole and embedded in antifade solution before analysis on an Axioskope epifluorescence microscope (Zeiss, Thornwood, NY) equipped with a 50 W fluorescence bulb and single- and dual-band filters for FITC and rhodamine.

High molecular weight DNA was prepared from tumor tissues using a standard phenol and chloroform extraction protocol, followed by ethanol precipitation. Illegitimate genetic recombinations between Igh and Myc, the molecular indicators of T(12;15), were detected by long-template or high-fidelity PCR methods as described previously (22, 39). PCR fragments were gel purified with the QIAquick gel extraction kit (Qiagen, Valencia, CA) and sequenced directly using the Femtomole cycle sequencing kit (Promega, Madison, WI).

To detect rearrangements of Myc and clonotypic VDJ recombinations by Southern blot hybridization, genomic DNA was digested with BssI or EcoRI, fractionated by electrophoresis on 0.7% agarose gels, transferred onto a nitrocellulose membrane, and hybridized to a 2.2-kb NheI/SpeI fragment of Myc including exon 2, or to a 1.5-kb HindIII/EcoRI fragment of Igh spanning J_H2 and Eμ. Probes were labeled with ³²P-dCTP by random priming using a commercial kit. The Igh probe cohybridized to the EμSV-bcl-2-22 transgene, which resulted in four transgene-derived restriction fragments on EcoRI digestion.

Three i.p. injections of 0.1–0.5 ml of pristane into conventionally maintained C mice typically induce PCT in 60% of the mice after an average latency of 220 days (40). PCT incidence drops below 5% when the mice are kept in a SPF colony (15). To evaluate whether the enforced expression of BCL2 in B cells overrides the requirement for conventional mouse maintenance and promotes PCT under SPF conditions, 34 C.BCL2 mice kept in a SPF colony were treated with three i.p. injections of 0.3 ml of pristane spaced 1 month apart. Tumor development was monitored by examining ascite cell specimens for the presence of neoplastic cells in the peritoneal cavity. Twenty-four of 34 (70.5%) pristane-treated C.BCL2 mice developed B-cell tumors (Fig. 1,A, column 1). The predominant tumor type was PCT, which occurred in 19 of 34 (55.9%) mice with a mean latency of 113 ± 28 days (Fig. 1,B). Five of 34 (14.7%) mice developed B lymphomas with a mean onset of 109 ± 24 days (Fig. 1,B) and a distinct morphology compared with PCT (Fig. 1,C). At autopsy, both PCT and lymphomas were mainly found in the pristane granuloma; however, advanced tumors had frequently spread to the spleen (splenomegaly up to 800 mg) and the mesenteric lymph node. No tumor was observed in control groups of 13 untreated C.BCL2 mice and 30 untreated C mice (Fig. 1,A, columns 5 and 6) after an observation period of 300 days. One PCT was detected in 44 pristane-treated C mice (2.3%; Fig. 1,A, column 2). In pristane-treated B6.BCL2 mice, the sole tumor type was lymphoma (11 of 20, 55%), in two cases with significant plasmacytic differentiation (Fig. 1,A, column 3). Mean tumor latency was 185 ± 49 days (Fig. 1,B). All 16 mice from the B6.BCL2 control group (no pristane) and all nontransgenic littermates (12 treated with pristane and 15 left untreated) remained tumor free after 300 days (Fig. 1 A, columns 7, 4, and 8). These findings established that transgenic expression of BCL2 in B cells creates a model of peritoneal PCT development that is still dependent on the C background but does not require that the mice be maintained in a conventional colony. Furthermore, tumor development in C.BCL2 mice was significantly faster than in B6.BCL2 mice (P = 0.003, two-sided Student’s t test), indicating that the C genotype not only modified the tumor phenotype but also accelerated tumor development.

Previous studies on lymphoma progression in two independently developed strains of BCL2 transgenic mice suggested that BCL2 selects B cells harboring Myc-activating T(12;15) translocations for outgrowth and malignant transformation (28, 29). To examine this, the cytogenetic make-up of 25 BCL2 transgenic tumors was determined. The sample was representative of the phenotypes and genotypes in this study, because it included 12 PCTs and 2 lymphomas from the C.BCL2 group and 11 lymphomas from the B6.BCL2 group. The chromosome complement of most tumors (18 of 25, 72%) was near diploid (<40 chromosomes), pseudodiploid (40 chromosomes), or hyperdiploid (41–45 chromosomes), whereas the remaining tumors (7 of 25, 28%) were near tetraploid (56–62 chromosomes) or tetraploid (∼80 chromosomes). Importantly, all 25 tumors, including those of B6 origin, contained T(12;15) translocations in their trypsin/Giemsa-stained metaphase plates. The karyotype of one near diploid B6.BCL2 lymphoma is shown in Fig. 2,A. Chromosome painting by fluorescence in situ hybridization confirmed the presence of T(12;15) in this tumor (Fig. 2 B) and all other tumors (data not shown). The present findings in 11 B6 lymphomas provide the first cytogenetic evidence that the T(12;15) translocation is not specific for plasma cell tumors of PCT-susceptible C mice (10, 41, 42, 43) but occurs also in lymphomas of PCT-resistant B6 mice. This is consistent with previous PCR studies demonstrating the occurrence of Igh-Myc junction fragments in B6 and other PCT-resistant strains of mice (23, 44). The reason why T(12;15)-carrying B cells evolve into PCT in C mice but lymphomas in B6 mice is not known. One possibility is that some alleles of strain C facilitate the differentiation of aberrant B cells with deregulated Myc into neoplastic plasma cells.

The examination of 15–25 metaphase plates from all 25 BCL2 transgenic tumors revealed that 5 of 12 PCT (42%) contained in addition to T(12;15) a reciprocal T(6;15) translocation. Tumor oligoclonality [the presence of independent T(12;15)-bearing or T(6;15)-bearing cell clones admixed in the peritoneal fluid] was ruled out as a potential explanation, because both translocations were repeatedly observed in the same trypsin/Giemsa-stained metaphase plates (Fig. 3). In addition, fluorescence in situ hybridization painting of chromosomes 6, 12, and 15 confirmed the coexistence of T(6;15) and T(12;15) in all five PCTs. Fluorescence in situ hybridization images of one tumor, C-PCT-24, are shown in Fig. 4, A and B, as an example. All PCTs with coexisting T(12;15) and T(6;15) translocations exhibited additional karyotypic changes, notably numerical aberrations (supplementary Table 1). Aberrations of this sort were random, independent of tumor type (lymphoma or PCT) and genetic background, and also observed in the other 20 karyotyped tumors carrying only T(12;15) (data not shown). It is likely that the aberrations reflected general genomic instability, similar to the instability observed in pristane-induced PCT in conventionally maintained C mice (45) and Burkitt-like lymphomas in λ–MYC transgenic B6 mice (46). Genomic instability may also be the underlying reason for the apparent clonal diversification of some tumors. One example is depicted in Fig. 4 C, which shows three subclones of tumor C-PCT-24 characterized by trisomies 15 and 19 or T(8;12) and T(11;12) translocations.

The karyotypic analysis of PCT with coexisting T(12;15) and T(6;15) translocations suggested a sequence of cytogenetic changes that begins with the T(12;15), proceeds with the duplication of chromosome 15, and ends with the T(6;15). First, in tetraploid metaphase plates of three of five tumors with coexisting T(12;15) and T(6;15) translocations, chromosomes with T(12;15) occurred as two copies, whereas chromosomes with T(6;15) occurred as single copies. This indicated that the T(12;15) took place in a diploid cell, whereas the T(6;15) occurred in a tetraploid cell. Tetraploidization is a common event in peritoneal PCTG. The proportional increase in the number of translocated chromosomes during genome duplication events was sometimes propagated to hypertetraploid state, in which four copies of T(12;15) but only two copies of T(6;15) were found in the karyotypes (Fig. 4,B). Second, all five tumors with coexisting T(12;15) and T(6;15) translocations also contained intact copies of chromosome 15 in their metaphase plates. This would not be the case if T(6;15) occurred in a T(12;15)-bearing diploid cell before tetraploidization, because this would result in the rearrangement of both copies of chromosome 15. Instead, the presence of normal 15 in a near diploid cell that also contains T(12;15) and T(6;15) requires that the duplication of 15 (nondisjunction during mitosis resulting in trisomy 15) precedes one of the translocation events. Although the temporal relationship of the two translocations with respect to +15 cannot be decided in a near diploid or pseudodiploid cell, the above-stated findings in the tetraploid tumors strongly suggested that T(12;15) usually precedes T(6;15). It follows that trisomy 15 must occur before T(6;15). A hypothetical scheme of the sequence of cytogenetic changes leading to PCT with coexisting T(12;15) and T(6;15) translocations plus trisomy 15 is shown in Fig. 5. The scheme is consistent with empirical evidence suggesting that trisomy 15 is a frequent event in mouse PCT (47) and other mouse neoplasms (48, 49), in which it can effect the overexpression of Myc.

To evaluate the fine structure of the T(12;15) translocation in BCL2 transgenic tumors, the reciprocal Igh-Myc junction fragments of five C.Bcl-2 PCT and five B6.Bcl-2 lymphomas were amplified by PCR and analyzed by DNA sequencing, as described previously (22, 39). The results are summarized in supplementary Table 2 and Fig. 6,A. The breakpoints on the Myc-activating product of translocation, chromosome T(12; 15) were distributed across the entire Igh locus, specifically in the switch (S) regions of Cα (three cases), Cε (one case), Cγ2b (five cases), and Cμ (one case; Fig. 6,A, left). In two cases, C-PCT-33 and B6-Ly-1, the breaksites occurred in the Sμ portion of a composite Sμ/Sγ2b region (supplementary Table 2). In the other eight cases, the breaks took place in the (unitary) switch region that corresponded to the indicated C_H locus. The breakpoints in Igh on the reciprocal product of translocation chromosome T(15;12), were located in Sμ (five cases), the 5′ flank of Sμ (three cases), or the vicinity of J_H4 (one case; Fig. 6 A, data not shown, and supplementary Table 2). Note that in one case, B6-Ly-11, the reciprocal product was not detected. The distribution of the translocation breaksites in Igh in BCL2 transgenic tumors was similar to that in pristane-induced PCT in conventionally maintained C mice, which tend to use the upstream portion of C_H (Sμ, J_H, or recombined DJ/VDJ genes) for breakpoints on chromosomes T(15;12) but the downstream portion of C_H, particularly Sα, for breakpoints on chromosome T(12;15) (21).

The breakpoints in Myc in 9 of 10 BCL2 transgenic tumors were tightly clustered in a 1-kb region in the near 5′ flank of the gene (Fig. 6 A, right). The exception was tumor C-PCT-32, which had a breakpoint in exon 1 in the Myc promoter. In the nine tumors with breaksites in the 5′ flank of Myc, the resulting exchanges [class II rearrangements of Myc according to Cory (50)] allocated the intact Myc gene, including the noncoding regulatory first exon with the two major Myc promoters, P1 and P2, to the Myc-deregulated product of translocation, chromosome T(12;15). The high prevalence of class II rearrangements of Myc in BCL2 transgenic tumors (90%) was unexpected, because rearrangements of this sort are rare in conventional PCT, IL-6 transgenic PCT (9), and presumptive precursors of IL-6 transgenic PCT (39). However, class II rearrangements have been observed in a limited study on precursors of conventional PCT (51). In 8 of 10 tumors, the translocation resulted in a loss of Myc sequence during the genetic exchange [compare breaksites in Myc on chromosome T(12;15) (supplementary Table 2) with the breaksites on chromosome T(15;12) (supplementary Table 2)]. The average size of this loss was 74 ± 114 bp (range, 1–340), similar to previous findings in PCT of conventional C mice (52). One BCL2 tumor, C-PCT-30, exhibited a precise exchange. One tumor, B6-Ly-11, was not evaluated because the reciprocal product of translocation was not detected. Two tumors demonstrated internal deletions in Myc in the vicinity of the Igh-Myc junction: C-PCT-33 (26 bp) and B6-Ly-1 (236 bp; see Myc/Myc junctions in supplementary Table 2).

Southern blot hybridization of six randomly chosen tumors not included in the PCR study showed clonotypic V(D)J rearrangements in all tumors (Fig. 6,B, bottom) and Myc rearrangements consistent with T(12;15) translocations in five of six tumors (Fig. 6 B, top). The reason why the exceptional tumor (B6-Ly-2) had no Myc rearrangement by Southern blotting is not known. The ready detection of Myc rearrangements in five of six cytogenetically positive tumors strengthened the contention that the great majority of BCL2 transgenic tumors carry T(12;15).

This study has demonstrated that the enforced expression of a human BCL2 transgene creates a mouse model of peritoneal PCT that is less dependent on environmental antigen stimulation than the classic model of pristane-induced PCT in conventionally maintained C mice. BCL2 transgenic PCT share several features with classic PCT. These include the dependence on the inflammatory granuloma for a source of IL-6 in situ, the requirement for the PCT susceptibility alleles of the C background, and the acquisition of chromosomal T(12;15) translocations that deregulate Myc. Features distinguishing BCL2 transgenic PCT from classic PCT are the speed of transgenic tumor development (mean latency, 112 days) and their strong bias for the subset of T(12;15) translocations that result in class II rearrangements of Myc (50). Furthermore, 5 of 12 (42%) karyotyped tumors contained a T(6;15) translocation in addition to the T(12;15). The coexistence of T(6;15) and T(12;15) in the same metaphase plates of primary PCT has never been observed in >300 PCTs karyotyped in our laboratory (reviewed in Ref. 1).³ Finally, BCL2 transgenic PCT developed under SPF conditions, whereas classic PCT requires that the mice are maintained in a conventional colony. This indicated that the transgenic expression of BCL2 substitutes for the positive growth and survival signals that are delivered to B cells by antigen under conventional maintenance conditions.

Our observation that all 25 karyotyped BCL2 transgenic tumors harbored the Myc-activating T(12;15) translocation indicated that BCL2 collaborates with Myc to promote neoplastic B-cell and plasma cell development. Studies on Myc-driven B-cell neoplasms in Εμ-Myc transgenic mice (53) have strongly suggested that attenuation of Myc-dependent apoptosis by BCL2 is the underlying mechanism of the Myc-BCL2 collaboration in oncogenesis (27, 54). The apparent ability of BCL2 to rescue aberrant B cells harboring T(12;15) is consistent with reports that overexpression of BCL2 rescues B cells that would normally be eliminated because of developmental problems (55), autoreactivity (56), dysfunctional B-cell receptors (57), or illegitimate genetic rearrangements (58). However, protection from apoptosis by BCL2 does not exclude the possibility that BCL2 uses alternative mechanisms to accelerate PCT development in mice. The ability of BCL2 in normal cells to slow entry into the cell cycle and promote exit from the cycle is sometimes lost in tumor cells, whereas the antiapoptotic effect of BCL2 is retained (reviewed in Ref. 59). Furthermore, although BCL2 has been shown to posses antimutagenic activities (60, 61), it can also exert powerful promutagenic and genome-destabilizing effects [e.g., by inhibiting DNA repair (62) and enhancing oxidative mutagenesis (63)]. The latter may be particularly relevant for peritoneal PCT, which may be facilitated by oxidative mutagenesis (64). Thus, although the mechanism of the Myc-BCL2 collaboration may be complex and although differences in Myc deregulation in Eμ-Myc transgenic B cells and T(12;15)-harboring B cells may further modify this collaboration, the findings in the Eμ-Myc model strongly suggest that protection from Myc-induced apoptosis is a critical component of the ability of BCL2 to promote B-cell and plasma cell tumor formation in mice.

The preference in BCL2 transgenic B-cell tumors for the subset of T(12;15) that results in class II rearrangements of Myc is not understood. In class II translocations, the intact Myc gene, including the first regulatory exon with the major Myc promoter, P1 and P2, is retained on the Myc-activating product of translocation chromosome T(12;15). Class II rearrangements of Myc are uncommon in pristane-induced PCT in conventionally maintained C mice and “spontaneous” PCT in IL-6 transgenic C mice (9). In these tumors, translocation breakpoints usually occur in exon 1/intron 1 of Myc, which “decapitate” the gene and allocate the P1/P2 promoter to the reciprocal product of translocation, chromosome T(15;12). It has long been suspected (51, 65) but not shown that tumors with class II translocations have more elevated levels of Myc mRNA and/or different mechanisms of Myc activation than tumors with class I translocations. Because it is likely that elevated levels of Myc are accompanied by increased proclivity to undergo Myc-dependent apoptosis, class II translocations may be poorly tolerated in mice unless the threshold of apoptosis is raised, in this model by enforced expression of BCL2. The antiapoptotic effect of BCL2 may also be key for the development of PCTs that contain a secondary T(6;15) translocation. This translocation is cytogenetically indistinguishable from the Myc-activating T(6;15) in classic PCT, but additional studies are warranted to evaluate its fine structure (Pvt1-Cκ juxtaposition?) and possible impact on Myc levels (additional up-regulation of Myc?). Secondary Myc-activating translocations [t(8;14)] have been identified as tumor progression events in human B-cell neoplasms (66, 67, 68).

In conclusion, protection from Myc-induced apoptosis (BCL2) collaborates with PCT susceptibility alleles (C genotype), IL-6 signaling (pristane granuloma), and deregulated Myc (chromosomal translocation) to promote pristane-induced PCT in mice. BCL2-overexpressing PCTs are still dependent on IL-6, indicating that BCL2 and IL-6 function in different pathways of tumor development. Accelerated PCT development in BCL2 transgenic C mice may be relevant for human plasma cell tumors overexpressing BCL2 (33, 34, 35), particularly for preclinical intervention studies aimed at inhibiting BCL2 in these tumors (69, 70).

We thank Drs. Andreas Strasser, Alan Harris, and Jerry Adams of Walter and Eliza Hall Institute (Melbourne, Australia) for kindly providing C57BL/6 EμSV-BCL2-22 transgenic mice, Katalin Benedek (Karolinska Institute, Stockholm, Sweden) for excellent technical assistance, Dr. Lynne Rockwood (National Cancer Institute, NIH, Bethesda, MD) for reading this manuscript and making helpful editorial suggestions, and Dr. Beverly Mock (National Cancer Institute, NIH, Bethesda, MD) for stimulating scientific discussion.