tet-LIN28B mice were generated by intercrossing Col1a^tetO-LIN28B (JAX: #023911) mice, carrying Doxycyclin (Dox) inducible Flag-tagged human Lin28b cDNA under the endogenous Col1a promoter, to Rosa26^rtTA*m2 (JAX: #006965) mice and were originally obtained from the laboratory of George Daley (Harvard Medical School) (33). Monoallelic transgene expressing mice were used for experiments. Rosa26^rtTA*m2 heterozygous littermates were used as control mice. For in vivo experiments, Doxycycline (Dox) chow (200 mg/kg, Ssniff, cat#A112D70203) was fed to tet-LIN28B and littermates for 10 days before analysis. B6.Cg-Lin28b^tm1.1Gqda /J (Lin28b KO) (JAX: #023917), C57BLKS-Rpl24^Bst /J (Bst) (JAX: #000516) and Nur77-GFP (Jax #016617) were from the Jackson laboratory. Wildtype mice were from Taconic (B6NTAC, Taconic). The Lin28b-eGFP knock-in strain was generated as previously described (34). Adult mice were used at 10 to 16 weeks of age, neonates were analyzed at the indicated days post-birth. Male and female mice were used interchangeably throughout the experiments. Animal husbandry and experimental procedures were performed in accordance with ethical permits issued by the Swedish Board of Agriculture.

Adult bone marrow B cell precursors from untreated adult mice were expanded and enriched ex vivo, as previously described (35). Briefly, single cell suspension of red blood cell lysed bone marrow cells from untreated adult mice were plated in complete RPMI-1640 (Fischer Scientific, cat#21875091) (supplemented with 10% FBS, HEPES, NEAA, Sodium Pyruvate, Penstrep, 2-mercaptoethanol) and incubated for 15 min at 37 °C. The non-adherent cells were isolated and seeded at a cell density of 0.75-1.5 million/mL using complete RPMI-1640, IL-7 (20 ng/mL, Peprotech, cat#217-17), and Dox (0.1 µg/ml, Sigma-Aldrich, cat#D3072). Note that Dox treatment was always initiated ex vivo at the start of the culture. The cell culture medium was refreshed every second day.

Cells from day 5 bone marrow cultures from tet-LIN28B or littermate control mice were used for lysis and RIP using Magna RIP RNA-Binding Protein Immunoprecipitation kit (Millipore, cat#17-700) according to the manufacturer’s instructions. Briefly, 10% of the total lysate was kept as input sample. 4 μg of anti-Flag antibody (Sigma) and 50 μl of A/G beads (Millipore, cat#16-663) were added to the lysate and incubated for 3 hours. Beads were washed 6 times with ice-cold RIP wash buffer (Millipore, cat#17-700). RNA was extracted using RNAzol (Sigma-Aldrich, cat#R4533) and Direct-zol RNA MicroPrep kit (Zymo Research, cat#R2060). For RNA quantification and quality control respectively, Qubit with RNA HS kit (Thermofisher Scientific, cat# Q32852) and Agilent Bioanalyzer with RNA Pico kit (Agilent, cat#5067-1513) were used according to the manufacturer’s instructions. 20ng of total RNA was used for the cDNA synthesis and amplification with the SMART-Seq v4 Ultra Low Input RNA Kit for Sequencing (Takara, cat#634889) according to instructions. Library tagmentation and indexing was performed using Nextera XT DNA Kit (Illumina, cat#FC-121-1030) and Nextera XT v2 set A Indexing kit (Illumina, cat#15052163). Library quantification and library size were measured using Qubit DNA HS kit (Thermofisher Scientific, cat#Q32851) and BioAnalyzer DNA HS kit (Agilent, cat#5067-4626). The samples were run in NextSeq 550 using NextSeq High output 150 cycles 2x76bp.

The sequencing was aligned to mm10 using the STAR aligner (v2.7.1) and counted using RSEM (v1.28). After mapping, the RSEM result files were loaded into R and their FPKM values were used for the analysis. To filter out lowly expressed genes, the initial data (covering 54,752 genes) was filtered for genes that had at least 50 FPKM value in the input and IP of both tet-Lin28b samples (1,459 genes), which were then used for all the analysis presented in Figure 1 . The individual sample FPKM values before and after filtering are included below ( Table 1 ).

Enrichment was calculated as IP over Input FPKM for each of the two tet-LIN28B samples, individually. 238 transcripts had an average enrichment of at least 2-fold and were considered for subsequent analyses. Littermate control sample was not considered when generating this list. Gene ontology analysis was done using the Panther database (pantherdb.org, release 17.0). Comparable results were obtained using the KEGG pathway and BioPlanet. All the sequencing data is accessible through GEO, GSE217788.

B cell precursor subsets were FACS sorted from adult BM of Dox induced mice of the indicated genotypes. RNA sequencing libraries were generated as previously described (19) and sequenced on Illumina NextSeq and NovaSeq. Note that FSC-A low, small Pro-B cells were excluded from the FACS sorting gate. All data was processed and analyzed as previously described (19). Data is accessible through GSE217788 except for Imm-B cells that was previously generated (19) and is accessible through GSE135603. GSEA used Wald's stat generated by DESeq2 as a ranking metric. MSigDB Mouse collections Hallmark_MYC_target_v1.

Cells were depleted of endogenous methionine by pre-culturing in methionine free RPMI (Fisher Scientific, cat#A1451701). After 45 minutes, the medium was supplemented with L-azidhomoalanine (AHA, 1 mM final concentration, Fisher Scientific, cat#C10102) and labeled for one hour. AHA uptake was visualized using Click-iT Plus Alexa Fluor 555 Picolyl Azide Toolkit (Fisher Scientific, cat#C10642), following the manufacturer’s protocol. Briefly, up to 3 x 10⁶ cells were fixed in 400 µL of 4% formaldehyde (Fischer Scientific, cat#28906) for 20 minutes at room temperature, washed, and then permeabilized in equal volume of 0.1% saponin (Sigma, cat#S4521) for 20 minutes. Alexa Fluor 488 Alkyne (Fisher Scientific, cat#A10267) was used in place of Alexa Fluor 555 for FACS analysis.

Analysis was performed in accordance with the manufacturer’s instructions (Invitrogen, cat#C10634). Briefly, single cell suspension of red blood cell lysed bone marrow cells was plated in complete RPMI-1640 (Fischer Scientific, cat#21875091) and labeled with EdU (10 µM) for 30 minutes. Cells were then stained for surface proteins before fixation and permeabilization. The EdU was Click-iT ligated to Alexa Fluor 555 and analyzed with DAPI for separation according to cell cycle phase.

Bone marrow cells were extracted by crushing bones from hind- and front limbs, hips, and sternum, using mortar and pestle. Peritoneal cavity was isolated by flushing adult mice with 8 mL of FACS buffer (Hank’s Balanced Salt Solution (Gibco, cat#14175-053) supplemented with 0.5% Bovine Serum Albumin (Sigma, cat#A9647) and 2 mM EDTA (Invitrogen, cat#15575-038)), or neonates 1-8 mL depending on the age. Red blood cell lysis was performed on bone marrow and spleens using ACK lysis buffer (Fischer Scientific, cat#A1049201), only for mice that were at least 19 days old. Lineage depletion was carried out before cell sorting by MACS Cell Separation and LS columns (Miltenyi Biotec, cat#130-042-401) to deplete CD3+, TER119+, or Gr-1+ cells, according to manufacturer’s instructions.

Antibody staining of surface antigen was performed using FACS buffer, at a cell density of 1 x 10⁷ cells per 100 µL of FACS buffer with antibodies, for 30 minutes at 4 °C. Analysis of intracellular antigens was done by using the same fixation and permeabilization as was described under “Metabolic labeling of nascent proteins” or using BD Cytofix/Cytoperm fixation/permeabilization kit (BD Biosciences, cat#554714). Maximum of 3 x 10⁶ cells per reaction. All flow cytometry experiments were gated on live CD19+ B cells and the following immunophenotypes: Pro-B cells (CD93+IGM–cKIT+), Pre-B cells (CD93+IGM–cKIT– and FSC-A to separate large from small), Imm-B cells (CD93+IGM+), peritoneal cavity B-1a B cells (CD5+CD23–B220low and CD43+), and peritoneal cavity B-2 B cells (B220high CD23+ and CD43–). All flow cytometry experiments were performed at the Lund Stem Cell Center FACS Core Facility (Lund University, Lund, Sweden) using BD Fortessa, BD Fortessa X20, BD FACS Aria III, or BD FACS Aria II instruments. Cell sorting was performed using a 70 µm nozzle, 0.32.0 precision mask, and a flow rate below 6000 events/s. Analysis was performed using FlowJo version 10. Antibodies are listed in Supplemental Table S1 .

4000 bone marrow Lineage^– Sca1⁺ cKit⁺ (LSK) donor cells (CD45.1) from day 10 neonatal Rpl24^WT/Bst or WT mice were FACS sorted and mixed at a 1:2 ratio with 8000 sorted CD45.1 competitor LSK cells and 200,000 CD45.1 total BM support cells. Cells were transplanted into lethally irradiated (2 × 450 cGy) and CD45.1+ CD45.2+ double positive adult recipients by tail vein injection. Endpoint analysis was performed 13–14 weeks after transplantation.

Sorted B cells were lysed in RIPA buffer (CST, cat#9806S), protease inhibitor (Roche, cat#4693132001), and PMSF (CST, cat#8553S). Loading samples were prepared with Laemmli buffer (Bio-Rad, cat#1610747) with β-mercaptoethanol (Bio-Rad, cat#1610710). Proteins were separated by SDS–polyacrylamide gel electrophoresis (Any kD Mini-Protean TGX gel, Bio-Rad, cat#4568125) and transferred to a nitrocellulose membrane (Trans-Blot Turbo Mini Nitrocellulose Transfer Packs, Bio-Rad, cat#1704158). Membranes were blocked for 1 hour at RT in 5% skimmed milk. Antibodies are listed in Supplemental Table S1 . The ECL Select kit (GE Healthcare, cat#RPN2235) was used to detect protein bands with a ChemiDoc XRS+ system (Bio-Rad). Band intensity was quantified by Image Lab Software v6.1(BioRad). Each lane was automatically detected with manual adjustments to subtract the background and each band was individually selected followed by quantification of peak area of obtained histograms.

For RIPseq, 10% of the each indicated protein extract was collected to validate the immunoprecipitation efficiency. Proteins were separated by SDS–polyacrylamide gel electrophoresis (12% Mini-Protean TGX gel, Bio-Rad, cat#4561043) and transferred to a nitrocellulose membrane (Trans-Blot Turbo Mini Nitrocellulose Transfer Packs, Bio-Rad, cat#1704158). Membranes were blocked for 1 hour at RT in 5% skimmed milk. Antibodies are listed in Supplemental Table S1 .

50.000 sorted cells were lysed in RNAzol (Sigma-Aldrich, cat#R4533) and total RNA was extracted using Direct-zol RNA MicroPrep kit (Zymo Research, cat#R2060) according to the manufacturer’s instructions. Total RNA (8 μl) was used in random primed cDNA synthesis with TaqMan Reverse Transcription Reagents (Applied Biosystem, cat#N8080234). Diluted cDNA 1:2 (4 μl) was used in every qPCR reaction using KAPA probe Fast qPCR kit (Kapa Biosystem, cat#KK4706) with primers targeting 45S (Thermofisher Scientific, cat# Mm03985792_s1) and 18S (Thermofisher Scientific, cat# 4333760T) rRNA transcripts. The relative expression levels of rRNA were evaluated using log2 (2ΔCt) calculation.

All statistical analysis was carried out in R (version 4.1.2) and RStudio (version 1.4.1717), using functions from the following packages: tidyverse (version 1.3.1), DESeq2 (version 1.32.0), rstatix (version 0.7.0), fgsea (version 1.18.0), and ggpubr (version 0.4.0). Statistics for bulk RNA-seq was generated using the DESeq2 package without custom settings. Data from non-sequencing methods were imported into R and statistics calculated using Wilcox or T-test, as indicated by the figure legends. Asterisks indicate the P-value, *<0.05, **<0.01, ***<0.001, ****<0.0001.

Using an inducible tet-LIN28B mouse model expressing Flag-tagged human Lin28b cDNA upon doxycycline (Dox) induction (Col1a^tetO-LIN28B Rosa26^rtTA*m2 ) (33), we and others have previously demonstrated that LIN28B expression is sufficient to re-initiate fetal-like hematopoietic output in adult life (14, 19, 36, 37). To better understand the underlying mechanism, we performed immunoprecipitation of its mRNA interactome in developing B cells. To obtain sufficient cell numbers for interactome profiling, adult B cell precursors from uninduced tet-LIN28B and control mice were expanded for 5 days by culturing whole bone marrow cells in the presence of IL-7 and supplemented with Dox ex vivo to induce LIN28B expression. After five days with IL-7, the cultures reached above 80% B cell purity ( Supplemental Figure 1A ). Cells were harvested and subjected to anti-FLAG RNA immunoprecipitation (IP) coupled with poly-A selected RNA-sequencing (RIP-seq) ( Figure 1A ). Immunoprecipitation efficiency of LIN28B was validated by western blot analysis ( Supplemental Figure 1B ). Total RNA sequencing of corresponding input samples was used for calculating RIP-seq enrichment. LIN28B RIP/input ratio of the technical replicates showed a high degree of correlation (R=0.93, p < 2.2e-16) ( Figure 1B ) and were averaged for downstream analyses. 238 transcripts displayed a higher than two-fold RIP/input enrichment in tet-LIN28B samples but showed no trend in the control ( Supplemental Table S2 ; Figure 1C ). Gene ontology (GO) analysis revealed a prominent enrichment for ribosomal protein coding transcripts ( Figures 1D, E ), suggesting that ribosomal function might be a core target pathway of LIN28B during B cell development. Interestingly, similar observations were previously made in other cell types, including 293T cells (22) and BE (2)-C neuroblastoma cells (31). Indeed, cross comparison of our identified targets with these studies yielded cytoplasmatic translation as the predominant overlapping GO term ( Figures 1F, G ). Our analysis also identified other GO terms involved in RNA splicing and oxidative phosphorylation, consistent with previous LIN28A/B interactome studies in various cell types (23, 38). We conclude that the direct LIN28B mRNA regulon in the context of primary bone marrow B cell progenitor cells is dominated by transcripts implicated in universal biosynthetic pathways, which are shared across a wide range of cell types. In particular, our findings implicate the ribosome as a potential target for LIN28B regulation in B cell progenitors.

LIN28B bound mRNA targets have been linked to both translational increase and decrease in a context dependent manner (22–24, 27, 31, 39). In the case of ribosomal protein transcripts, enhanced protein translation was shown by SILAC (stable isotope labeling using amino acid in cell culture) metabolic labeling (22) as well as polysome profiling (31). To assess ribosomal protein levels during B cell development, we performed western blot analysis on sorted B cell precursors from tet-LIN28B and control mice following 10 days of Dox induction in vivo ( Supplemental Figure 2A ). We selected RPL36, RPL23 and RPS9 from the RIP-seq enriched transcript list and found increased protein expression upon LIN28B induction in the Pro-B and large Pre-B subsets ( Figures 2A, B ). Differences were less apparent for the small Pre-B and Imm-B subsets ( Supplemental Figure 2B ), which have lower cellular protein content. Overall, these results are consistent with enhanced translation of LIN28B bound ribosomal protein encoding transcripts.

Given this increase in ribosomal proteins, we expected a corresponding increase in overall ribosomal content upon tet-LIN28B induction. To this end, we measured the 18S rRNAs by qRT-PCR of FACS sorted B cell precursors. To our surprise, Pro and large-Pre-B cells did not exhibit any LIN28B dependent increase in 18S levels. Instead, ribosomal content first began to increase in a LIN28B dependent manner in the quiescent small Pre-B cell stage and reached statistical significance in Imm-B cells ( Figure 2C ). Additional analysis of the 45S precursor rRNA showed the same trend ( Figure 2D ). Thus, an increase in ribosome biogenesis takes place subsequent to the observed increase in ribosomal protein expression during LIN28B induced B cell development. To evaluate whether this observation correlated with changes in global protein synthesis rates, we measured newly synthesized proteins in freshly isolated B cell precursors using metabolic labelling with the methionine analog L-azidohomoalanine (AHA) (40). As expected, FACS readout of AHA uptake from Dox induced control and tet-LIN28B mice showed highly dynamic protein synthesis levels during B cell development that peaks in the large Pre-B cell subset ( Figure 2E ). Pro-B cells exhibited a bimodal profile with respect to protein synthesis that correlated with cell size (AHA^high or AHA^low, Supplemental Figure 2C ). Although there was a tendency towards an increased frequency of AHA^high Pro-B cell in tet-LIN28B induced mice, the changes were not statistically significant ( Figures 2F, G ; Supplemental Figures 2C, D ). tet-LIN28B induction increased global protein synthesis was first observed as cells transitioned from the large to the small Pre-B stage and peaked in the Imm-B cell stage ( Figures 2F, G ), mirroring the observed effects on ribosome biogenesis. This was accompanied by increased cell size as measured by FACS (FSC-A) ( Figure 2H ), and a moderate increase in the fraction of proliferating cells within the Imm-B subset ( Supplemental Figure 2E ). These findings are consistent with elevated protein synthesis boosting biomass accumulation and cell cycle entry following the pre-BCR checkpoint. Interestingly, analysis of mature B cells showed that protein synthesis in peritoneal cavity B-1 and B-2 cells were not affected by ectopic LIN28B ( Figure 2I ). Taken together, we conclude that LIN28B enhances ribosome content and global protein synthesis levels specifically during the later stages of bone marrow B cell development.

We sought to understand the stage dependent effects of LIN28B expression on protein synthesis by considering the signals that control dynamic changes in cellular metabolism and biosynthesis during B cell development. We have previously shown that Lin28B enhances the output of CD5+ Imm-B cells, and that CD5 levels correspond to BCR self-reactivity (19). To assess whether enhanced protein synthesis in Imm-B cells is linked to higher signaling levels through the newly selected BCR we compared CD5 expression with AHA uptake by tet-LIN28B induced Imm-B cells. Indeed, we observed a positive correlation between the two readouts suggesting that enhanced protein synthesis is dictated by BCR signaling strength ( Figures 3A, B ). To strengthen this interpretation, we confirmed that CD5+ Imm-B cells correspond to those with the highest level of Nur77-GFP reporter expression ( Figures 3C, D ; Supplemental Figure 3 ), a well-established measurement of antigen receptor stimulation in mice (41, 42). Interestingly, the percentage of CD5+ Nur77-GFP+ cells was already enhanced in tet-LIN28B induced small Pre-B cells where a ~10% increase in protein synthesis was observed ( Figure 2F ), although the overall MFI was much lower compared to the Imm-B cells. Taken together, our data show that LIN28B induced protein synthesis precedes the positive selection of CD5+ Imm-B cells and is in line with its permissive effect on the developmental progression from the Pre-B cell stage onwards.

The lack of LIN28B induced effects on protein synthesis in Pro-B cells was puzzling given the Lin28b dependent increase in ribosomal protein expression at this stage ( Figure 2A ). Considering that both IL-7 and LIN28B can act through elevating c-MYC expression, and that the c-MYC transcriptional program is a critical driver of cellular protein synthesis (2, 43), we hypothesized that potent IL-7 signaling might mask any LIN28B mediated effects on the c-MYC/protein synthesis axis in Pro-B cells. To test this, we assessed LIN28B induced protein synthesis in Pro-B cells following IL-7 withdrawal ex vivo. B cell precursors were expanded from control and tet-LIN28B adult bone marrow under saturating IL-7 conditions for five days, before culturing under limiting doses of IL-7 ( Figure 3E ) as previously described (3). Two days following IL-7 withdrawal, cells were harvested and FACS analyzed for LIN28B-dependent effects on c-MYC expression and global protein synthesis. Interestingly, although overall c-MYC and protein synthesis levels of Pro-B cells decreased upon IL-7 withdrawal, the decrease was partially alleviated by tet-LIN28B induction ( Figure 3F left). A similar trend was observed for protein synthesis ( Figure 3F right) at limiting IL-7 concentrations. Thus, LIN28B has the capacity to enhance the c-MYC/protein synthesis axis in Pro-B cells, but only under limiting IL-7 availability. Taken together, our results provide a likely explanation to the stage dependent effects of LIN28B expression on protein synthesis during B cell development ( Figure 3G ). In Imm-B cells, enhanced protein synthesis is linked to the increased signaling strength of BCR specificities licensed by LIN28B expression. Prior to the Imm-B cell stage, effects on protein synthesis is dictated by IL-7 responsiveness. During the Pro-B cell stage, potent IL-7 mediated protein synthesis masks the milder effects by LIN28B which become apparent upon entry into the small Pre-B cell stage whereby IL-7Ra expression is fully silenced. This scenario is supported by RNAseq data from sorted B cell precursor subsets from tet-LIN28B and control adult bone marrow, which showed a LIN28B induced increase in c-MYC target genes specifically upon entry into the small Pre-B cell stage ( Figure 3H ). Taken together, our results suggest a critical role for LIN28B in the small Pre-B cell stage as cells loose IL-7 responsiveness and point toward a key role for c-MYC in LIN28B induced protein synthesis.

Endogenous LIN28B expression is developmentally restricted. FACS analysis of LIN28B-eGFP reporter mice (34) captured its window of expression as eGFP gradually diminished to background levels after postnatal day 10 in B cell precursors and postnatal day 19 in hematopoietic stem and progenitor cells (HSPCs) ( Figure 4A ). This time-window correlates with the output of B-1a cells which ceases between day 10 and day 19 during unperturbed hematopoiesis (44). During the same window, protein synthesis in neonatal bone marrow small Pre-B and Imm-B cells was elevated compared to the adult ( Figure 4B ). This was dependent on endogenous Lin28b, as 3-day-old KO neonates had a significantly lower rate of protein synthesis and decreased cell size during the small Pre-B and Imm-B cell stages ( Figures 4C, D ). These defects were primarily qualitative and not accompanied by changes in subset frequency ( Figure 4E ). We conclude that Lin28b dependent increases in protein synthesis during the small Pre-B and Imm-B cell stages represent a hallmark of early life B cell development.

To investigate the importance of elevated protein synthesis for neonatal B cell development, we took advantage of the ribosomal protein hypomorphic Rpl24^WT/Bst mouse model. This model carries a mutation that disrupts Rpl24 mRNA splicing and protein expression (45). The resulting defect in ribosome biogenesis causes a moderate reduction in overall cellular protein synthesis which has no apparent impact on normal adult B cell development (40), but can diminish Eμ-MYC driven B lymphomagenesis known to be ‘addicted’ to heightened ribosomal function (46). Neonatal B cell development has not previously been examined in Rpl24^WT/Bst mice. Given our findings of higher protein synthesis rates, we hypothesized that neonatal B cell development would be more sensitive to constraints in protein synthesis capacity ( Figure 5A ). Indeed, analysis of Rpl24^WT/Bst bone marrow revealed a decrease in the overall CD19+ B cell frequency on postnatal day 3 and day 10 that was resolved by day 19 ( Figure 5B ). This translated into a concomitant decrease in the frequency of B cell precursor subsets in the bone marrow of Rpl24^WT/Bst neonates ( Figure 5C ) and a time-limited decrease in the frequency of mature B cells in both the spleen ( Figure 5D ) and the peritoneal cavity of neonatal mice ( Figure 5E ). Thus, neonatal B cell development is more sensitive to constrained protein synthesis with signs of ribosome ‘addiction’. To verify that these effects are cell intrinsic during B cell development, we performed competitive transplantations of neonatal Rpl24^WT/Bst or littermate bone marrow HSPCs into lethally irradiated congenic recipients ( Figure 5F ). Analysis of donor versus competitor reconstitution of recipient PerC B cells at 14 weeks post-transplantation showed a selective decrease in B-1a cells from Rpl24^WT/Bst donors. In contrast, contribution to the B-2 compartment, which relies on continuous HSPC dependent influx, was comparable between the Rpl24^WT/Bst and wildtype donor HSPCs ( Figure 5G ). Taken together, these findings demonstrate that neonatal B cell output is selectively sensitive to reduced protein synthesis in a cell intrinsic manner.

The B-1a cell subset is enriched for the VH11 and VH12 encoded reactivity against the self-membrane component phosphatidyl choline (PtC) known to originate early in life. To investigate the qualitative impact of constrained protein synthesis on B-1a cell composition, we analyzed the frequency and absolute number of PtC reactive B-1a cells in 19-day-old Rpl24 ^WT/Bst mice using fluorescently labelled PtC liposomes. This timepoint was selected to minimize influence by peripheral homeostatic expansion (47). Indeed, FACS analysis revealed a dramatic decrease in the frequency and absolute number of PtC reactive B cells in the peritoneal cavity of Rpl24^WT/Bst neonates ( Figures 5H–J ). We therefore conclude that elevated protein synthesis is a central distinguishing parameter of neonatal B lymphopoiesis that potentiates the output of CD5+ B-1a cells carrying self-reactive BCRs.