Male and female mice were used in all experiments. Mice analyzed in this study were at least 4 weeks old and were kept under specific-pathogen-free conditions. All animal procedures conformed to EU Directive 2010/63EU and Recommendation 2007/526/EC regarding the protection of animals used for experimental and other scientific purposes, enacted in Spanish law under RD 53/2013. The procedures were reviewed by the CNIC Institutional Animal Care and Use Committee and approved by the Community of Madrid.

AicdaCre^+/ki mice (15, 21) and Rosa26tdTom^+/ki mice were obtained from the Jackson Laboratory (007770 and 007909) and were crossed to generate AicdaCre^+/ki;R26tdTom^+/ki mice. λ-MYC mice (λ-MYC^+/TG ) (20) were then bred to AicdaCre^+/ki;R26tdTom^+/ki mice to generate λ-MYC^+/TG ;AicdaCre^+/ki;R26tdTom^+/ki mice, hereafter λ-MYC Tom^+/ki mice. For the kinetics experiments, we included only pre-tumor stage mice without signs of disease. For all other experiments, mice were monitored twice a week for lymphoma development until visible signs of disease appeared. Mice were sacrificed when moribund, and the spleen and lymph nodes were collected and analyzed. We define tumor cells as those originated from sick mice harboring severe enlargement of lymph nodes with massive B cell infiltration.

Mice were sacrificed, and spleen, lymph nodes, femurs, and tibias were extracted and conserved on ice in PBS supplemented with 2% FBS. Cell suspensions were generated, depleted of erythrocytes by incubation with ACK Lysing Buffer (Lonza, 10-548E), and then incubated with anti-mouse CD16/CD32 antibodies (eBiosciences) to block Fc receptors. Cells were stained with combinations of the following fluorochrome-conjugated antibodies: B220-BV421/FITC (RA3-6B2, BD Horizon), CD19-APCeF780 (ID3, eBiosciences), CD25-APC (PC61.5, Tonbo Biosciences), GL7-FITC (BD GL-7, Pharmingen), GL7-PerCP-CY5.5 (GL-7, Biolegend), IgD-Biotin (11-26, Southern Biotech), IgD-FITC (11-26c.2a, BD Pharmingen), IgM-APC/FITC (IL/41, BD Pharmingen), ST-BV421/PE-CY7 (BD Biosciences) Finally, a viability marker was added to the labeled cell suspension.

Intracellular stainings were performed using the Foxp3/Transcription Factor Staining Buffer Set (eBioscience, 00-5523-00). After FC blocking, cell-surface antigens were stained with selected fluorochrome-conjugated Abs and the viability marker. Cells were additionally fixed with 1% paraformaldehyde for 40 minutes on ice to prevent leakage of Tomato fluorescent proteins. Subsequently, cells were fixed and permeabilized according to the manufacturer’s standard protocol, and stained intracellularly with c-MYC (D84C12, Cell Signaling) or KI67 (SP6, Abcam) and GARAB-AF647 (Life Technologies) secondary antibody. Samples were acquired with LSRFortessa or FACSCanto instruments (BD Biosciences), and data were analyzed with FlowJo V10.4.2 software.

Bone marrow (BM) and spleens were harvested from four 8-week-old mice (two λ-MYC and two WT controls), and RNA was extracted from FACS-sorted BM populations (Pre-Pro-B [B220+CD19-], Pro-B [B220+CD19+IgD-IgM-CD25-], Pre-B [B220+CD19+IgD-IgM-CD25+], and immature B cells [B220+CD19+IgD-IgM+CD25-]) and from recirculating splenic B cells obtained by immunomagnetic negative selection with CD43 MACS Microbeads (Miltenyi Biotec-CD43 (Ly-48) Microbeads Mouse). cDNA was synthesized using the High-Capacity cDNA Reverse Transcription Kit (Applied Biosystems) and quantified by SYBR green assay (Applied Biosystems) with normalization to GAPDH expression. The following primers were used: GAPDH (forward) 5′-TGA AGC AGG CAT CTG AGG G-3′, (reverse) 5′-CGA AGG TGG AAG AGT GGG AG-3′; mouse c-Myc (forward) 5’ GAG CTC CTC GAG CTG TTT GA 3’, (reverse) 5’ AGT TCA CGT TGA GGG GCA TC 3’; human c-MYC (forward) 5’ GGA CCC GCT TCT CTG AAA GG 3’, (reverse) 5’ GAG GCT GCT GGT TTT CCA CTA 3’.

Two λ-MYC Tom^+/ki mice harboring IgM+Tom+ tumors were analyzed by RNA-sequencing. From each mouse, B220+IgM+Tom+ (Tom+) and B220+IgM+Tom- (Tom-) splenocytes were purified using a Sy3200 Cell Sorter (SonyBiotechnology). Total RNA was extracted using the PicoPureTM RNA Isolation Kit (ThermoFisher Scientific), and barcoded RNA-seq libraries were generated using the NEBNext RNA Library preparation kit (Illumina). Library size was measured using the Agilent 2100 Bioanalyzer DNA 1000 chip. Libraries were sequenced with a HiSeq2500 sequencer (Illumina) to generate 60-base single reads. FastQ files for each sample were obtained using CASAVA v1.8 software (Illumina). Sequencing reads were pre-processed with a pipeline that used FastQC (22) to assess read quality and Cutadapt (23) to trim sequencing reads (eliminating Illumina adaptor remains) and to discard reads < 30 bp. The resulting reads were mapped against the mouse transcriptome (GRCm38, release 76) and quantified using Salmon algorithm (24). Data were then processed with a differential expression analysis pipeline that used Bioconductor package EdgeR (25) for normalization and differential expression testing of paired samples with quasi-likelihood F-test (QLF). Differentially expressed genes (DEGs) between Tom+ and Tom- were selected based on an adjusted P-value < 0.05. Pathway enrichment analysis was done with Ingenuity Pathway Analysis with DEGs with an absolute FC>1.2. Hierarchical clustering was performed using the average linkage method based on Euclidian distance (Genesis 1.7.7) and the significantly DEGs were identified by ANOVA (26). GSEA was performed with pre-ranked log fold-change values between Tom+ and Tom- lymphoma cells and the gene set PASQUALUCCI_LYMPHOMA_BY_GC_STAGE_DN (17).

Data are presented as means ± standard deviation. Statistical analysis was performed in GraphPad Prism (version 8.0.2 for Windows, GraphPad Software, La Jolla California USA, www.graphpad.com). Normality of the data was assessed with the D’Agostino & Pearson normality test. For data with a Gaussian distribution, two-tailed unpaired Student t test was used when comparing two experimental groups and one-way ANOVA when comparing three experimental groups. For survival analysis, Kaplan-Meier curves were generated and compared using the log-rank (Mantel-Cox) test. Differences were considered statistically significant at P < 0.05.

To study the role of AID in MYC-driven lymphomas, we made use of the λ-MYC model, a transgenic mouse model of MYC-driven lymphomagenesis where a translocated MYC gene from a human BL line was placed under the control of λ chain regulatory sequences (20). To evaluate the effect of c-MYC overexpression in this mouse model, we first studied B cell development in the bone marrow (BM) of young, pre-tumor stage λ-MYC mice, defined as mice that do not show tumoral masses or signs of disease ( Figure 1 ). WT littermates (λ-MYC^+/+ ) were used as controls (hereafter, controls). At 4-8 weeks of age, λ-MYC mice had a significantly higher proportion of B220+ B cells, but this difference was not seen in older mice ( Figure 1A ). At 8 weeks of age, λ-MYC mice had a markedly elevated proportion of Pro-B cells, whereas all other populations were underrepresented ( Figure 1B ). Very similar differences were found when absolute numbers, rather than proportions, were assessed ( Figure 1C ), indicating that B cell differentiation is severely altered in λ-MYC mice.

To further characterize BM differentiation in pre-tumor stage λ-MYC mice, we analyzed the expression of c-MYC protein and RNA ( Figures 2A-C ). The proportion of c-MYC-expressing B cells was 5-fold higher in λ-MYC BM, as measured by flow cytometry ( Figure 2A ). c-MYC overexpression was detectable as early as the Pre-Pro-B cell stage, and the proportion of c-MYC expressing cells differed among the different B cell differentiation subsets ( Figure 2B ). RT-qPCR analysis detected c-MYC transgene expression from the Pre-Pro-B cell stage, with a near constant expression level at all differentiation stages in the BM ( Figure 2C ). In contrast, transcription of endogenous c-Myc in control mice showed a transient sharp decline at the Pre-B cell stage ( Figure 2C ). Thus, transgenic c-MYC overexpression occurs very early and is sustained in BM B cell precursors of λ-MYC mice, while endogenous c-MYC protein is regulated during B cell differentiation.

We next analyzed cell death and proliferation during early B cell differentiation in pre-tumor stage λ-MYC mice. Pro-B cells from λ-MYC mice showed increased survival compared with controls ( Figure 2D ), but the proportion of proliferating Pro-B cells was unaltered ( Figure 2E ). In contrast, viability and proliferation were both sharply reduced in λ-MYC Pre-B cells ( Figures 2D, E ). B cell survival was also compromised at immature and recirculating B cell subsets ( Figure 2D ) of λ-MYC mice. These data indicate that pre-tumor stage λ-MYC mice have dramatic alterations in B cell differentiation, at least in part due to a strong block at the Pro-B cell stage. This block associates with increased survival of Pro-B cells, together with compromised survival at later developmental stages and reduced Pre-B cell proliferation.

We next analyzed the effect of c-MYC overexpression on peripheral B cell populations. The splenic B cell population in pre-tumor stage λ-MYC mice was significantly reduced relative to controls, both as a proportion of the total splenocytes ( Figures 3A ; S1A ) and in terms of absolute numbers ( Figure 3B ). The lymphopenia in λ-MYC mice was accompanied by a lack of IgM and IgD expression in a considerable proportion of splenic B220+ cells (hereafter, Ig- cells), with the Ig- population being more prominent at 4 weeks of age and decreasing thereafter ( Figures 3C ; S1B ). In 4-week-old λ-MYC mice, all Ig- splenic B cells expressed either GL7 or CD43 ( Figure S2A ), markers of distinct subsets during early B cell differentiation in the BM (27, 28). These results indicate that early B cell progenitors accumulate in the spleens of young λ-MYC mice.

Flow cytometry analysis of c-MYC expression in peripheral B cells identified two clearly distinct c-MYC+ cell populations in 8-week-old λ-MYC mice that differed in the surface expression of IgM/ (MYC+Ig- and MYC+Ig+), whereas wild type controls were lacking these populations ( Figure 3D ). The c-MYC+ population accounted for about 10% of total B cells in the spleen ( Figure 3E ) and close to 5% in the lymph nodes ( Figure S2B ), with c-MYC expression more prevalent in the Ig- splenic B cell population ( Figures 3E ; S2B ).

To gain insight into the phenotypic differences between splenic B cells of λ-MYC and control mice, we analyzed cell death and proliferation by flow cytometry ( Figures 3F, G ). In contrast to BM ( Figure 2D ), there were no between-genotype differences in B cell survival in the spleen ( Figure 3F ). However, the proportion of proliferating Ig- cells was significantly higher in λ-MYC mice ( Figure 3G ), consistently with the idea that they contain a population of BM precursors. Together, these data indicate that deregulated c-MYC expression in λ-MYC mice leads to peripheral B cell lymphopenia resulting from defects in B cell generation at two different levels: impaired BM differentiation and aberrant migration of B cell precursors from the BM.

To assess the specific role of AID in MYC-driven lymphomagenesis, we combined the λ-MYC lymphoma model with the AicdaCre^+/ki and R26tdTom^+/ki alleles (15, 21) ( Figure S3A ). The resulting AicdaCre^+/ki;R26tdTom^+/ki; λ-Myc^+/TG mice (hereafter λ-MYC Tom^+/ki) allowed genetic tracing of cells that had expressed AID throughout MYC-driven lymphoma generation. As controls, we used AicdaCre^+/ki;R26tdTom^+/ki; λ-Myc^+/+ littermates (hereafter, controls). Flow cytometry analysis showed the presence of Tom+ cells in the spleen and lymph nodes of pre-tumor stage λ-MYC Tom^+/ki mice and littermate controls ( Figure S3B ).

Flow cytometry analysis of splenocytes from sick λ-MYC Tom^+/ki mice revealed tumor heterogeneity, with one group of animals harboring tumors dominated by Ig- B cells (Ig- tumors), whereas others had tumors containing mostly Ig+ cells (Ig+ tumors) ( Figure 4A ). Ig- tumors presumably originated from early B cell precursors, and these mice had shorter survival, indicating a more aggressive phenotype ( Figure 4B ). Approximately one third of the Ig+ tumors (6/17) contained a fraction of AID-experienced cells, detected by Tomato expression (Ig+Tom+ tumors; >10% Tom+ B cells) ( Figure 4C ). The mice with Ig+Tom+ tumors had better overall survival ( Figure 4D ), suggesting that AID expression is associated with better prognosis in λ-MYC mice.

To characterize the impact of AID expression on lymphoma B cells, we performed RNA-sequencing of IgM+Tom+ (Tom+) and IgM+Tom- (Tom-) B lymphoma cells isolated by flow cytometry from two tumors from two independent mice (tumors #1 and #2) ( Figures 5A, B ). Differential gene expression analysis between the Tom+ and Tom- samples was then used as the input for hierarchical clustering ( Figure 5C ) and pathway enrichment analysis ( Figure 5D ) with Ingenuity Pathway Analysis (IPA) software.

Pathways significantly activated in Tom+ cells included the apoptotic signaling pathway and cell death receptor signaling ( Figure 5D ). The phosphatidyl-inositol tri-phosphate (PI3K) signaling pathway—an important signaling pathway activated in BL (29)—was activated in Tom+ cells. In addition, the negative regulator of the PI3K pathway PTEN was significantly deregulated. Other important signaling pathways such as ERK/MAPK were significantly activated in Tom+ cells. Likewise, B cell receptor signaling, and B cell development pathways were active in Tom+ cells. On the other hand, PAK signaling pathway was inhibited in Tom+ lymphoma cells compared to Tom- cells.

To further study the molecular identity of Tom+ and Tom- cells, we performed GSEA analysis using a set of genes that were downregulated in tumors from λ-MYC-IμBCL6 mice, defined as tumors of GC or post-GC origin, compared to tumors from λ-MYC, defined as pre-GC-origin tumors (17). Interestingly, these genes were significantly downregulated in Tom+ cells compared to Tom- cells ( Figure 5E ), indicating that Tom+ cells are transcriptionally similar to GC-derived tumors. Together, these results indicate that AID expression alone determines the transcriptional identity of the tumor cells with a hallmark signature distinctive of mature subtypes of B cell lymphomas.