Conditional inactivation of p53 in mature B cellspromotes generation of nongerminal center-derivedB-cell lymphomasMonica Gostissaa,b,c, Julia M. Biancoa,b,c, Daniel J. Malkina,b,c, Jeffery L. Kutokd, Scott J. Rodigd, Herbert C. Morse IIIe,Craig H. Bassingf, and Frederick W. Alta,b,c,1

aHoward Hughes Medical Institute, bProgram in Cellular and Molecular Medicine, Boston Children’s Hospital, and cDepartment of Genetics, Harvard MedicalSchool, Boston, MA 02115; dDepartment of Pathology, Brigham and Women’s Hospital, Harvard Medical School, Boston, MA 02115; eLaboratory ofImmunogenetics, National Institute of Allergy and Infectious Diseases, National Institutes of Health, Rockville, MD 20852; and fDepartment of Pathologyand Laboratory Medicine, Center for Childhood Cancer Research, Children’s Hospital of Philadelphia, Philadelphia, PA 19104

Contributed by Frederick W. Alt, December 28, 2012 (sent for review November 18, 2012)

The p53 tumor suppressor exerts a central role in protecting cellsfrom oncogenic transformation. Accordingly, the p53 gene is mu-tated in a large number of human cancers. In mice, germ-line in-activation of p53 confers strong predisposition to development ofdifferent types of malignancies, but the early onset of thymiclymphomas in the majority of the animals prevents detailed studiesof tumorigenesis in other tissues. Here, we use the Cre/Loxapproach to inactivate p53 in mature B cells in mice (referred to as“CP” B cells) and find that such p53 inactivation results in theroutine development of IgM-positive CP peripheral B-cell lympho-mas. The CP lymphomas generally appear to arise, even in micesubjected to immunization protocols to activate germinal centerreaction, from naive B cells that had not undergone immuno-globulin (Ig) heavy chain gene class switching or somatic hyper-mutation. In contrast to thymic lymphomas that arise in p53-deficient mice, which generally lack clonal translocations, nearlyall analyzed CP B-cell tumors carried clonal translocations. How-ever, in contrast to spontaneous translocations in other mouse B-celltumor models, CP B-cell tumor translocations were not recurrent anddid not involve Ig loci. Therefore, CP tumors might provide modelsfor human lymphomas lacking Ig translocations, such as splenic mar-ginal zone B-cell lymphoma or Waldenstrom macroglobulinemia.Our studies indicate that deletion of p53 is sufficient to trigger trans-formation of mature B cells and support the notion that p53 defi-ciency may allow accumulation of oncogenic translocations in B cells.

The p53 tumor suppressor is a transcription factor that regu-lates a large array of genes involved in control of cell cycle

and apoptosis (1, 2). Transactivation-independent activities ofp53 have also been described, ranging from transcriptional re-pression (3) to cytoplasmic and mitochondrial functions (2). Levelsof p53 protein are extremely low in normal conditions, but p53becomes stabilized and activated by a variety of posttranslationalmodifications in cells subjected to different types of DNA damageas well as upon overexpression of oncogenes (1, 2). As a result ofp53 activation, cells carrying potentially harmful lesions, such asDNA double-strand breaks (DSBs) or mutations that activateoncogenes, initiate cell cycle arrest to repair the lesion or un-dergo programmed cell death. Germ-line p53 mutations in humanscause Li-Fraumeni syndrome, a familial condition characterizedby early onset of different tumors (4, 5). Moreover, the p53 geneis somatically mutated or deleted in a large number of humancancers, indicating that this tumor suppressor exerts its protec-tive role against oncogenic transformation in multiple tissues (5).Targeted disruption of the p53 gene in mice, however, results ina strong predisposition for early-onset thymic lymphomas (6). Asmall percentage of germ-line p53-deficient mice succumb to Blineage lymphomas (7, 8), but the short lifespan of these animalresulting from thymic lymphoma prevented more detailed studiesof the effects of p53 deficiency in different stages of B-cell dif-ferentiation as well as in other tissues. By using the Cre/Lox ap-proach in mice with p53-conditional alleles (6, 9), several studies

demonstrated that somatic inactivation of p53 is sufficient topromote tumor formation in some, but not all, tissues examined.Thus, for example, p53 deletion per se results in development ofbreast tumors (10) and osteosarcomas (11, 12), whereas devel-opment of ovarian or prostate cancers requires simultaneous de-letion of other tumor suppressors (13, 14).In humans, many B- and T-cell lymphomas are characterized

by clonal translocations that usually juxtapose an oncogene toantigen receptor loci (15, 16). Translocations in progenitors oflymphoid tumors involve on one partner programmed DSBsthat are generated in the context of Ig gene assembly in B cells,and T-cell receptor (TCR) assembly in T cells. This process,called V(D)J recombination, takes place in early stages of B- andT-cell differentiation and is initiated by the RAG endonuclease,formed by the products of recombination activating gene (RAG)-1and -2. RAG introduces DSBs at target V, D, and J segments inthe Ig and TCR loci, which are then joined by the classical non-homologous DNA end-joining pathway (C-NHEJ) (16, 17). Uponantigen stimulation of mature B cells, the constant (C) region ofthe Ig heavy chain (IgH) molecule, which initially is encoded bythe Cμ exons, can be exchanged to another heavy chain isotype,with different effector functions, by class switch recombination(CSR) (16, 17). CSR is initiated by activation-induced cytidinedeaminase (AID) activity, which leads to DSBs within large re-petitive sequences (S regions) that flank each set of IgH C regionexons with the breaks subsequently being joined by C-NHEJ oralternative end-joining (16, 17). AID is also responsible for an-other Ig diversification process, somatic hypermutation (SHM),which introduces of mutations in the variable region exons,allowing the selection of B cells that produce Ig molecules withhigher affinity for antigen (18, 19). SHM takes place in special-ized structures, called germinal centers (GCs) that organize inperipheral lymphoid organ following antigen encounter,whereas CSR can also occur outside of the GC reaction (20).Human B-cell lymphomas can originate at different stages

of B-cell differentiation, as can be inferred by examining thepattern of Ig loci rearrangements. Many human B-cell lym-phomas, such as follicular lymphomas or large B-cell lym-phomas, are of GC or post-GC origin and accordingly carryswitched and hypermutated IgH alleles (21). These tumorsroutinely harbor translocations between IgH and oncogenessuch as c-myc and Bcl6, with breakpoints that can be ascribedto aberrant CSR or SHM (16, 17). Although true pre-GC lym-phomas, such as mantle cell lymphoma, are relatively rare, other

Author contributions: M.G., C.H.B., and F.W.A. designed research; M.G., J.M.B., D.J.M.,and J.L.K. performed research; M.G., J.L.K., S.J.R., and H.C.M. analyzed data; and M.G. andF.W.A. wrote the paper.

The authors declare no conflict of interest.

Freely available online through the PNAS open access option.1To whom correspondence should be addressed. E-mail: alt@enders.tch.harvard.edu.

This article contains supporting information online at www.pnas.org/lookup/suppl/doi:10.1073/pnas.1222570110/-/DCSupplemental.

2934–2939 | PNAS | February 19, 2013 | vol. 110 | no. 8 www.pnas.org/cgi/doi/10.1073/pnas.1222570110

Dow

nloa

ded

by g

uest

on

Sep

tem

ber

14, 2

020

cases, including splenic marginal zone lymphoma (MZL) andWaldenstrom macroglobulinemia, derive from B cells that mayhave undergone SHM but not CSR and usually do not harborIgH translocations (22–24).Despite the frequency of translocations involving antigen re-

ceptor loci in human lymphomas, T-cell tumors arising in germ-line p53-null animals harbor clonal translocations only in a mi-nority of the cases. Moreover, the observed translocations arenot recurrent and do not involve TCR loci (25). When both p53and C-NHEJ factors are deleted in the germ line or in B-lineagecells, mice invariably develop B-cell lymphomas with character-istic translocations between the IgH and the c-myc oncogene loci(16), and p53 deficiency allows accumulation of IgH/c-myc trans-locations in normal B cells stimulated to undergo CSR (26). Tobetter characterize the role of p53 in B-cell lymphoma formation,given that the onset of these tumors is masked by the fast ap-pearance of thymic lymphomas in p53-null mice, we condition-ally inactivated the p53 gene in mature B cells by mean of theCre/Lox approach.

ResultsMice with Mature B Cell–Specific Inactivation of p53 Develop B-CellLymphomas. To specifically delete the p53 gene in mature B cells,we bred a previously generated p53-floxed conditional allele(p53F, Fig. 1A) (9) into the CD21-Cre background (27). Cre re-combinase expression from the CD21 promoter takes place duringdifferentiation from immature transitional B cells to mature B cells;accordingly, efficient p53 deletion was observed by Southern blot-ting in peripheral B cells (Fig. 1A). CD21-Cre may become non-specifically active in the germ line; therefore, we generated ourexperimental cohort by crossing CD21-Cre/p53F/F males to p53F/Ffemales. CD21-Cre/p53F/F and CD21-Cre/p53F/- offspring (col-lectively referred to as “CP,” with the latter deriving from non-specific Cre-mediated deletion of one floxed p53 allele) wasmonitored for tumor development. Because deficiency of thehistone variant H2AX does not confer tumor susceptibility (28, 29),we similarly generated a cohort of CD21-Cre/H2AXF/F or CD21-Cre/H2AXF/- mice (referred to as “CH”) as controls.To promote immune responses, half of the mice in each cohort

were immunized by injection of sheep RBCs, according to stan-dard protocols (30). No difference in overall survival and causeof death were observed between the nonimmunized and immu-nized groups (Table S1); therefore, results from the twogroups are presented together. CP mice became moribundbetween 8 and 12 mo of age, whereas control CH mice usuallylived up to 2 y of age (Fig. 1B). About 50% of CP mice succumbedto B-cell lymphomas, which likely originated in the spleen (af-fected in 10/10 mice and usually greatly enlarged, Fig. 1C). Thetumor often involved peripheral lymph nodes (5/10 mice), andmore rarely mesenteric lymph nodes (3/10 mice) and thymus(3/10 mice). Metastasis to the liver was observed in four animals.Development of B-cell lymphomas was more frequent in p53F/Fthan in p53F/- mice, because the latter were also susceptible toother types of cancers, likely from p53 heterozygosity (Table S1).Indeed, three CP mice of the p53F/- genotype developed thymiclymphomas and sarcomas, which are typical in p53 heterozygousmice (31). Mice 187 and 307 presented with normal spleens andtumor masses, which histologically were of mixed lymphoid andmyeloid lineages; therefore, these two tumors were not analyzedfurther. In the CH cohort, only one mouse developed a B-celllymphoma at 21 mo of age, and two mice died of ovarian tumorsat more than 27 mo of age (Table S1).

CP B-Cell Lymphomas Are Surface IgM+ and of Pregerminal CenterOrigin. Histopathologic studies of sections from CP tumor sam-ples revealed a relatively complex picture. Tumor CP220 (Fig.2A, Upper Left) was characterized by a nodular growth patternand a mixed centrocytic/centroblastic population typical of mu-rine follicular lymphomas (32). Other cases (tumors CP245 andCP569; Fig. 2A, Upper Right) showed histologic and cytologicfeatures characteristic of murine splenic MZL (33) and were

remarkably similar to MZL previously identified in mice witha conventional knockout of p53 (8) or an altered p53 exon 1 (34).Some cases had a more diffuse and aggressive phenotype,characterized by high mitotic rates (tumors CP239 and CP301;Fig. 2A, Lower Left), larger cells (tumor CP166), or anaplasticcells (tumor CP277; Fig. 2A, Lower Right). Analysis of surfacemarker expression by flow cytometry revealed that CP lympho-mas were invariably B220+/IgM+, with most of the tumors alsobeing Igκ+. Tumor CP301 was the only IgM+/Igλ+ case (Fig. 2B).These results suggest that CP B-cell lymphomas most likelyoriginated from B cells that had not undergone CSR. Moreover,immunohistochemical analysis showed that a subset of the tumorswere Bcl6-negative, consistent with a pre-GC origin (Table S2).To further characterize these p53-deficient B-cell tumors, we

analyzed rearrangements at the IgH and Igκ loci in CP tumorDNA by Southern blotting. We first used probes hybridizingdownstream of the JH region (JH4–3 probe; Fig. 3A, Left) andof the Jκ region (Jκ probe; Fig. 3A, Right), and found that allsamples contained distinct, rearranged bands (Fig. 3B). Theseresults indicated that the tumors are monoclonal with one or twonon–germ-line bands (e.g., CP236, CP301) or oligoclonal withmore than two non–germ-line bands (e.g., CP220, CP250) andusually originated from single cells that had undergone V(D)J

A

B

C

Fig. 1. CP mice develop B-cell lymphomas. (A) Southern blot analysisdemonstrating deletion of p53 gene in mature B cells. A schematic map withthe position of relevant restriction sites and probes used is shown at top.Position of the bands corresponding to the wild-type (wt), floxed (fl), anddeleted (del) p53 alleles is indicated. Bc-day4, purified splenic B cells culturedfor 4 d with anti-CD40/IL4; Bcells, purified splenic B cells; Tcells, purifiedsplenic T cells. (B) Kaplan-Meier curve of the CP (n = 22) and control CH (n =17) cohorts. Curves represent total survival. (C) (Left) Example of enlargedspleen frequently observed in CP mice succumbing to B-cell lymphomas;(Right) normal control spleen.

Gostissa et al. PNAS | February 19, 2013 | vol. 110 | no. 8 | 2935

IMMUNOLO

GY

Dow

nloa

ded

by g

uest

on

Sep

tem

ber

14, 2

020

recombination at both IgH and Igκ loci. We next examined thestatus of the Sμ region by using an Iμ probe from the JH to Cμintron, which on an EcoRI digest recognizes a fragment that isnot within the IgH V, D, or J segments and is not altered by V(D)J recombination (Fig. 3A, Left). Sμ is the donor switch region forthe vast majority of IgH CSR events; therefore, detection of arearrangement with this probe would indicate that the cell oforigin of the tumor had undergone CSR. All but one (CP246) ofthe analyzed CP tumor samples carried germ-line, unrearrangedSμ bands (Fig. 3C), further confirming that they derive from cellsthat had not undergone CSR. We also analyzed SHM at the IgHlocus by PCR amplification and sequencing of the region en-compassing the intron between the intronic Eμ enhancer and therearranged JH segment (Fig. 3D). None of the lymphomas ana-lyzed except CP246 had any mutations, consistent with the no-tion that they arose from naive pre-GC B cells (Fig. 3D).

CP B-Cell Lymphomas Harbor Clonal Nonrecurrent Translocations.B-cell lymphomas in humans and mice routinely harbor clonaltranslocations that involve Ig loci and different oncogenes, suchas c-myc, Bcl2, or Bcl6 (16). To determine if this was also thecase for CP lymphomas, we performed spectral karyotyping (SKY)on metaphase spreads from short-term tumor cell cultures. Mostof the tumors showed variable degrees of aneuploidy, a charac-teristic associated with p53 deficiency (35). All but one of theanalyzed CP lymphomas carried multiple clonal translocationsthat were nonrecurrent, involving different chromosomes (chr)in each tumor analyzed (Fig. 4A and Table 1). Tumor CP569harbored reciprocal T(14;2) and T(2;14) translocations; tumor

CP239 harbored T(3;1), T(4;3), and T(17;2); and tumor CP166carried reciprocal T(1;15) and T(15;1) plus T(13;10) and T(10;9).Two tumors also harbored complex translocations: tumorCP220 had a complex translocation involving chr 8, 14, and 12in addition to a T(2;5) and a T(17;4), whereas tumor CP277had a complex T(5;3;5) translocation in addition to clonal T(6;19), T(4;6) and frequent T(17;10), and T(X;3). Another tumorgenerated outside of the experimental cohort, tumor CP752,similarly carried multiple clonal translocations including T(10;6),T(9;1), T(15;16), T(14;1), and T(11;14). Only tumor CP246 lackedany translocations that could be identified by SKY. This was alsothe only tumor exhibiting SHM and it developed with a shorterlatency than any of the other tumors (Table S1), suggesting itmay represent a different tumor type than other CP tumors.Mouse models deficient for p53 and C-NHEJ or DSB-

response factors usually carry translocations involving Ig loci(16). The IgH locus lies on the telomeric portion of chr 12,whereas the Igκ and Igλ loci lie on chr 6 and 16, respectively.Tumor CP220 has a complex T(8;14;12) translocation. However,the chr 12 breakpoint did not involve the IgH locus, as demon-strated by FISH analysis with IgH-flanking probes (Fig. 4B,Left, and Fig. S1A). Surprisingly, FISH analyses with probesflanking the TRCα/δ locus on chr 14 showed that the complextranslocation in this B-cell tumor splits the 5′ and 3′ ends of theTCRα/δ locus, which normally rearranges during V(D)J re-combination in pro-T cells. However, other translocations in-volving chr 14 in tumors CP569 and CP752 did not involveTCRα/δ (Fig. S1 C and D). Tumors CP277 and CP752 hadtranslocations involving chr 6 and tumor CP752 also had a chr 16translocation. FISH analysis with probes flanking Igκ and Igλshowed that these loci were not involved in the translocations(Fig. 4B and Fig. S1 B and D). Chr 6 also carries the TCRβ locus,which was not translocated in tumor CP752 but was partiallyduplicated in tumor CP277 (Fig. 4B and Fig. S1 B and D). To-gether, these analyses indicated that CP tumors do not carrytranslocations involving Ig loci.Mouse B-cell lymphomas often rearrange or amplify the c-myc

oncogene, resulting in high levels of c-myc expression that con-tribute to transformation (36–38). This seems not to be the casefor CP lymphomas because only one tumor, CP166, had atranslocation involving chr 15 that bears the c-myc locus. Toconfirm that c-myc overexpression was not a key factor duringlymphomagenesis in CP mice, we performed Northern blottingon RNA from selected CP tumors. High levels of c-myc tran-scripts were not detected (Fig. 4C), indicating that other, un-known mechanisms promote transformation of CP B cells.

DiscussionHere we show that specific inactivation of the p53 tumor sup-pressor gene in mature B cells is sufficient to promote oncogenictransformation and results in the development of splenic matureB-cell lymphomas. By histopathologic criteria, several CP B-celltumors resemble murine splenic marginal zone lymphomas(SMZL), with some cases showing a more aggressive phenotypeconsistent with diffuse large B-cell lymphomas characteristic ofhigh-grade MZL in mouse (33). Indeed, a previous study iden-tified the prevalent B-cell tumor in germ-line p53 knockout miceas SMZL and found progression from marginal zone hyper-plasia to invasive lymphomas composed of more pleomorphiccells (8). More recently, it was found that mice with an alteredfirst exon resulting in B cell–specific deletion of p53 also de-veloped SMZL (34). These studies, however, were limited tohistological analyses and did not examine cytological abnormali-ties nor determine the cellular origin of the tumors. Nonetheless,the combined results of these studies clearly indicate that, at leastin mice, splenic marginal zone B cells are uniquely sensitive to thetransforming effects of p53 deficiency.We found that all CP B-cell tumors were IgM+ and that only

one exhibited somatic hypermutation, indicating that their cell oforigin was a mature B cell that had not passaged the germinalcenter or undergone CSR. Accordingly, CP tumors lacked clonal

A

B

Fig. 2. CP B-cell lymphomas are IgM+ and show features of MZL. (A) Rep-resentative histologic sections from indicated CP tumors, stained with H&E.(B) Representative FACS analysis on CP220 and CP301 tumors, using anti-bodies against B220 and IgM, Igκ, or Igλ, as indicated. (Left) Normal spleensample. pLN, peripheral lymph nodes; Spl, total spleen.

2936 | www.pnas.org/cgi/doi/10.1073/pnas.1222570110 Gostissa et al.

Dow

nloa

ded

by g

uest

on

Sep

tem

ber

14, 2

020

translocations involving Ig loci or the c-myc oncogene, whichare believed to originate from mistakes in rejoining DSBs gen-erated during CSR and SHM (16, 17). These results are inkeeping with the fact that splenic marginal zone B cells are al-most universally IgM-positive with BCRs enriched for germ-linesequences (39).Given that half of the CP mice analyzed had been immunized

to stimulate T cell–dependent immune responses, our resultssuggest that naive B cells are better targets for transformationinduced by p53 deficiency in the absence of other deficiencies,such as C-NHEJ deficiency. Because p53-dependent respon-ses have been shown to be reduced during the GC reaction byBCL6 (40), it is possible that other p53-independent mecha-nisms are operational in cells undergoing CSR and SHM tosafeguard genomic stability. A previous study used mb1-Cre todelete p53 from early developing B cells and showed that thisresulted in generation of tumors from multiple stages of B-celldifferentiation including pro-B cells and mature B cells (41). Incontrast to our results, a subset of mice from this earlier studydeveloped IgM-negative B-cell lymphomas with S regionrearrangements and T(12;15) translocations. Given that mb1-Cre–mediated deletion of p53 takes place at the progenitor B-cellstage, the precise origin of these tumors remains unclear. More-over, the exact nature of T(12;15) translocations in tumors thatderived from mb1-Cre deletion of p53 were not determined, andit is possible that they originated from V(D)J recombinationbreaks at the IgH locus that persisted from the pro-B-cell stage.Alternatively, lack of p53 from early B-cell development couldhave favored accumulation of other mutations, which in turnallowed survival of cells carrying IgH/c-myc translocations. Itis also possible that development of mb1-Cre–deleted p53-deficient tumors that appear to derive from B cells undergoingCSR was influenced by other factors, including differences inthe housing conditions of the experimental animals, resulting inexposure to different types of antigens, and potential differencesbetween the genetic backgrounds of the experimental animals.Clonal translocations, even those not involving Ig loci, were

present in all but one of the CP tumors as well as in many of theIgM+ tumors from mb1-Cre–deleted p53-deficient mice (41).These translocations were not recurrent and involved differentchromosomes in each tumor. Moreover, most of the tumorscarried multiple, sometimes complex translocations, which thatmade it most unlikely that any single oncogenic event was

responsible for transformation of CP B cells. Such high fre-quency of translocations in CP B-cell tumors may reflect a pro-pensity of B cells to tolerate high levels of DSBs. Indeed, recentwork from our laboratory suggests that translocations in acti-vated cycling B cells are more frequent than in G1-arrested pro-B-cell lines (42, 43). However, metaphases from p53-deficientsplenic B cells stimulated in vitro to undergo CSR do not showhigh levels of chromosomal breaks and translocations, suggestingthe activity of other checkpoint mechanisms (44, 45). It is pos-sible that splenic B cells activated in vitro are not representativeof the population of B cells from which CP tumors arise in vivo.An alternative, nonexclusive explanation is that p53 deficiency inB cells allows accumulation of oncogenic translocations thatoccur at very low levels because of a defective apoptotic response.Indeed, a specific increase in oncogenic IgH/c-myc translocationscan be detected in p53-deficient cultured splenic B cells (26).Moreover, evidence suggests that the proapoptotic functions ofp53 are responsible for protecting against B-cell lymphoma de-velopment, whereas its cell-cycle arrest functions are more im-portant for suppression of T-cell lymphomas, which lack clonaltranslocations (6).Given that organization and function of primary and second-

ary lymphoid organs are significantly different between mice andhumans, it is often difficult to compare B-cell malignancies inthese species (32). Indeed, only few CP tumors closely resemblehuman MZL by histopathologic criteria, with most of themshowing a more diffuse pattern. However, for several charac-teristics, CP tumors might represent a model for human SMZLor other IgM-positive human malignancies, such as Waldenstrommacroglobulinemia. First, somatic p53 inactivation is found in about20–30% of human SMZL. Moreover, although translocationsbetween oncogenes and Ig loci are present in most human B-celllymphomas, splenic MZL or Waldenstrom macroglobulinemiamost often lack these characteristic aberrations (22–24). Thesetumors usually also carry evidence of SHM (22, 23, 46), butmouse SMZLs are not hypermutated (33), again indicating anintrinsic difference between mouse and human B-cell biology.Finally, similarly to human MZL and Waldenstrom macroglob-ulinemia, which are slowly progressing diseases, CP tumors arisewith long latencies, suggesting they may initially develop as in-dolent disease and become more aggressive after accumulationof additional mutations.

A

B C

D

Fig. 3. CP B-cell tumors harbor clonal IgH and Igκ loci rear-rangements, but are not somatically hypermutated. (A) Sche-matic of the IgH and Igκ loci showing restriction sites andprobes used for Southern blot analyses. RI, EcoRI; HIII, HinDIII.(B) Southern blot analysis of DNA from CP tumors indicated ontop demonstrating clonal rearrangements in the JH (Upper)and Jκ (Lower) regions. (C) Southern blot analysis of DNA fromindicated CP tumors with the Iμ probe, which detects rear-rangements in the Sμ region. (B and C), probes and restrictionenzyme used are indicated at the bottom of each panel. Po-sition of the germ-line bands (gl) is shown. DNA from normalspleen was used as control. (D) Table summarizing the resultsof experiments to verify levels of SHM in DNA from CP tumors.The diagram on the top shows the region of the IgH locus usedfor PCR amplification and sequencing.

Gostissa et al. PNAS | February 19, 2013 | vol. 110 | no. 8 | 2937

IMMUNOLO

GY

Dow

nloa

ded

by g

uest

on

Sep

tem

ber

14, 2

020

Materials and MethodsGeneration of CP and CH Mice. CD21-Cre, H2AXF/F, and p53F/F mice have beenpreviously described (9, 27, 47). Because the CD21-Cre transgene deletesnonspecifically more in the female than in the male germ line, CD21-Cre/H2AXF/F or CD21-Cre/p53F/F males were crossed with H2AXF/F and p53F/F

females, respectively, to generate the experimental cohorts. Mice weremaintained in a pathogen-free environment. For immunization, 1 × 108 sheepRBCs resuspended in 100 μL of PBS were injected intraperitoneally; boosterinjections were performed every 2 wk. Induction of a robust GC reaction10 d after immunization was confirmed in selected mice by histological ex-amination of the spleen and by an increase in PNAhi splenic B cells according

to published protocols (30). Experimental animals were monitored for tumorformation and killed for analysis when clear signs of disease appeared.

All animal experiments were performed under protocols approved by theinstitutional Animal Care and Use Committee of Boston Children’s Hospital(Protocol 11-11-2074R).

FACS Analysis. Single-cell suspensions from tumor masses and control organswere stained with CyChrome (CyC)-labeled anti-mouse B220 (eBiosciences),FITC-labeled anti-mouse CD43 (BD Biosciences), and RPE-labeled anti-mouseIgM (Southern Biotech) antibodies or with CyC-labeled anti-mouse B220(eBiosciences), FITC-labeled anti-mouse Igκ (BD Biosciences), and PE-labeledanti-mouse Igλ (BD Biosciences).

C B

A

CP239

CP277

CP220

+

chr8

pai

nt C(14;8;12) n12

+ ch

r14

+ ch

r12

C(14;8;12) n14

T(4;6) n6

CP277

+ ch

r4+

chr6

T(6;19)

+ ch

r4+

chr6

CP752 n6

+ ch

r4+

chr6

T(10;6)

+ ch

r4+

chr6

Fig. 4. CP B-cell tumors harbor clonal, nonrecurrenttranslocations that do not involve Ig loci and c-myc.(A) SKY analysis of selected CP tumors. One repre-sentative metaphase is shown. The arrows indicatechromosomes involved in clonal translocations. De-tailed view of these chromosomes is presented in thepanels at the side, showing DAPI, spectral, andcomputer-classified staining for each chromosome.(B) FISH and chromosome paint analyses on CP tumorscarrying chr 12 and chr 6 translocations. Sequentialhybridization with the set of probes indicated onthe left was performed. Only chromosomes involvedin translocations are shown, with correspondingnormal (n) counterparts. The whole metaphases arepresented in Fig. S1. (C) Northern blot analysis on RNAfrom indicated CP tumors with probes specific for c-myc (Upper) or GAPDH (Lower) as loading control.RNA from normal spleen (norm spl) and from a pro-B-cell tumor with c-myc overexpression (+ control) is in-cluded for comparison.

Table 1. SKY and FISH analysis of CP tumors

SKY FISH

Tumor no. Clonal tr. Nonclonal tr. IgH Igκ Igλ TCRα/δ TCRβ

220 T(2:5), T(17;4), C(8;14;12) Normal ND ND C(14;8;12) ND277 T(4;6), T(6;19), C(5;3;5) T(17;10, TX;3) ND Normal ND ND Duplicated166 T(15;1), T(10;9), T(13;10) T(1;15), T(10;13)239 T(3;1), T(4;3), T(17;2)246 No translocations569 T(10;5), T(14;10) ND ND ND Normal ND752 T(10;6), T(9;1), T(15;16), T(14;1), T(11;14) ND Normal Normal Normal Normal

At least 10 metaphases were analyzed per each sample. Clonal translocations (tr.) are present in more than 50% of metaphases analyzed; nonclonaltranslocation are present in less than 50% of metaphases analyzed. ND, not done.

2938 | www.pnas.org/cgi/doi/10.1073/pnas.1222570110 Gostissa et al.

Dow

nloa

ded

by g

uest

on

Sep

tem

ber

14, 2

020

Data acquisition was performed on a FACSCalibur flow cytometer equippedwith CellQuest software (Becton Dickinson). Analysis was performed withFlowJo software (Tree Star).

Histological Analysis. Tumor tissues were fixed in 10% (vol/vol) bufferedformalin and stored in 70% (vol/vol) ethanol. Paraffin-embedded tissueswere sectioned and stained with H&E. Histologic diagnoses were made onthe basis of established criteria (32).

Southern Blotting. Genomic DNA isolated from tumor masses or normalcontrol tissues were separated on 0.8% agarose gel and transferred to Zeta-Probe GT (Biorad) nylon membrane. Hybridization was performed in 50%(vol/vol) formamide/SScPE at 42 °C. The 5′ p53 probe was a 600-bp XbaIfragment upstream of p53 gene exon1 and has been described previously.The JH4–3 probe was a 1.6-kb HindIII/EcoRI fragment downstream of JH4;the Jκ probe was a 1-kb fragment downstream of Jκ5; and the Iμ probe wasa 1.2-kb XbaI/PstI fragment encompassing part of the Iμ promoter. The c-mycprobe used for Northern blot was generated by PCR-amplifying c-myc exon2.

Somatic Hypermutation Analysis. The genomic region encompassing JH1 to JH4and part of the intron downstream of JH4 was PCR-amplified from tumorDNA using degenerate oligonucleotides corresponding to the different VH

families (48) as forward primers and oligonucleotides downstream of JH4(5′AGGCTCTGAGATCCCTAGACAG3′ or 5′CCTCTCCAGTTTCGGCTGATCC3′)as reverse primers. Proofreading polymerase (iProof High-Fidelity DNA Poly-

merase, Biorad) was used for amplification and PCR conditions were aspreviously published (48). Amplification products were purified from aga-rose gel and submitted to sequencing. Sequences were compared with thepublished 129/Sv and C57/B6 sequences (accession nos. NT_114985.2 andNT_166318.1, respectively). PCR amplification and sequencing was repeatedtwo or three times for each sample.

Metaphase Preparation, SKY, and FISH. Tumor cell suspensions were culturedfor overnight and Colcemid (KaryoMAX Colcemid Solution; GIBCO) wasadded at the final concentration of 50 ng/mL for 3–5 h. Metaphase spreadswere prepared according to standard protocols (44). Spectral karyotypingwas performed with a mouse SKY paint kit (Applied Spectral Imaging) fol-lowing manufacturer’s indications. Images were acquired with BX61 Micro-scope (Olympus) equipped with a motorized automatic stage, a cooled-CCDcamera, and an interferometer (Applied Spectral Imaging). A 63× objectivewas used. Analysis was performed with the HiSKY software (Applied SpectralImaging). At least 15 metaphases per each sample were analyzed.

ACKNOWLEDGMENTS. We thank Roberto Chiarle for helpful suggestions.This work was supported by National Institutes of Health Grants5P01CA92625 and CA098285 and a Leukemia and Lymphoma Society ofAmerica (LLS) Specialized Center of Research grant (to F.W.A.). This workwas supported in part by the Intramural Research Program of the NationalInstitutes of Health, National Institute of Allergy and Infectious Diseases (toH.C.M.). M.G. was an LLS senior fellow. C.H.B. is an LLS Scholar. F.W.A. is aninvestigator of the Howard Hughes Medical Institute.

1. Vousden KH, Prives C (2009) Blinded by the light: The growing complexity of p53. Cell137(3):413–431.

2. Brady CA, Attardi LD (2010) p53 at a glance. J Cell Sci 123(Pt 15):2527–2532.3. Böhlig L, Rother K (2011) One function—multiple mechanisms: The manifold activities

of p53 as a transcriptional repressor. J Biomed Biotechnol 2011:464916.4. Palmero EI, Achatz MI, Ashton-Prolla P, Olivier M, Hainaut P (2010) Tumor protein 53

mutations and inherited cancer: Beyond Li-Fraumeni syndrome. Curr Opin Oncol 22(1):64–69.

5. Olivier M, Hollstein M, Hainaut P (2010) TP53 mutations in human cancers: Origins,consequences, and clinical use. Cold Spring Harb Perspect Biol 2(1):a001008.

6. Attardi LD, Donehower LA (2005) Probing p53 biological functions through the use ofgenetically engineered mouse models. Mutat Res 576(1-2):4–21.

7. Donehower LA, et al. (1995) Effects of genetic background on tumorigenesis in p53-deficient mice. Mol Carcinog 14(1):16–22.

8. Ward JM, et al. (1999) Splenic marginal zone B-cell and thymic T-cell lymphomas inp53-deficient mice. Lab Invest 79(1):3–14.

9. Jonkers J, et al. (2001) Synergistic tumor suppressor activity of BRCA2 and p53 ina conditional mouse model for breast cancer. Nat Genet 29(4):418–425.

10. Liu X, et al. (2007) Somatic loss of BRCA1 and p53 in mice induces mammary tumorswith features of human BRCA1-mutated basal-like breast cancer. Proc Natl Acad SciUSA 104(29):12111–12116.

11. Berman SD, et al. (2008) Metastatic osteosarcoma induced by inactivation of Rb andp53 in the osteoblast lineage. Proc Natl Acad Sci USA 105(33):11851–11856.

12. Walkley CR, et al. (2008) Conditional mouse osteosarcoma, dependent on p53 loss andpotentiated by loss of Rb, mimics the human disease. Genes Dev 22(12):1662–1676.

13. Quinn BA, et al. (2009) Induction of ovarian leiomyosarcomas in mice by conditionalinactivation of Brca1 and p53. PLoS ONE 4(12):e8404.

14. Zhou Z, et al. (2006) Synergy of p53 and Rb deficiency in a conditional mouse modelfor metastatic prostate cancer. Cancer Res 66(16):7889–7898.

15. Küppers R, Dalla-Favera R (2001) Mechanisms of chromosomal translocations in B celllymphomas. Oncogene 20(40):5580–5594.

16. Gostissa M, Alt FW, Chiarle R (2011) Mechanisms that promote and suppress chro-mosomal translocations in lymphocytes. Annu Rev Immunol 29:319–350.

17. Zhang Y, et al. (2010) The role of mechanistic factors in promoting chromosomaltranslocations found in lymphoid and other cancers. Adv Immunol 106:93–133.

18. Di Noia JM, Neuberger MS (2007) Molecular mechanisms of antibody somatic hy-permutation. Annu Rev Biochem 76:1–22.

19. Maul RW, Gearhart PJ (2010) AID and somatic hypermutation. Adv Immunol105:159–191.

20. MacLennan ICM, et al. (2003) Extrafollicular antibody responses. Immunol Rev194:8–18.

21. Küppers R, Klein U, Hansmann ML, Rajewsky K (1999) Cellular origin of human B-celllymphomas. N Engl J Med 341(20):1520–1529.

22. Issa GC, Leblebjian H, Roccaro AM, Ghobrial IM (2011) New insights into the patho-genesis and treatment of Waldenstrommacroglobulinemia. Curr Opin Hematol 18(4):260–265.

23. Oscier D, Owen R, Johnson S (2005) Splenic marginal zone lymphoma. Blood Rev19(1):39–51.

24. Traverse-Glehen A, Baseggio L, Salles G, Felman P, Berger F (2011) Splenic marginalzone B-cell lymphoma: A distinct clinicopathological and molecular entity. Recentadvances in ontogeny and classification. Curr Opin Oncol 23(5):441–448.

25. Liao MJ, et al. (1998) No requirement for V(D)J recombination in p53-deficient thymiclymphoma. Mol Cell Biol 18(6):3495–3501.

26. Ramiro AR, et al. (2006) Role of genomic instability and p53 in AID-induced c-myc-Ightranslocations. Nature 440(7080):105–109.

27. Kraus M, Alimzhanov MB, Rajewsky N, Rajewsky K (2004) Survival of resting mature Blymphocytes depends on BCR signaling via the Igalpha/beta heterodimer. Cell 117(6):787–800.

28. Bassing CH, et al. (2003) Histone H2AX: A dosage-dependent suppressor of oncogenictranslocations and tumors. Cell 114(3):359–370.

29. Celeste A, et al. (2003) H2AX haploinsufficiency modifies genomic stability and tumorsusceptibility. Cell 114(3):371–383.

30. Shinall SM, Gonzalez-Fernandez M, Noelle RJ, Waldschmidt TJ (2000) Identification ofmurine germinal center B cell subsets defined by the expression of surface isotypesand differentiation antigens. J Immunol 164(11):5729–5738.

31. Jacks T, et al. (1994) Tumor spectrum analysis in p53-mutant mice. Curr Biol 4(1):1–7.32. Morse HC, 3rd, et al.; Hematopathology subcommittee of the Mouse Models of Hu-

man Cancers Consortium (2002) Bethesda proposals for classification of lymphoidneoplasms in mice. Blood 100(1):246–258.

33. Fredrickson TN, Lennert K, Chattopadhyay SK, Morse HC, 3rd, Hartley JW (1999)Splenic marginal zone lymphomas of mice. Am J Pathol 154(3):805–812.

34. Chiang YJ, Difilippantonio MJ, Tessarollo L, Morse HC III, Hodes RJ (2012) Exon 1disruption alters tissue-specific expression of mouse p53 and results in selective de-velopment of B cell lymphomas. PLoS ONE 7(11):e49305.

35. Aylon Y, Oren M (2011) p53: Guardian of ploidy. Mol Oncol 5(4):315–323.36. Zhu C, et al. (2002) Unrepaired DNA breaks in p53-deficient cells lead to oncogenic

gene amplification subsequent to translocations. Cell 109(7):811–821.37. Wang JH, et al. (2008) Oncogenic transformation in the absence of Xrcc4 targets

peripheral B cells that have undergone editing and switching. J Exp Med 205(13):3079–3090.

38. Gostissa M, et al. (2009) Long-range oncogenic activation of Igh-c-myc translocationsby the Igh 3′ regulatory region. Nature 462(7274):803–807.

39. Martin F, Kearney JF (2000) B-cell subsets and the mature preimmune repertoire.Marginal zone and B1 B cells as part of a “natural immune memory.” Immunol Rev175:70–79.

40. Phan RT, Dalla-Favera R (2004) The BCL6 proto-oncogene suppresses p53 expression ingerminal-centre B cells. Nature 432(7017):635–639.

41. Rowh MA, et al. (2011) Tp53 deletion in B lineage cells predisposes mice to lym-phomas with oncogenic translocations. Oncogene 30(47):4757–4764.

42. Chiarle R, et al. (2011) Genome-wide translocation sequencing reveals mechanisms ofchromosome breaks and rearrangements in B cells. Cell 147(1):107–119.

43. Zhang Y, et al. (2012) Spatial organization of the mouse genome and its role in re-current chromosomal translocations. Cell 148(5):908–921.

44. Franco S, et al. (2006) H2AX prevents DNA breaks from progressing to chromosomebreaks and translocations. Mol Cell 21(2):201–214.

45. Franco S, et al. (2008) DNA-PKcs and Artemis function in the end-joining phase ofimmunoglobulin heavy chain class switch recombination. J Exp Med 205(3):557–564.

46. Hockley SL, et al. (2010) Insight into the molecular pathogenesis of hairy cell leu-kaemia, hairy cell leukaemia variant and splenic marginal zone lymphoma, providedby the analysis of their IGH rearrangements and somatic hypermutation patterns. BrJ Haematol 148(4):666–669.

47. Bassing CH, et al. (2002) Increased ionizing radiation sensitivity and genomic in-stability in the absence of histone H2AX. Proc Natl Acad Sci USA 99(12):8173–8178.

48. Ehlich A, Martin V, Muller W, Rajewsky K (1994) Analysis of the B-cell progenitorcompartment at the level of single cells. Curr Biol 4(7):573–583.

Gostissa et al. PNAS | February 19, 2013 | vol. 110 | no. 8 | 2939

IMMUNOLO

GY

Dow

nloa

ded

by g

uest

on

Sep

tem

ber

14, 2

020