Research ArticleLYMPHOCYTES

Lin28b controls a neonatal to adult switch in B cell positive selection

See allHide authors and affiliations

Science Immunology  27 Sep 2019:
Vol. 4, Issue 39, eaax4453
DOI: 10.1126/sciimmunol.aax4453

Sending B cells back in time

B-1 cells are a subset of self-reactive B cells that arise in early life. Precisely how and why the immune system permits the development of self-reactive B cells in neonates remains a mystery. Here, by studying B cell development in neonatal mice, Vanhee et al. have uncovered the importance of RNA binding protein Lin28b in facilitating positive selection of self-reactive B-1 cells in neonates. Further, they found that ectopic expression of Lin28b was sufficient to promote selection of self-reactive B-1 cells in adult mice. The authors propose that Lin28b functions as a cell-intrinsic switch that jumpstarts the generation of B cells in early life.


The ability of B-1 cells to become positively selected into the mature B cell pool, despite being weakly self-reactive, has puzzled the field since its initial discovery. Here, we explore changes in B cell positive selection as a function of developmental time by exploiting a link between CD5 surface levels and the natural occurrence of self-reactive B cell receptors (BCRs) in BCR wild-type mice. We show that the heterochronic RNA binding protein Lin28b potentiates a neonatal mode of B cell selection characterized by enhanced overall positive selection in general and the developmental progression of CD5+ immature B cells in particular. Lin28b achieves this by amplifying the CD19/PI3K/c-Myc positive feedback loop, and ectopic Lin28b expression restores both positive selection and mature B cell numbers in CD19−/− adult mice. Thus, the temporally restricted expression of Lin28b relaxes the rules for B cell selection during ontogeny by modulating tonic signaling. We propose that this neonatal mode of B cell selection represents a cell-intrinsic cue to accelerate the de novo establishment of the adaptive immune system and incorporate a layer of natural antibody-mediated immunity throughout life.


Whereas T cell development relies on self-peptide–major histocompatibility complex (MHC) ligand-mediated positive selection to effectively contribute to host defense (1), B cell function is not similarly constrained and adult immature B (ImmB) cells are purged for self-antigen reactivity even at low binding affinities (2, 3). Instead, developmental progression of newly formed ImmB cells carrying innocuous B cell receptors (BCRs) relies on ligand-independent tonic signaling mediated by the BCR and CD19 (47). Defying this censorship of self-reactivity are B-1 cells, an innate-like B cell subset harboring an oligoclonal and weakly self-reactive repertoire, which is excluded from germinal center reaction and is responsible for a functionally distinct layer of antibody-mediated immunity through the secretion of protective natural antibodies (8). Their positive selection was elegantly demonstrated using an anti-thymocyte autoantigen-specific BCR transgenic model, in which the presence of the Thy-1 self-antigen was a prerequisite for the establishment of transgene-positive CD5+ B-1 cells of this naturally occurring specificity (9). In the near two decades that followed this discovery, the basis for the ability of B-1 cells to escape central tolerance and undergo positive selection has remained unresolved.

Considering the predominant early-life origin of CD5+ B-1 cells, it is conceivable that B cell positive selection stringency may be temporally controlled during ontogeny. Transgenic expression of BCRs directed against self-antigens such as phosphatidylcholine (PtC) are efficiently incorporated into the neonatal B cell repertoire while being excluded in favor of endogenous BCRs in the adult (912). However, although the use of BCR transgenic models has been instrumental in unveiling unique aspects of B-1 cell selection, conclusions were limited to a few B-1–restricted antigen specificities in a nonphysiological setting. To date, the underlying molecular mechanisms for such a putative ontogenic switch in B cell selection remain unknown.

Lin28b is a mammalian paralog of the heterochronic RNA binding protein Lin28, first described in Caenorhabditis elegans to block the biogenesis of the let-7 family of microRNAs (miRNAs) and thereby control the timing of developmental events (13). We have previously shown that Lin28b exhibits a fetal restricted expression pattern during murine hematopoiesis and that ectopic expression in the adult promotes key aspects of fetal-like lymphopoiesis, including the efficient production of CD5+ B-1 cells (14, 15). Since then, additional evidence from the lymphoid, erythroid, and megakaryocyte lineages has cemented the role of Lin28b as a multilineage molecular switch for fetal hematopoiesis (1620). However, the requirement for endogenous Lin28b during early-life B lymphopoiesis has not been explored.

CD5 is a negative regulator of antigen receptor signaling and was first identified as a surface molecule expressed on human T cells and B chronic lymphoblastic leukemia cells (21). Parallel observations in mice and men established a positive correlation between the frequency of CD5+ B cells and antibody polyreactivity (22). More recently, it was shown in antigen receptor transgenic mice that CD5 levels reflect the degree of self-antigen reactivity in T and B cells (9, 2326). In this study, we use surface CD5 levels to interrogate ontogenic changes in B cell positive selection in BCR wild-type (WT) mice. Our results demonstrate that B cell positive selection is temporally controlled during ontogeny by endogenous Lin28b by amplifying a previously reported positive feedback loop involving CD19 and c-Myc (2729). Thus, Lin28b acts as a cell-intrinsic enhancer of overall ImmB cell positive selection, including the efficient developmental progression of weakly self-reactive B-1 cells early in life.


Establishing a positive correlation between CD5 expression and self-reactivity in BCR WT B-1 cells

The murine adult B-1 cell compartment displays a range of CD5 surface levels, with those originating from fetal liver (FL) hematopoietic stem and progenitor cells (HSPCs) displaying the most surface expression. Although B-1 cells can mature in low numbers from adult HSPCs (3032), these exhibit markedly lower surface CD5 expression (Fig. 1A) (33). Thus, we reasoned that uncovering the significance and developmental control of B cell CD5 levels would be critical in understanding the ontogenic switch in B cell output, leading to reduced B-1 cell generation. Although CD5 expression has been linked to self-reactivity in BCR transgenic models, this link has not been empirically established in a nontransgenic BCR setting. To this end, we took advantage of the high degree of self-antigen–driven clonal dominance in the murine B-1 compartment as a measure of self-reactivity and analyzed the BCR repertoires of sorted CD5hi, CD5int, CD5low, and CD5neg peritoneal cavity B-1 cells from WT adult C57BL/6 mice (Fig. 1B). The successful separation based on CD5 surface levels was confirmed by fluorescence-activated cell sorting (FACS) analysis (Fig. 1C). Subsequent IgHM BCR repertoire sequencing analysis (VDJseq; see data file S2) demonstrated a correlation between increasing CD5 expression and CDR3 clonal dominance (Fig. 1D). This was quantified by the inequality index (Gini coefficient) of clonal representation (Fig. 1E). In addition, we observed a gradual increase in the representation of IGHV-11 and IGHV-12 gene segment usage, known to primarily encode the self-reactive specificity against PtC (Fig. 1F). This observation is consistent with FACS data displaying increasing PtC liposome reactivity with rising CD5 levels (Fig. 1G). Consistent with the superior ability of fetal HSPCs to generate CD5+ B-1 cells (Fig. 1A) and the lack of terminal deoxynucleotidyl transferase (TdT) expression during fetal life (34), the presence of N-nucleotide additions inversely correlated with CD5 levels (Fig. 1H). These data establish that surface CD5 expression correlates with B cell self-reactivity and early ontogeny, providing a practical readout for self-reactivity in a non-BCRtg setting.

Fig. 1 Establishing a positive correlation between CD5 expression and self-reactivity in BCR WT B-1 cells.

(A) Flow cytometric analysis of peritoneal cavity (PerC) B cells 16 weeks after transplantation of E14.5 FL or ABM LinSca1+cKit+ (LSK) HSPCs into lethally irradiated CD45.1+CD45.2+ congenic recipients. Lineage panel for E14.5 FL LSK: Ter119B220Gr1CD3e; ABM LSK: Ter119B220Gr1CD11bCD3e; PerC B cells: Ter119Gr1CD3e. (B) CD5neg, CD5low, CD5int, and CD5hi gating strategy for WT adult PerC B-1 cells (LinCD19+CD43+CD23). (C) Histogram overlay showing post sort analysis of the populations defined in (B). (D) Relative distribution of IgHM CDR3 sequences as determined by high-throughput VDJseq of the indicated populations. (E) Gini index of IgHM CDR3 sequence reads in sorted CD5neg, CD5low, CD5int, and CD5hi populations (left to right). (F) Stacked bar graph indicates the combined frequency of IGHV-11 (white) and IGHV-12 (gray) containing IgHM CDR3 sequence reads. (G) Frequency of PtC liposome–reactive cells in the indicated B-1 populations as assessed by FACS. *P ≤ 0.05, ***P ≤ 0.001, ****P ≤ 0.0001 (n = 9 from three experiments). (H) Frequency of CDR3 containing ≥1 N-nucleotide additions at the N1 and N2 junctions combined. VDJseq data (B to F and H) are representative of two biological and technical replicates.

Lin28b potentiates an elevated mode of positive selection unique to neonatal mice

During adult B cell development in WT mice at steady state, CD5 expression is restricted to self-reactive transitional B cells in the spleen doomed for anergy and exclusion from the long-lived B cell pool (25). This is in contrast to the T cell lineage, where reactivity toward self-peptide–MHC complex is a desirable feature and CD5 is developmentally induced during positive selection in the CD4+CD8+ double-positive thymocyte stage (23, 24). Considering that CD5+ B-1 cells are allowed to mature early in life, we hypothesized that neonatal self-reactive CD5+ ImmB cells would be positively selected into the mature B cell pool, analogous to CD4+CD8+ double-positive thymocytes, rather than being purged from the naïve B cell repertoire. In line with this hypothesis, we found that both the median fluorescence intensity (MFI) and the spread in surface CD5 levels as measured by interquartile range (IQR) were elevated as neonatal bone marrow (NBM) pre-B cells entered the IgM+CD93+ ImmB cell stage (Fig. 2A and fig. S1A). This emergence of CD5hi cells was apparent in the neonate but not adult ImmB cells and coincides with the timing of central tolerance establishment (35). Although only present at low frequencies (0.45 ± 0.25%) among neonatal ImmB cells, PtC-reactive cells reside within this CD5hi fraction (Fig. 2B), suggesting that CD5 levels are induced during B cell maturation on the basis of self-reactivity. The neonatal specific expression of CD5 on ImmB cells was mirrored by the presence of CD5+ transitional T1 cells in the neonatal but not adult spleen (fig. S1B) known to be destined for the B-1 lineage (36). Considering the link between CD5 expression and BCR self-reactivity, these results are consistent with an increase in the tolerated spectrum of self-reactive specificities and an age-restricted licensing of an enhanced mode of positive selection in neonatal ImmB cells.

Fig. 2 Lin28b promotes the positive selection of CD5+ ImmB cells in neonatal mice.

(A) Representative histogram overlay of CD5 surface expression on 2-day-old NBM and 4-month-old ABM small pre-B (LinCD19+CD93+IgMCD43CD24highFSClow) and ImmB cells (ImmB) (LinCD19+CD93+IgM+CD24high). Lower panel shows the CD5 fluorescence intensity IQRs of the same populations. Q1, 25th percentile; Q3, 75th percentile; M, median; FMO, fluorescence minus one staining control. (B) CD5 levels on PtC liposome–reactive and total ImmB cells from 2-day-old WT NBM. (C) Histogram overlay of CD5 levels on 2-day-old NBM and ABM of the indicated genotypes. Lower panel shows the CD5 IQR. (D) Representative FACS plots showing the frequency of PerC CD5+ B-1 cells from 10-day-old mice. (E) Quantification of data shown in (D) (n = 4 to 7 from three experiments). (F) Quantification of relative forward scatter (FSC-A) of ImmB from mice of the indicated genotypes and ages. Data are shown as a value relative to the FSC-A of a representative Lin28b+/+ 2-day-old sample (n = 3 to 7 from three experiments). ns, not significant. **P ≤ 0.01, ***P ≤ 0.001, ****P ≤ 0.0001. (G) Representative histogram overlays of CD5 surface expression on WT BM ImmB of the indicated ages.

We and others have previously established Lin28b as a molecular switch capable of reinitiating fetal-like hematopoiesis including B-1 cell output (14, 15, 20). However, these data were exclusively based on ectopic expression of either Lin28a or Lin28b. To investigate whether endogenous Lin28b is required for the maturation of CD5+ B-1 cells during neonatal life, we analyzed Lin28b−/− neonatal mice (37). We found that Lin28b deficiency led to decreased CD5 MFI and IQR in the ImmB compartment in a dose-dependent manner (Fig. 2C), resulting in the subsequent reduction of CD5+ transitional T1 B cells in the spleen (fig. S1C) and mature CD5+ B-1 cell output in the peritoneal cavity (Fig. 2, D and E). Furthermore, we observed a positive correlation between increased Lin28b dosage and ImmB cell size as measured by forward scatter (Fig. 2F), a hallmark of B cell positive selection (6). These data suggest that permissiveness toward CD5+ ImmB cells and developmental progression of CD5+ B-1 cells rely on endogenous Lin28b in neonatal mice. In line with this notion, we observed a postnatal decline in ImmB cell surface CD5 levels that coincides with the timing of the attenuation in Lin28b mRNA expression in hematopoietic stem cells (15) and becomes indistinguishable from the levels in adult ImmB cells by day 19 of age (Fig. 2G). We conclude that Lin28b expression early in life potentiates an elevated mode of positive selection characterized by the tolerance of CD5+ ImmB cells and the output of CD5+ B-1 cells during the first weeks of life.

Ectopic expression of Lin28b is sufficient to augment overall B cell positive selection in adult mice

To address whether ectopic Lin28b expression during adult B cell maturation is sufficient for potentiating the generation of CD5+ ImmB cells, we put tet-Lin28b mice (38) on a doxycycline (DOX) diet for a minimum of 10 days. Lin28b transgene expression resulted in a potent increase in CD5 MFI and IQR upon pre-B to ImmB cell transition (Fig. 3A). Increased CD5 levels were accompanied by a tet-Lin28b–induced increase in ImmB cell size (Fig. 3B) and the emergence of CD5+ transitional T1 cells in the spleen (fig. S1D), consistent with the neonatal mode of enhanced positive selection (6, 36). Despite the observed increase in cell size, we did not observe apparent changes in the percentage of replicating cells among tet-Lin28b pre-B and ImmB cells (fig. S2, A and D), consistent with a modest increase in cellular anabolism. To track the maintenance of CD5 expression in the periphery, we adoptively transferred the 20% highest and lowest CD5-expressing ImmB cells from DOX-treated tet-Lin28b mice into individual non–DOX-treated Rag1KO recipients. Three weeks after adoptive transfer, CD5 levels were largely maintained in most donor-derived mature B cells, demonstrating that CD5+ ImmB cells are predisposed to give rise to CD5+ B-1 cells in a cell-intrinsic manner (Fig. 3, C to E). Furthermore, PtC-reactive specificities exclusively derived from the CD5+ donor population (Fig. 3F), consistent with CD5 marking self-reactive ImmB cells.

Fig. 3 Ectopic expression of Lin28b is sufficient to support enhanced B cell positive selection in the adult.

(A) Representative histogram overlay of CD5 surface expression on ABM small pre-B cells and ImmB cells from tet-Lin28b and WT ABM mice fed a DOX diet for at least 10 days. Lower panel shows the CD5 IQR of the populations in the upper panel. (B) Quantification of relative FSC-A of pro-B, pre-B, and ImmB cells from tet-Lin28b ABM. The ratio of tet-Lin28b/WT cells from each experiment was calculated and plotted (n = 5 from three experiments). Statistics of the ratios against the value 1 were performed. ***P ≤ 0.001, ****P ≤ 0.0001. (C) Sort strategy for isolation of the 20% highest and lowest CD5-expressing ABM ImmB cells from DOX-fed tet-Lin28b mice. Sorted cells were transplanted intraperitoneally (I.P.) into untreated RAG1KO recipient mice on a normal diet. (D) Frequency of PerC CD5+ B-1 cells 3 weeks after transplantation (n = 3 from two experiments). (E) Representative CD5 surface expression on PerC B-1 cells 3 weeks after transplantation. (F) Representative PtC liposome reactivity of PerC B-1 cells.

To establish at which B cell developmental stage Lin28b is required to mediate efficient CD5+ B-1 cell output, we adoptively transferred pro-B, pre-B, or ImmB cells from untreated tet-Lin28b donor mice into DOX-fed Rag1KO recipients (fig. S3A). CD5+ B-1 cell output was most efficient when transgene expression is initiated at the pro-B cell stage and declined when initiated later at the pre-B and ImmB stages (fig. S3B). We subsequently performed the reverse experiment in which tet-Lin28b transgene expression in DOX-treated donor mice was turned off at different stages of B cell differentiation upon adoptive transfer into non–DOX-treated Rag1KO recipients. Our results demonstrate that cessation of Lin28b transgene expression before the ImmB stage is detrimental to developmental progression of progenitors destined for the CD5+ B-1 fate (fig. S3B). Although there is a delay in transgene expression/cessation after transfer, these data firmly establish a need for Lin28b expression during the central tolerance checkpoint of B cell maturation (fig. S3C).

To quantitatively test whether the efficiency of B cell positive selection is enhanced by Lin28b, we measured mature B cell output by single-cell lineage tracing of preselection pro-B cells using lentiviral cellular barcoding (15, 39). Adult bone marrow (ABM) pro-B cells were FACS-sorted from uninduced tet-Lin28b mice and transduced with a lentiviral library (Barcode-GFP-LV) encoding high-complexity DNA barcodes (Fig. 4A) (40). After transduction, the cells were divided into two equal halves that were adoptively transferred into Rag1KO recipients fed either DOX or normal diet. Green fluorescent protein–positive (GFP+) mature B cells were isolated from the recipient spleen 2 weeks after transfer and analyzed for unique barcode content as a measure of relative selection efficiency (fig. S4, A to E). The number of unique progenitors that contributed to mature B cells was significantly increased upon ectopic Lin28b expression (Fig. 4, B and C). Barcode read frequency analysis did not reveal any unbalanced clonal expansion over the course of the experiment (Fig. 4B). Because barcode labeling efficiency was directly comparable, our finding is consistent with an increase in progeny:precursor ratio and thereby selection efficiency. In addition, both splenic mature B-1 and follicular B-2 cells from DOX-treated recipients displayed increased CD5 levels (Fig. 4, D and E). These data are reminiscent of previous observations that the majority of B cells in the human fetal spleen and cord blood are CD5+ (4143) and implicate Lin28b in the mechanism underlying this developmentally restricted expression pattern. We conclude that Lin28b augments overall B cell positive selection, including selection into the B-1 lineage.

Fig. 4 Lin28b alters the efficiency of overall B cell selection.

(A) Schematic showing the cellular barcoding setup of pro-B cells. Tet-Lin28b pro-B cells (1 × 105) (LinCD19+IgMCD93+cKit+CD43+CD24loCD25) were sorted from untreated mice and transduced with Barcode-GFP-LV. Transduced cells were divided into two equal parts that were individually adoptively transferred into DOX-fed and untreated Rag1KO recipients, respectively. Barcode representation of mature splenic B cell progeny (CD19+CD93GFP+) was assessed 2 weeks after adoptive transfer. I.V., intravenously. (B) Stacked bars show representative read frequencies of individual barcodes retrieved. Number on top of bars indicates the number of unique barcodes detected in both PCR technical replicates (for details, see fig. S3). (C) Enumeration of barcodes retrieved (n = 5 from two experiments). Significance was tested using a paired t test. (D) Representative dot plots showing splenic mature B cells 2 weeks after transfer. (E) Histogram overlay shows representative CD5 expression of splenic mature B-1 (GFP+CD19+CD93CD43+CD23) and follicular B-2 subsets (GFP+CD19+CD93CD43CD23+). (F) Kinetics of EdU label progression during B cell development after administration of a single pulse of EdU into tet-Lin28b and WT mice fed a DOX diet for at least 10 days before labeling. Labeling of the indicated populations at the indicated time points was assessed by FACS (n = 3 to 5 per data point from two to three experiments). Error bars show the SD of the mean. *P ≤ 0.05.

To track the kinetics of developmental progression through the central tolerance checkpoint in the presence or absence of tet-Lin28b transgene expression in vivo, we pulsed DOX-fed tet-Lin28b and WT adult mice with a single dose of 5-ethynyl-2′-deoxyuridine (EdU) and assessed labeled B cells after 2, 24, 48, 72, and 96 hours of chase (Fig. 4F). Whereas initial labeling efficiency of IgM B cell progenitors was comparable, labeled tet-Lin28b cells displayed accelerated progression through the B cell developmental stages and emergence into the splenic transitional T1 stage. Together, our data demonstrate that Lin28b promotes overall positive selection efficiency, leading to accelerated developmental progression of bone marrow egress.

Lin28b promotes B cell positive selection by amplifying the CD19/PI3K/c-Myc feedback loop

To explore the molecular pathway by which Lin28b regulates B cell positive selection, we performed RNA sequencing (RNA-seq) analysis of DOX-fed tet-Lin28b and WT ABM ImmB cells (see data file S3). Differential gene expression analysis identified an increase in the expression of several previously known Lin28b/let-7 target genes (e.g., Igf2bp3, Myc, and Lin28b itself) (Fig. 5A) as well as increased Cd5 and Nr4a1 (Nur77) transcript levels (Fig. 5B), consistent with enhanced self-reactivity (Fig. 3) (44). Unsupervised gene set enrichment analysis (GSEA) of the hallmark gene set collection from the Molecular Signature Database (45) identified c-Myc–induced targets as the top up-regulated molecular signature upon tet-Lin28b transgene expression (Fig. 5, C and D, and fig. S5). let-7 and interferon-γ response target genes served as positive and negative controls of our GSEAs, respectively (Fig. 5D). In line with this, we observed a consistent increase in Myc mRNA and protein levels upon ectopic tet-Lin28b expression in ABM ImmB cells (Fig. 5, E and F). This finding is consistent with previous reports establishing c-Myc as a let-7 target that is derepressed upon Lin28b expression (4648).

Fig. 5 Lin28b promotes B cell positive selection by activating the c-Myc transcriptional program.

(A) RNA-seq of sorted ABM ImmB cells from WT and tet-Lin28b mice after 2 weeks of DOX diet (n = 3). Volcano plot shows differential gene expression analysis (DESeq). (B) Normalized DESeq counts for the indicated genes. (C) Unsupervised GSEA for significantly enriched gene sets among the hallmark gene set collection from the Molecular Signature Database (45) (left) and top enriched gene sets (right). (D) Leading edge plots show enrichment analysis of the indicated gene sets. Top 100 predicted let-7 targets are extracted from TargetScan7.2 (83). (E) Normalized DESeq counts and qPCR analysis of Myc transcript levels in ABM ImmB cells from WT or tet-Lin28b mice (n = 3). (F) Western blot analysis of c-Myc and Lin28b transgene protein levels in sorted ABM ImmB cells. (G) Histogram showing c-Myc protein expression as measured by intracellular FACS in 3-day-old NBM and ABM ImmB cells from the indicated genotypes. (H) Western blot analysis showing c-Myc protein expression in 3-day-old NBM and spleen cells from the indicated genotypes. (I) RNA-seq of WT (n = 3) and Lin28b−/− (n = 4) 3-day-old NBM. Left: Top enriched hallmark gene sets. Right: Leading edge plot showing a depletion of c-Myc targets in Lin28b-deficient ImmB cells based on DESeq analysis.

Next, to assess whether c-Myc protein levels are controlled by endogenous Lin28b during neonatal B cell maturation, we performed intracellular FACS staining (Fig. 5G). We observed an elevated level of c-Myc protein in WT neonatal ImmB cells compared with their adult counterparts. This elevated c-Myc expression is decreased in neonatal Lin28b−/+ and Lin28b−/− ImmB cells in a dose-dependent fashion. Immunoblotting of total neonatal splenic lysates also confirmed reduced c-Myc expression in Lin28b−/− mice (Fig. 5H). In line with this finding, RNA-seq analyses of sorted neonatal ImmB cells revealed a depletion of c-Myc target genes among Lin28b−/− compared with WT mice (Fig. 5I) (see data file S4). Thus, Lin28b critically maintains an elevated level of c-Myc protein levels and function during neonatal B cell maturation.

To assess whether transgenic c-Myc overexpression mediates increased positive selection, we analyzed ImmB cells of precancerous Eu-Myc transgenic mice. Our findings demonstrate decreased λ light chain usage (fig. S6, A and B), increased cell size (fig. S6C), and a trend, although not significant, toward increased surface CD5 expression (fig. S6, D and E), consistent with increased positive selection and a partial recapitulation of tet-Lin28b–induced B cell maturation. We also observed a marked increase in CD19 surface expression on Eu-Myc ImmB cells, consistent with the previously reported CD19/c-Myc positive feedback loop (fig. S6F) (28). We did not, however, detect any reactivation of endogenous Lin28b expression as has been reported in several human and murine tumor models (fig. S6, G and H) (49, 50). Together, our evidence suggests that c-Myc is a key mediator of Lin28b-mediated positive selection. It is important to note that c-Myc is not the only downstream mediator of Lin28b. We observed several other significantly enriched molecular signatures in our GSEA (Fig. 5C and fig. S5)—including E2F, oxidative phosphorylation, and mammalian target of rapamycin complex 1 (MTORC1) signaling—consistent with previous reports of Lin28b action (38, 47, 51).

Previous studies have implicated c-Myc as a key mediator of the CD19/PI3K (phosphatidylinositol 3-kinase) signaling pathway that critically controls ImmB cell positive selection and mature B cell survival (2729, 5254). CD19−/− mice display reduced c-Myc protein levels (28), extensive receptor editing and developmental arrest of ImmB cells (57), as well as decreased peripheral B cell numbers and B-1 cell representation (5557). We crossed the tet-Lin28b transgene onto a CD19−/− background (56) to test whether Lin28b-induced Myc expression can genetically rescue the observed defects in CD19−/− mice. After a 3-week DOX treatment, we observed a near-complete rescue of splenic B cell numbers in tet-Lin28b transgenic CD19−/− mice (Fig. 6, A and B) as well as a partial rescue of B-1 cell representation (figs. S6, I to K). FACS analysis of CD19−/− ImmB cells revealed the expected characteristics of impaired positive selection—including decreased CD5 levels, enhanced λ:κ light chain usage ratio (Fig. 6, C to E), and reduced ImmB cell representation (fig. S6L). These defects were also restored to normal adult levels by tet-Lin28b expression. These results demonstrate a potent ability of Lin28b to functionally compensate for CD19 during B cell selection and maintenance.

Fig. 6 Ectopic Lin28b functionally replaces CD19 in B cell development and maintenance.

(A) Representative FACS plots showing frequency of splenic B cell adult mice of the indicated genotypes fed a DOX diet for 3 weeks (n = 5 to 7 from three experiments). (B) Quantification (left) and absolute B cell numbers (right) in spleen of mice from (A). (C) CD5 levels on ABM ImmB cells from (A). (D) Representative plots of κ and λ light chain usage in ABM ImmB cells. (E) Quantification of (D) (n = 4 to 9 from three experiments). (F) Representative Western blot of pPDK1, pGSK3b, c-Myc, human Lin28b (tg-Lin28b), pS6, and b-actin levels in magnetically lineage-depleted (CD3Gr1CD11bTer119) splenic B cells (n ≥ 3). (G) Representative histogram of pS6 levels in ABM ImmB cells from WT and tet-Lin28b mice (n ≥ 3). (H) Proposed CD19/c-Myc/Lin28b positive feedback loop in neonatal and adult mice.

To assess whether tet-Lin28b expression up-regulates c-Myc expression in CD19−/− mice, we performed Western blot analyses. c-Myc protein in total splenic B cells increased upon tet-Lin28b transgene expression. In addition, Lin28b also augmented the PI3K signaling pathway in CD19+/+ and CD19−/− mice as measured by phospho-PDK1, phospho-GSK3b, and phospho-S6 (Fig. 6, F and G), consistent with a previously reported role for Lin28b in the positive regulation of the PI3K/MTORC1 signaling pathway (38), as well as our own RNA-seq analyses (Fig. 5C). We conclude that Lin28b promotes B cell positive selection at least in part through the amplification of a previously reported c-Myc/PI3K positive feedback loop (2729).

Lin28b-induced positive selection in the adult produces fully functional CD5+ B-1 cells

To assess the functionality of CD5+ B-1 cells induced by tet-Lin28b–mediated positive selection in the adult, we compared their in vivo and in vitro effector functions with natural CD5+ B-1 cells found in unperturbed adult WT mice (nB-1). To this end, we first established a transplantation-based model system that circumvents a critical pitfall of previous approaches (14, 20), in which constitutive overexpression of Lin28a/b in mature B-1 cells failed to resolve developmental effects on positive selection from those on mature B-1 cell maintenance and function. Donor tet-Lin28b ABM HSPCs were DOX-treated for 4 weeks to allow for a transient wave of fetal-like CD5+ B-1 cell (L28B-1) output upon transfer into preconditioned Rag1KO recipient mice (Fig. 7A). WT FL and ABM LinSca1+cKit+ (LSK) HSPCs were similarly transplanted, yielding FLB-1 and ABMB-1 control populations, respectively. Donor-derived B-1 cells were analyzed after a 12-week chase period during which DOX treatment was removed. We confirmed the cessation of ectopic Lin28b protein expression and the subsequent decline of ongoing B-1 maturation (fig. S7, A and B), which mimics the natural down-regulation of endogenous Lin28b during postnatal life (14, 17). Whereas WT ABM HSPCs produced a low frequency of CD5+ B-1 cells, transient Lin28b expression resulted in the long-term reconstitution of L28B-1 cells with a similar CD5 expression profile as nB-1 cells (Fig. 7, B and C, and fig. S7, C to E). To more rigorously assess the long-term survival of L28B-1 cells in the absence of ABM influx, a hallmark of B-1 cell biology (58), we performed competitive transfer of FACS-sorted test B-1 cell subsets at a 1:2 ratio with nB-1 congenic competitor cells into Rag1KO mice. At 12 weeks after transfer, the L28B-1 cells displayed long-term fitness comparable with FLB-1 and nB-1 cells (Fig. 7D). In contrast, ABMB-1 cells were markedly outnumbered by competitor cells. Thus, continuous Lin28b expression is not required for the bone marrow–independent long-term maintenance of L28B-1 cells in the periphery. Instead, our data suggest that the initial Lin28b-dependent mode of B-1 cell selection is the main predictive parameter of longevity. The stability of surface CD5 levels closely mirrored our observations of peripheral maintenance. While L28B-1 cells maintained their CD5 levels long term, the remaining ABMB-1 cells, initially sorted for CD5 positivity, displayed a lower CD5 expression profile at the time of analysis (Fig. 7E). Corroborating these data, CD5 levels on ABMB-1 cells were destabilized upon in vitro lipopolysaccharide (LPS) stimulation but remained constant on L28B-1 and nB-1 cells (fig. S7F). We conclude that the long-term fitness and stability in CD5 expression critically depend on the Lin28b-mediated mode of positive selection.

Fig. 7 Tet-Lin28b–induced positive selection during adult B lymphopoiesis produces fully functional CD5+ B-1 cells.

(A) Schematic showing the DOX-STOP experimental setup. A total of 10,000 LSK HSPCs were sorted from the indicated donors and transplanted into sublethally irradiated Rag1KO recipients. Recipients were kept on a DOX diet for 4 weeks, allowing for transient B-1 cell output by tet-Lin28b HSPCs, followed by a normal diet for 12 weeks (fig. S6). (B) Gating strategy for FACS sorting of PerC CD5+ B-1 cells isolated from untransplanted adult (16-week-old) WT mice (nB-1) or the indicated reconstituted recipients 16 weeks after transplantation. (C) Frequency of PerC CD5+ B-1 cells 16 weeks after transplantation (n = 7 to 14 from three experiments). (D) A total of 2 × 104 sorted CD5+ B-1 cells from the indicated mice were competitively transferred by intraperitoneal injection along with congenic nB-1 competitors at a 1:2 ratio. Bar graph shows donor:competitor recovery ratio 12 weeks after transfer (n = 2 to 6 from two experiments). (E) Histogram showing CD5 expression levels of transferred nB-1, FLB-1, ABMB-1, and L28B-1 cells and their congenic competitors. (F) Frequency of CDR3 containing one or more N-nucleotide additions at the N1and N2 VDJ junctions combined. n = 3 to 7 biological replicates from three experiments. *P ≤ 0.05, **P ≤ 0.01, ****P ≤ 0.0001. Error bars show the SD of the mean. (G) Graphical summary. Our data put forward a unified model for the augmented B-1 cell output early in life. Lin28b-dependent ImmB cell positive selection early in life endows B-1 cells with their hallmark characteristics, including a stable CD5 expression profile and long-term fitness. Ectopic Lin28b in adult life can efficiently induce B-1 cell positive selection on the backdrop of a highly diverse adult type BCR junctional diversity. Adult-derived B-1 cells that develop independently of Lin28b are functionally comparable in terms of IL-10 and spontaneous IgM secretion but display reduced CD5 expression and impaired long-term fitness. We propose that Lin28b-mediated enhancement of B cell tonic signaling licenses self-reactive B-1 cell output during the neonatal period when the risk for harmful autoreactivity is constrained by limited junctional diversity, thereby contributing to an important layer of natural antibody-mediated immunity.

One hallmark of B-1 cells is their semi-invariant repertoire stemming from the lack of TdT expression early in life (34). To address the diversity of the L28B-1 BCR repertoire, we performed VDJseq of the immunoglobulin heavy chain repertoire. Our results demonstrate that L28B-1 cells, like ABMB-1 cells, exhibit high junctional diversity (Fig. 7F). These results are in line with the observation that Lin28b does not abrogate TdT expression in ABM pro-B cells (fig. S7G) and is consistent with a previous report based on constitutive Lin28b overexpression (20). Thus, we conclude that Lin28b allows for the positive selection of a highly diverse repertoire of CD5+ ImmB cells and that germline-encoded specificities prevalently used early in life are not a prerequisite for neonatal-like B cell positive selection or long-lived B-1 cell fate.

Last, we assessed IgM and interleukin-10 (IL-10) secretion as two important B-1 effector functions. Our results show that both L28B-1 and ABMB-1 isolated from the spleen cells are capable of spontaneous IgM secretion by enzyme-linked immunospot (ELISPOT) assay. Furthermore, IL-10 production upon in vitro LPS stimulation of both L28B-1 and ABMB-1 was comparable with nB-1 and FLB-1 levels (fig. S7, H and I). We conclude that Lin28b-dependent positive selection produces functionally competent B-1 cells and is a prerequisite for the B-1 signature characteristics of long-term survival and stable CD5 expression. However, the Lin28b-dependent mode of positive selection is not required for other critical innate-like functional properties such as IL-10 production and spontaneous IgM secretion (Fig. 7G) (8).


In this study, we demonstrate that the early-life restricted expression pattern of Lin28b potentiates an enhanced mode of B cell positive selection during a limited window of time. Numerous studies have previously cemented a role for Lin28b as a multilineage master regulator of fetal-like hematopoiesis (14, 1620). Our finding that Lin28b promotes positive selection by amplifying a positive feedback loop involving CD19 and c-Myc represents a highly B lineage–specific mode of Lin28b action. CD19 overexpression has previously been demonstrated to enhance positive selection, B-1 cell numbers, and autoantibody production by shifting the selection criteria for newly formed B cells (55, 59, 60). Thus, the ability of Lin28b to replace CD19-mediated tonic signaling during B cell maturation represents a cell-intrinsic mechanism of altering the threshold for B cell selection early in life and positions the Lin28b/let-7 axis among a growing list of miRNA-dependent modes to fine-tune the PI3K pathway and B cell selection (29, 61, 62).

Our results suggest that the proto-oncogene c-Myc, which promotes B lymphomagenesis through the CD19/PI3K pathway (28, 63), is one key mediator of Lin28b action in B cells. In addition to being a let-7 target that is derepressed upon Lin28 expression (4649), the Myc transcript was recently found to be bound by Lin28b, which may result in stabilization of the transcript (64). These findings are consistent with a conserved functional synergy between Lin28b, itself a c-Myc transcriptional target, and c-Myc during normal development, induced pluripotent stem cell reprogramming, as well as oncogenic transformation. Previously, another let-7 target, Arid3a, was shown to be ontogenically controlled by Lin28b (20) and to promote B-1 cell output, in part, through up-regulation of Myc expression (65). These findings implicate Arid3a as an additional player in the molecular network controlling the neonatal mode of positive selection. Many questions remain as to how Lin28b achieves enhanced B cell positive selection including to what extent death by neglect and negative selection is suppressed and how Lin28b deploys let-7–dependent versus let-7–independent mechanisms. Lin28b-positive (66) and CD5-positive (67) subtypes of pediatric acute lymphoblastic leukemias have been independently described and highlight the need to dissect whether or not endogenous Lin28b is complicit with c-Myc in the oncogenesis of B cell cancers.

Previously, the topic of B-1 cell selection has mainly been studied through the narrow lens of a small number of germline-encoded BCR specificities using BCR transgenic mice. Here, we used several innovative approaches to tackle this question in BCR WT mice. First, we establish CD5 FACS analysis as a useful tool for studying ImmB cell positive selection in general and neonatal mice in particular. Second, our use of cellular barcoding represents a suitable method for quantifying the relative precursor-progeny frequency of B cell selection efficiency. Third, our use of transient tet-Lin28b induction in a transplantation setting allowed us to tease apart transgene effects on B cell positive selection from those on mature B cell function and maintenance. Together, these approaches have allowed us to demonstrate that Lin28b expression during the central tolerance checkpoint is necessary and sufficient for the generation of fully functional long-lived B-1 cells on the backdrop of a highly diverse adult type BCR repertoire. Although the degree and specificity of self-antigen reactivity of tet-Lin28b–induced CD5+ B-1 cells that are not PtC reactive remain to be empirically determined, our findings indicate that the Lin28b-dependent mode of B-1 cell selection into the mature B cell pool is a predictive feature of bone marrow–independent self-renewal.

Despite the many remaining questions, our study establishes an undeniable link between developmental timing and B cell signaling strength. This is a substantial advancement in our understanding of B-1 cell fate determination that reconciles two compelling yet previously disjointed bodies of work emphasizing the importance of early progenitor ontogeny on the one hand (6871) and BCR-driven events on the other (21, 7275). Most recently, elegant work from Graf et al. (75) demonstrating that the Cre recombinase–mediated exchange of BCR transgene identity in mature naïve follicular B-2 cells confers B-1 functionality cemented the instructive power of BCR signaling. However, CD5 expression was not fully restored in these BCR-switched cells, leaving open the possibility that additional developmental determinants are required for complete reprogramming to natural B-1 fate (75). To this end, our findings that CD5 levels are set in the ImmB cell stage and that transient Lin28b expression during, but not beyond, bone marrow B cell selection is a requirement for stably maintained CD5 expression provide missing links. Our findings represent a mechanistic explanation to the early-life bias of CD5+ B-1 cell output, providing a unified model for B-1 lineage choice in which Lin28b acts as a developmentally restricted regulator of B cell signaling, which, in turn, instructs B-1 cell fate during positive selection.

The risks for incorporating harmful autoreactivity raise the stakes for Lin28b-dependent enhancement of positive selection early in life. Still, when restricted to a narrow developmental window during ontogeny, this seemingly reckless behavior may confer significant evolutionary gains for the host. First, this may represent a critical mechanism to accelerate the expansion phase of the adaptive immune system during the sensitive neonatal period to mitigate the potential risks of lymphopenia. Second, it represents an opportunity for the incorporation of useful polyreactive, and by definition somewhat self-reactive, specificities into the mature B cell pool that could contribute to tissue homeostasis as well as close any gaps in our immune system exploitable by pathogens. Moreover, continuous encounter of self-antigen primes innate-like B cells for rapid and intense responses before the delayed response of conventional follicular B-2 cells. Given that the lack of TdT expression severely restricts CDR3 diversity early in life, the benefits of Lin28b-dependent positive selection may outweigh the limited risks for generating harmful autoimmunity. The unique role of polyreactive antibodies in humoral immunity is widely recognized in both the broad antimicrobial function of natural antibodies and the highly specific protection responses to thymus-dependent antigens (76, 77). Together, our work provides key insights into the phenomenon of B cell positive selection from a developmental biology perspective and highlights the contribution of the ontogenically restricted expression pattern of Lin28b to the establishment of peripheral B cell heterogeneity and lifelong antibody diversity.



Col1a1tm2tetO-LIN28B mice crossed to R26m2rtTA were obtained from the laboratory of G. Daley (Harvard Medical School) (38). Transheterozygous mice are herein referred to as tet-Lin28b. In all experiments where tet-Lin28b mice were used, control mice harboring the R26m2rtTA allele were used and are designated as “WT.” The following mouse strains were obtained from the Jackson Laboratory: Lin28b−/− (B6.Cg-Lin28btm1.1Gqda/J strain number 023917) (37), Rag1KO (B6.129S7-Rag1tm1Mom/J strain number 002216), CD45.1 (B6.SJL-Ptprca Pepcb/BoyJ strain number 002014), CD19−/− [B6.129P2(C)-Cd19tm1(cre)Cgn/J strain number 006785], and EuMYC [B6.Cg-Tg(IghMyc)22Bri/J strain number 002728]. CD45.2 WT mice were from Taconic (B6NTac, Taconic). Adult mice were 10 to 16 weeks of age unless specified differently. DOX diet (200+ mg/kg) was obtained from ssniff GMBH. All animal procedures were performed in accordance with ethical permits approved by the Swedish Board of Agriculture.

HSPC transplantations

For HSPC transplantations, 10,000 LSK HSPCs from CD45.2 E14.5 FL or ABM were FACS-sorted and transplanted into either Rag1KO or congenic CD45.2+CD45.1+ recipient mice together with 200,000 CD45.1 total BM support cells. Recipient mice were irradiated (1 × 450 cGy for Rag1KO and 2 × 450 cGy for CD45.1) and transplanted by tail vein injection. All mice were given ciproxin (125 mg/liter) containing water for 2 weeks after irradiation. Endpoint analysis was performed 16 to 20 weeks after transplantation as indicated for each experiment. For the DOX-STOP regimen in Fig. 7, recipient mice were fed a DOX diet for 7 days before transplantation with and 4 weeks after transplantation. Mice were analyzed after a DOX-free chase period of 12 weeks.

Lentiviral cellular barcoding and transplantation of pro-B cells

A total of 100,000 pro-B cells were sorted and transduced with Barcode-GFP-LV (40), as previously described (15). The equivalence of 50,000 sorted and transduced pro-B cells was transplanted per Rag1KO recipient 12 hours after Barcode-GFP-LV transduction. Recipients for cells in the +DOX condition were maintained on a DOX diet starting 7 days before transplantation and throughout the course of the experiment. Splenic mature B cells (CD19+CD93GFP+) were FACS-sorted 2 weeks after transplantation and analyzed for barcode content, as previously described (15). Samples were indexed using the Nextera XT Indexing Kit (Illumina) and sequenced using the Illumina NextSeq platform (Illumina). Two technical polymerase chain reaction (PCR) replicates were performed from the genomic DNA of each population, and only barcodes found in both technical replicates were considered for further analysis, as shown in fig. S4.

Flow cytometry

Antibodies are detailed in table S2. BM cells were extracted by crushing the bones from hindlimbs, hips, and spine with mortar and pestle (78). Adult, but not neonatal, BM and spleens were subjected to red blood cell lysis. Peritoneal lavage was performed using 10 ml (adult) or 2 ml (10-day neonate) of FACS buffer [Hanks’ balanced salt solution supplemented with 0.5% bovine serum albumin (BSA) and 2 mM EDTA]. Antibody staining was performed in FACS buffer at a density of maximum 1 × 107 cells/100-μl volume for 30 min on ice. For intracellular staining, a maximum of 10 × 106 cells were fixed and permeabilized either in 100 μl of 4% formaldehyde at room temperature (RT) for 15 min and in 0.5% Triton-X in phosphate-buffered saline (PBS) at RT for 15 min or using the Transcription Factor Buffer Set (BD Pharmingen) according to the manufacturer’s instructions.

All FACS experiments were performed at the Lund Stem Cell Center FACS Core Facility, Lund University, on FACSAria III, FACSAria IIu, Fortessa, and LSR II instruments (Becton Dickinson). Bulk populations were sorted using a 70- or 85-μm nozzle, 0.32.0 precision mask, and a flow rate of maximally 6000 events/s. For pro-B cell or ImmB cell isolation, total BM cells were pre-enriched by MACS (Miltenyi Biotec) depletion of Lineage+ cells (Ter119+Gr1+CD11b+CD3e+) according to the manufacturer’s instructions before FACS sorting.

All analyses were performed using FlowJo 9 or 10. For IQR plots, the 25th percentile, median, and 75th percentile of expression were determined using FlowJo. Values were then plotted using R and the ggplot2 package.


VDJseq on isolated complementary DNA (cDNA) was performed as previously described (79). All analyses were performed using R version 3.2.1 (2015) and R version 3.5.2 (2018) (R Core Team; Gini indexes were calculated using the R Reldist package 1.6-6. R scripts are available upon request.

EdU pulse label

Mice were injected intraperitoneally with a single dose of 5 mg of EdU (Abcam). At the indicated time points, 10 × 106 ABM or spleen cells were prepared and stained for surface markers as described above. Next, cells were fixed in 4% paraformaldehyde for 15 min at RT and permeabilized using a solution of 0.5% Triton-X in PBS for 15 min at RT. Subsequently, the Click-iT reaction between EdU and azide–Alexa Fluor 555 was performed according to the manufacturer’s protocol (Thermo Fisher Scientific).

Enzyme-linked immunosorbent assay

Enzyme-linked immunosorbent assay (ELISA) plates were coated with an IL-10 antibody in carbonate buffer [0.1 M Na2CO3:0.1 M NaHCO3 (pH 9.6) at a 1:3 ratio] overnight at 4°C. Plates were washed with PBS-T (PBS with 0.1% Tween 20) and blocked with 1% BSA in PBS and incubated at RT for 2 hours. After washing with PBS-T, plates were washed before adding appropriate sample dilutions. After washing with PBS-T, anti–IgM–horseradish peroxidase antibody was added for 2 hours at RT. After washing of plates with PBS-T, TMB substrate (Pierce) was added. The reaction was stopped by an equal amount of STOP solution (2 M H2SO4).

RNA sequencing

RNA-seq was performed on sorted ABM or 3-day-old NBM ImmB cells (CD19+B220loCD93+IgM+CD24hi). Libraries were generated using the SMART-Seq v4 Ultra Low Input RNA Library Prep Kit (Clontech) and the Nextera XT DNA Indexing Kit (Illumina) according to the manufacturer’s instructions. Libraries were indexed using the Nextera XT v2 Indexing Kit (Illumina) according to the manufacturer’s instructions. Libraries were sequenced on an Illumina NextSeq system using a NextSeq V500/550 150-cycle high-output kit. Reads were aligned using STAR aligner v2.7.1 (80) and counted using RSEM v.1.28. Differential gene expression analysis was performed using the DESeq2 R package v1.22.2 (81) in R 3.5.2. GSEA was performed using the fgsea R package v1.8.8 in R 3.5.2. (82) was run on DESeq2 normalized counts using standard parameters with 1000 permutations and the log2 fold change as ranking metric. H collection MSigDB hallmark gene sets were used for analysis (45). Data are made available in the Gene Expression Omnibus repository (GSE135603).

Quantitative PCR

Total RNA was isolated using RNAzol (Sigma-Aldrich). cDNA synthesis was performed using the TaqMan Reverse Transcription Reagent Kit (Life Technologies) according to the manufacturer’s instructions. Quantitative PCR (qPCR) analysis was done using probes from Integrated DNA Technologies (IDT) (Myc Mm.PT.39a.22214843.g and Actb Mm.PT.58.28494642, Lin28b Mm.PT.58.8558661) and master mix from Kapa Biosystems.

Western blot

Cells were lysed in 50 mM tris-HCl (pH 7.8), 150 mM NaCl, 1 mM EDTA, and 1% NP-40. Loading samples were prepared with Laemmli buffer (Bio-Rad) with β-mercaptoethanol (Scharlau Chemicals) and protease inhibitor (Roche). Proteins were separated by SDS–polyacrylamide gel electrophoresis (12% Mini-Protean TGX gel, Bio-Rad) and transferred to a nitrocellulose membrane (Trans-Blot Turbo Mini Nitrocellulose Transfer Packs, Bio-Rad). Membranes were blocked for 1 hour at RT in 5% BSA, 5% skimmed milk, or 5% enhanced chemiluminescence (ECL) blocking solution in tris-buffered saline (TBS)–0.05% Tween 20 depending on the manufacturer’s instructions. Antibodies are listed in table S2. The ECL Prime kit (GE Healthcare) was used to detect protein bands with a ChemiDoc XRS+ system (Bio-Rad). All reagents were used according to the manufacturer’s instructions.

Statistical analysis

All statistical analysis was performed using GraphPad Prism v7. Statistical significance was tested using a two-sided unpaired t test, unless otherwise specified in the figure legend. Statistical analysis of RNA-seq data was obtained through analysis in the cited R packages.


Fig. S1. Gating strategies for developing B cells.

Fig. S2. Cell cycle analysis of tet-Lin28b ABM pre-B and ImmB subsets.

Fig. S3. Determination of the minimal DOX treatment window during B cell maturation required for efficient CD5+ B-1 cell output from tet-Lin28b ABM.

Fig. S4. Barcode filtering and analysis strategy.

Fig. S5. Differentially expressed genes within top enriched Hallmark gene sets.

Fig. S6. Genetic evidence implicating Lin28b in the CD19/PI3K/c-Myc pathway.

Fig. S7. Transient tet-Lin28b expression of ABM progenitors promotes the output of functionally competent CD5+ B-1 cells.

Table S2. Antibodies used in this study.

Data file S1. Table S1: Raw data supplement.

Data file S2. VDJseq PerC B-1 CD5 slices.

Data file S3. DESeq2 normalized counts for adult tet-Lin28b and WT ImmB RNA-seq.

Data file S4. DESeq2 normalized counts for neonatal Lin28b+/+, Lin28b−/− ImmB RNA-seq.


Acknowledgments: We thank J. A. Daniel and T. P. Bender for their critical input. Funding: J.Y. was supported by the European Research Council (715313), Wallenberg Academy Fellows, the Swedish Research Council, StemTherapy, and the Wenner-Gren Foundation. J.Y. and E.J.G. were supported by the Swedish Cancer Society. M.S., J.Y., and S.S. were supported by the Knut and Alice Wallenberg Foundation. Author contributions: J.Y., S. Vanhee, and E.J.G. designed the study. S. Vanhee, E.J.G., H.Å., T.A.K., S.D., G.M., K.O., A.D., and S. Vergani performed the experiments. S. Vanhee performed the computational analyses together with S.L., S.S., and J.U. C.T.J., G.B., C.B., and M.S. provided critical expertise. J.Y. conceived the study and wrote the paper together with S. Vanhee. Competing interests: The authors declare that they no competing interests. Data and materials availability: The data for this study have been deposited in the database GSE135603.
View Abstract

Stay Connected to Science Immunology

Navigate This Article