Early B Cell Progenitors Deficient for GON4L Fail To Differentiate Due to a Block in Mitotic Cell Division

This article is featured in In This Issue, p.3763

B cell development sustains a pool of peripheral B cells that support Ab-mediated immunity. During the earliest stages of this process, a network of transcription factors and signaling pathways guide B cell progenitors through alternating phases of differentiation and proliferation (1–4). Differentiation requires the DNA-binding proteins E2A, EBF1, PAX5, and STAT5 (among others) (5–9), which form a transcriptional regulatory network that directs the formation of early B cell precursors. In the most primitive progenitors, E2A and EBF1 activate B-lineage genes (10–13), promoting specification toward a B cell fate (1, 2, 14, 15). EBF1 and PAX5 then activate additional B-lineage genes and repress others that promote alternative developmental programs, sealing commitment to a B cell fate (16–20). Additionally, the receptors c-Kit and FLT3 and that for IL-7 provide signals that are essential for the formation of early B cell progenitors (4).

The B cell transcription factor network and signaling pathways also control the proliferation of early-stage B cell precursors. A main driver of this process is IL-7 signaling, which activates the transcription factor STAT5 and the MAPK/ERK and PI3K signaling pathways (21), thereby promoting expression of proteins essential for survival and mitotic division. These include cyclin D3, which controls the G₁/S transition of the cell cycle and is essential for B cell development (22–24). Further, IL-7 signaling sustains expression of EBF1, which also activates mitotic genes (25–28). The roles of STAT5 and EBF1 in B cell development are well established (29–31), but less is known about pathways downstream of these proteins that control cell division by B cell progenitors.

In Justy mice, B cell development is blocked at an early stage due to a recessive point mutation in the Gon4-like (Gon4l) gene (32). This lesion disrupts splicing of Gon4l precursor mRNA in B cell progenitors, greatly reducing expression of full-length Gon4l transcript and protein. The function of GON4L is not understood, but studies in organisms ranging from plants to invertebrates to zebrafish have implicated this protein in pathways that control differentiation and cell division within developmental programs (33–37). For example, GON4L deficiency in zebrafish embryos blocks erythropoiesis, somite formation, and tail extension, which was correlated with cell cycle arrest and apoptosis (34, 37). Validating a role in cell division, other studies identified GON4L as important for the growth of cultured human cancer cells (38–40).

GON4L is a nuclear protein that is predicted to form domains characteristic of transcriptional regulators, including a highly acidic region, two paired amphipathic helix repeats, and a SANT-L domain (41). Further, molecular analysis showed that GON4L forms complexes with the transcriptional regulators YY1, SIN3A, and HDAC1, which have all been implicated in the regulation of cell division (41–45). Additionally, GON4L binds to NPAT, a transcriptional coactivator that regulates histone gene expression during DNA replication (46, 47), and to MCM3 and MCM4, components of the mini-chromosome maintenance complex required for DNA replication (37, 48). However, the importance of these interactions for GON4L function is poorly understood.

The findings outlined above suggest that GON4L is important for cell division during B cell development. Therefore, we determined how GON4L deficiency in B cell progenitors from Justy mice affected cell cycle progression, proliferation, and mitotic gene expression. In Justy B cell progenitors, the critical B-lineage transcription factor PAX5 was expressed normally, and the IL-7 signaling pathway was functional; nevertheless, these cells failed to proliferate. This proliferative arrest correlated with impaired cell cycle progression and DNA synthesis, as well as induction of apoptosis. Also, Justy B cell progenitors failed to activate genes needed for mitotic division. Enforced expression of proteins that regulate the G₁/S transition augmented B cell development from Justy cells, suggesting that GON4L is critical at this stage of the cell cycle. Together, our data indicate that GON4L regulates pathways that guide proliferation by early-stage B cell progenitors.

Mice were housed in a specific pathogen–free facility. Justy mice were described previously (32). p53-deficient mice (49) were from the Jackson Laboratory (Bar Harbor, ME). All procedures were approved by the University of Iowa Institutional Animal Care and Use Committee.

Abs were from BD Biosciences (San Jose, CA), BioLegend (San Diego, CA), eBioscience (San Diego, CA), and Cell Signaling Technology (Danvers, MA). Ab clones used were B220 (RA3-6B2), BP-1 (6C3), BrdU (BU20A), CD3 (145-2C11), CD4 (GK1.5), CD8a (53-6.7), CD11b/Mac-1 (M1/70), CD11c (N418), CD16/32 (2.4G2), CD19 (1D3), CD43 (S7), CD49b (DX5), CD127 (A7R34), CD135 (A2F10), Gr-1 (RB6-8C5), IgM (B76), Ly6C (AL-21), PAX5 (1H9), pSTAT5 (47), pERK (197G2), pAKT (D9E), and TER-119. Biotinylated IgH-specific Ab was obtained from Vector Laboratories (Burlingame, CA). Streptavidin Brilliant Violet 605 was obtained from BD Biosciences.

Flow cytometric analysis was performed as described previously (32). For sort purifications, target cells were enriched by negative selection by labeling cells with rat anti-mouse Abs specific for Gr-1, IgM, Ly6C, and TER-119 and selecting out Ab-bound cells using sheep anti-rat Dynabeads (Life Technologies, Carlsbad, CA). Recovered cells were stained for non-B lineage markers (CD3, CD4, CD8, CD11c, CD49b, Ly6C) and B-lineage markers (B220, BP-1, CD19, CD43, CD127, FLT3). Pre–pro-B cells (defined as Lin⁻B220⁺CD43⁺CD19⁻FLT3⁺CD127⁺) and pro-B cells (defined as Lin⁻B220⁺CD43⁺CD19⁺BP-1⁻FLT3^+/−CD127⁺) were isolated by two consecutive rounds of sort purification using a FACSAria. A representative gating strategy used for sort purification is shown in Supplemental Fig. 2.

Mice were injected with 1 mg of BrdU, and bone marrow was harvested 6 h later. Cultured cells were incubated for 2 h in 10 μM BrdU and harvested. Cells were stained for surface markers and analyzed as described previously (50). For DNA content analysis, cells were stained for surface markers and processed using buffers from BD Biosciences. Cells were incubated on ice for 15 min in Cytofix/Cytoperm, washed with Perm/Wash, and incubated on ice for 10 min in Cytoperm Plus. After washing with Perm/Wash, cells were incubated at room temperature for 5 min in Cytofix/Cytoperm and then stained with FxCycle Violet stain (Life Technologies) immediately prior to analysis.

RNA from two independent sort purifications was pooled and reverse transcribed using SuperScript III (Life Technologies). Quantitative real-time PCR was performed as described previously (32). Analysis of some cell cycle–associated genes was performed using the Mouse Cell Cycle PCR Array (QIAGEN, Germantown, MD).

Cells were digested with Proteinase K (Thermo Fisher Scientific, Pittsburgh, PA), and genomic DNA was recovered by precipitation. PCR primers used were described previously (51). Primer sequences were as follows: J_HR3: 5′-GTTCTAATGTCACCACAGACCAG-3′; V_H7183F: 5′-CGGTACCAAGAA(C/G)A(A/C)CCTGT(A/T)CCTGCAAATGA(C/G)C-3′; and V_HJ558F: 5′-CGAGCTC-TCCA(A/G)CACAGCCT(A/T)CATGCA(A/G)CTCA(A/G)C-3′.

All buffers were from BD Biosciences. Cells were stained for surface markers immediately prior to staining for intracellular proteins. To stain for PAX5, cells were incubated on ice for 15 min in Cytofix/Cytoperm, washed with Perm/Wash, incubated on ice for 10 min in Cytoperm Plus, washed with Perm/Wash, and then incubated again in Cytofix/Cytoperm at room temperature for 5 min. Cells were then washed with Perm/Wash, incubated on ice for 30 min in Perm/Wash containing PAX5 Ab, and washed again with Perm/Wash prior to analysis. To stain for pSTAT5, pERK, and pAKT cells were incubated at 37°C for 10 min in Phosflow Fix Buffer 1, washed with stain buffer (PBS containing 3% FBS), and incubated at −20°C for 30 min in Perm Buffer II. Cells were then washed with stain buffer, incubated on ice for 30 min in stain buffer containing Abs, and washed again with stain buffer prior to analysis.

Cells were stained for surface markers, resuspended in ice-cold PBS, and incubated on ice for 15 min. IL-7 was added to a final concentration of 400 ng/ml, followed by incubation at 37°C for 15 min. Cells were then fixed and processed for pSTAT5 staining, as described above.

Cytokines were from PeproTech (Rocky Hill, NJ). Multipotent progenitors (MPPs) were isolated as described previously (32). Cells were added to 24-well plates containing OP9 stromal cells in RPMI 1640 supplemented with 10% FBS, FLT3 ligand (FLT3L), IL-7, and stem cell factor (SCF). Cells were cultured for 7 d and then analyzed. For retroviral transduction, progenitor cells were added to OP9 cells in StemSpan Serum-Free Expansion Medium (STEMCELL Technologies, Vancouver, BC, Canada) containing FLT3L, IL-3, IL-11, SCF, and thrombopoietin and cultured for 24 h. Cells were then transferred to RetroNectin-coated plates containing retroviral particles. After 24 h, cells were transferred to wells containing OP9 cells in Opti-MEM (Life Technologies) supplemented with 5% FBS, FLT3L, IL-7, and SCF. Cells were cultured for 11 d and then analyzed.

Retroviral vectors were constructed using the plasmid pMIG (Addgene, Cambridge, MA) and cDNAs from GE Dharmacon (Lafayette, CO). pMIG vector and pCL-Eco (52) were transfected into 293T cells. After 24 h, supernatants were collected, filtered, and added to plates coated with RetroNectin (Clontech, Mountain View, CA).

Activated caspase levels were analyzed as described previously (50).

Data were analyzed using the unpaired, two-tailed Student t test in Prism software (GraphPad, San Diego, CA).

We previously showed that the Justy mutation causes a deficiency in GON4L in B cell progenitors, arresting B cell development at the pre–pro-B to pro-B cell transition (32). To better understand how this transition is affected, B cell progenitors from wild-type (WT) and Justy mice were separated into pre–pro-B and pro-B cell fractions based on the expression of FLT3 and CD19 (Fig. 1A). These markers were chosen because cells transitioning from the pre–pro-B to pro-B cell stage normally downregulate FLT3 and upregulate CD19 (53–56). As expected, FLT3⁺/CD19⁻ pre–pro-B cells and FLT3⁻/CD19⁺ pro-B cells were easily distinguishable in WT bone marrow. Additionally, some cells were detected that reached the pro-B stage, as characterized by CD19 and PAX5 expression (Supplemental Fig. 1A), but still expressed FLT3. We named these FLT3⁺ pro-B cells, which we reasoned are a transient intermediate between pre–pro-B and pro-B cells. In WT mice, low numbers of pre–pro-B and FLT3⁺ pro-B cells were observed, whereas the numbers of FLT3⁻ pro-B cells were much greater (Fig. 1B). In Justy mice, pre–pro-B and FLT3⁺ pro-B cell numbers were similar to WT mice, but the numbers of FLT3⁻ pro-B cells were greatly decreased (Fig. 1B). These data indicate that GON4L deficiency impairs accumulation of B cell progenitors at the FLT3⁻ pro-B cell stage.

The accumulation of FLT3⁻ pro-B cells in WT mice implies that this population is generated by a burst of proliferation. This idea is supported by previous studies that suggested that pro-B cells undergo division in vivo, as judged by DNA content profiles (23, 57). To assess proliferation by the B cell progenitor fractions that we defined based on FLT3 and CD19 expression, we analyzed DNA synthesis by injecting WT and Justy mice with the thymidine analog BrdU and measuring its incorporation into cells (Fig. 1C, 1D). In WT mice, FLT3⁺ and FLT3⁻ pro-B cells incorporated more BrdU than did pre–pro-B cells, suggesting that the former divide more rapidly. Compared with WT cells, Justy pre–pro-B and FLT3⁺ pro-B cells incorporated significantly less BrdU, indicating that DNA synthesis was impaired by GON4L deficiency. Surprisingly, no difference in BrdU incorporation was observed between WT and Justy FLT3⁻ pro-B cells. This may be attributable to selection for mutant cells that express enough GON4L to reach the FLT3⁻ pro-B cell stage but are still unable to accumulate or progress further in development.

We also analyzed the cell cycle distribution of B cell progenitors by staining for DNA content ex vivo (Fig. 1E, 1F). Less than 20% of WT pre–pro-B cells were in the S or G₂/M phases of the cell cycle, suggesting that these cells divide at a low rate. Compared with WT cells, significantly more Justy pre–pro-B cells were in the G₁ phase of the cell cycle, whereas fewer were in the S phase. A trend toward fewer Justy pre–pro-B cells in the G₂/M phase was also apparent. Consistent with high BrdU incorporation, most (∼60%) WT FLT3⁺ pro-B cells were in the S or G₂/M phase, indicating that these cells divide rapidly. As with pre–pro-B cells, significantly more Justy FLT3⁺ pro-B cells were in the G₁ phase relative to WT cells, and fewer were in the S or G₂/M phase relative to WT cells. Analysis of WT and Justy FLT3⁻ pro-B cells showed no differences in the frequencies of cells in the G₁ or S phase, but significantly fewer Justy cells were in the G₂/M phase, indicating that the mutation had an impact on this stage of the cell cycle. Collectively, these results suggest that GON4L deficiency disrupts the cell cycle in B cell progenitors as they make the transition from the pre-pro to the pro-B cell stage, thus preventing these cells from undergoing a proliferative burst that normally drives the accumulation of FLT3⁻ pro-B cells.

B cell progenitor proliferation requires IL-7 (21), which activates the STAT5, ERK, and PI3K signal-transduction pathways. Therefore, we analyzed the levels of phosphorylated (activated) STAT5, ERK, and AKT, with AKT being a surrogate for PI3K pathway activity. All were normal or elevated in cells from Justy mice that were directly analyzed ex vivo, indicating that the IL-7 signaling pathway is functional (Fig. 2A–F). To further assess responsiveness to IL-7, bone marrow cells from WT and Justy mice were incubated in the absence or presence of IL-7, after which the levels of pSTAT5 in B cell progenitors were analyzed (Fig. 2G). We found that exposure to IL-7 increased pSTAT5 levels in WT and Justy B cell progenitors, confirming that the IL-7 signaling pathway is functional in Justy cells.

Pro-B cell formation requires a transcription factor network that controls B-lineage commitment and cell division (1, 2, 31). Our previous study indicated that RNAs encoding key members of this transcription factor network (e.g., E2A, EBF1, and PAX5) are expressed at normal levels in Justy B cell progenitors (32). The DNA-binding protein PAX5 functions at the apex of the B-lineage transcription factor network, so we directly measured its expression in WT and Justy B cell progenitors (Supplemental Fig. 1A). In cells from WT mice, PAX5 expression was low in pre–pro-B cells, was much higher in FLT3⁺ pro-B cells, and was maintained at high levels in FLT3⁻ pro-B cells. Justy B cell progenitors showed the same expression pattern, and the absolute PAX5 levels in these cells were similar to WT cells. These data agree with other studies that showed that Justy B cell progenitors express normal levels of CD19 and other genes activated by the B-lineage transcription factor network (32).

PAX5 and other members of the B-lineage transcription factor network (e.g., E2A and STAT5) also promote Igh gene rearrangement in B cell progenitors (5, 58, 59), so this process was analyzed in WT and Justy pro-B cells. Similar to WT cells, Justy cells rearranged the Igh gene and expressed normal levels of IgH protein, which is only produced upon successful gene rearrangement (Supplemental Fig. 1B, 1C). These findings show that GON4L deficiency does not impair the regulation of PAX5 expression or Igh gene rearrangement and suggest that critical aspects of the B-lineage transcription factor network are functional in GON4L-deficient cells.

Prior studies suggested that GON4L coordinates cell division with cell differentiation (33–37), an idea in line with our data suggesting that GON4L-deficient pro-B cells fail to undergo a proliferative burst. Further, other studies suggested that GON4L drives proliferation by human cancer cells (38–40). Given these findings, we explored the possibility that GON4L deficiency due to the Justy mutation affects the expression of genes that control cell-cycle progression and division.

The cyclin proteins are key regulators of the cell cycle, so we measured expression of RNAs encoding these proteins in WT and Justy B cell progenitors using quantitative real-time (RT-PCR) analysis. RNA was isolated from sort-purified pre–pro-B and pro-B cells using the strategy shown in Supplemental Fig. 2. For pro-B cells, the FLT3⁻ and FLT3⁺ subsets were pooled to obtain enough cells from Justy mice. The D-type cyclins (cyclin D1, D2, and D3) were of particular interest because they control the G₁ and G₁/S phases of the cell cycle, although only cyclin D3 is essential for B cell development (23, 24). Correspondingly, Ccnd3 RNA expression was much higher in WT and Justy B cell progenitors relative to that for Ccnd1 or Ccnd2 (Fig. 3A). Compared with those in WT cells, Ccnd1 and Ccnd2 RNA levels were significantly elevated in Justy pro-B cells (Fig. 3A). More notably, Ccnd3 RNA levels were clearly and significantly decreased in Justy B cell progenitors and were not upregulated at the pro-B cell stage, as seen in WT cells. The E-type cyclins and cyclin A2 regulate the G₁/S and S phases of the cell cycle, whereas B-type cyclins and cyclin F control mitosis. Expression of RNAs for all of these proteins was significantly reduced in Justy cells, with the exception of Ccne1 and Ccne2, which was normal (Fig. 3B–E). Thus, expression of cyclins is reduced in GON4L-deficient B cell progenitors.

We also analyzed the expression of RNAs that encode E2F transcription factors, which are activated by cyclin D and E, and upregulate the genes needed for DNA replication and mitosis. E2f1 and E2f3 RNA levels were normal in Justy cells, but those for E2f2 and Tfdp1 were significantly decreased (Fig. 4A, 4B). E2F-regulated genes include those encoding cyclin A2 and D3, as well as those for E2F2 and DP1 themselves (60, 61). Expression of all of these genes was reduced in Justy cells (Figs. 3A, 3C, 4B), so we analyzed the expression of other E2F targets. RNA levels for some E2F targets were normal in Justy cells (data not shown), but those transcribed from the Myb, Rbl1, Skp2, and Rad21 genes (60–65), among several others, were significantly reduced (Fig. 4C). Further, in the case of Ki-67, RNA and protein expression was significantly lower in Justy cells compared with WT cells (Fig. 4D–F). These findings demonstrate that B cell progenitors deficient for GON4L express key cell cycle regulators at abnormally low levels.

The experiments described above analyzed WT and Justy B cell progenitors removed directly from bone marrow, an environment in which defective or dying B cell progenitors are likely rapidly eliminated. To assess the impact of GON4L deficiency in a simplified setting, we tracked pre–pro-B and pro-B cell development in a culture system. MPPs were isolated from bone marrow and cultured with stromal cells and cytokines for several days. Afterward, B cell progenitor development and accumulation were assessed by flow cytometry (Fig. 5A). WT cultures contained few pre–pro-B and FLT3⁺ pro-B cells but had large numbers of FLT3⁻ pro-B cells, reflecting normal differentiation from the pre–pro-B to pro-B cell stage (Fig. 5B, 5C). In comparison, Justy cultures generated abnormally high numbers of pre–pro-B cells, fewer FLT3⁺ pro-B cells, and dramatically less FLT3⁻ pro-B cells, indicating that the pre–pro-B to pro-B cell transition was disrupted. Similar to results from our ex vivo analysis, pSTAT5 levels were normal in Justy B cell progenitors generated in culture, confirming that IL-7 signaling was functional (Fig. 5D). Additionally, cultured Justy pro-B cells expressed significantly less Ccnd3 RNA relative to WT cells (Fig. 5E), as was observed in cells isolated directly from bone marrow. Thus, the Justy phenotype was faithfully reproduced in the culture system.

We next assessed BrdU incorporation by B cell progenitors generated in culture (Fig. 6A, 6B). Similar to findings from in vivo labeling studies (Fig. 1C, 1D), Justy pre–pro-B cells incorporated significantly less BrdU relative to WT cells. This was also observed for FLT3⁺ pro-B cells; however, in this case, the difference between WT and Justy cells was much more dramatic. BrdU incorporation by Justy FLT3⁻ pro-B cells was modestly, but still significantly, lower as well. Therefore, GON4L deficiency impairs DNA synthesis by B cell progenitors produced in vivo and in vitro.

The cell cycle profiles of cultured B cell progenitors were visualized by staining for DNA content (Fig. 6C, 6D). In this case, FLT3⁺ and FLT3⁻ pro-B cells were pooled for the analysis. Consistent with a decrease in BrdU incorporation, significantly fewer Justy pre–pro-B and pro-B cells were in the S phase. More strikingly, pre–pro-B and pro-B cells in Justy cultures had much greater numbers of cells with a sub-G₁ (<2N) DNA content, indicating DNA fragmentation and apoptosis (Fig. 6C, 6E). Given these results, we analyzed levels of activated caspase 8, an initiator of proapoptotic pathways (Fig. 6F, 6G). WT and Justy pre–pro-B cells contained low levels of activated caspase 8, but the amount detected in Justy cells was significantly increased. FLT3⁺ and FLT3⁻ pro-B cells contained higher levels of activated caspases 8, and these were significantly elevated in Justy cells. Activated levels of the proapoptotic caspases 3 and 7 were also elevated in Justy cells (data not shown). Combined, these findings indicate that deficiency in GON4L in B cell progenitors leads to the induction of proapoptotic pathways, which helps to explain why Justy pro-B cells fail to accumulate in bone marrow or in culture.

Given that Justy B cell progenitors are prone toward apoptosis, we asked how B cell development from Justy cells was affected by promoting survival. In zebrafish, defects caused by GON4L deficiency were partially rescued by suppressing p53 expression (34, 37). In Justy mice, however, Trp53 deficiency failed to restore B cell development (Supplemental Fig. 3A, 3B). We next asked whether enforced expression of prosurvival factors could rescue Justy pro-B cell development in vitro. WT and Justy MPPs were transduced with a control retroviral vector or that expressing BCL-2 or BCL-x_L and then cultured for several days. Afterward, the frequencies of pro-B cells within the Mac-1⁻Gr-1⁻GFP⁺ fraction were determined, with GFP expression marking transduced cells. Enforced expression of BCL-2 or BCL-x_L had no effect on pro-B cell accumulation in WT cultures and suppressed, rather than augmented, pro-B cell formation in Justy cultures (Supplemental Fig. 3C–E). Thus, the effects of GON4L deficiency in B cell progenitors cannot be overcome by promoting survival.

Our data showed that Ccnd3 RNA expression is decreased in Justy B cell progenitors and that events controlled by cyclin D3 (e.g., activation of E2F expression) are impaired (Figs. 3, 4). Therefore, we asked how pro-B cell development in vitro was affected by increasing cyclin D3 expression. MPPs were transduced with a control retroviral vector or that expressing cyclin D3, and the frequencies of pro-B cells within the Mac-1⁻Gr-1⁻GFP⁺ (transduced) populations were determined several days later. For WT cultures, enforcing cyclin D3 expression significantly reduced pro-B cell development (Fig. 7A, 7B). In contrast, enforced expression of cyclin D3 in Justy cells greatly increased the frequency of pro-B cells (Fig. 7A, 7C). Determining total cell yields confirmed that the changes in pro-B cell frequencies caused by cyclin D3 overexpression correlated with differences in the absolute numbers of pro-B cells generated (Supplemental Fig. 4). We then asked how enforced expression of other cyclins affected pro-B cell development (Fig. 7A–C). Retroviral-mediated expression of cyclin E1 had no effect on pro-B cell development from WT MPPs, but it dramatically increased that from Justy cells. In contrast, enforced expression of cyclin A2 or cyclin B1 had no effect on pro-B cell formation from WT or Justy MPPs.

Justy pro-B cells expressed significantly less E2f2 RNA, so we asked how enforcing E2F2 expression affected pro-B cell development. Retroviral E2F2 expression slightly decreased pro-B cell development from WT cells but significantly increased that from Justy cells (Fig. 7D, 7E). In this case, the effect was less dramatic than that caused by overexpression of cyclin D3 or E1. These data demonstrate that impairment of pro-B cell development caused by GON4L deficiency is overcome by increasing the expression of factors important for the G₁/S and S phases of the cell cycle.

We defined how GON4L deficiency affected the differentiation and proliferation of B cell progenitors. Our findings show that GON4L-deficient cells express the critical B-lineage transcription factor PAX5, undergo Igh gene rearrangement, and can signal through IL-7R. However, these cells had defects in cell cycle progression and DNA synthesis and underwent apoptosis. Further, GON4L deficiency impaired activation of cell cycle–associated genes. Transgenic expression of proteins that promote the G₁/S transition overcame the block in B cell development, allowing GON4L-deficient pro-B cells to accumulate. Combined, our data suggest that GON4L regulates proliferation by B cell progenitors, likely by acting during the early phases of the cell cycle.

In zebrafish embryos, loss of GON4L disrupts erythropoiesis and the formation of somites, with the latter giving rise to the vertebral column and other tissues (34, 37, 66). For both defects, GON4L deficiency prevented expression of transcription factors required for the respective developmental pathway (i.e., GATA-1 for erythropoiesis, MyoD for somite formation). These results suggest that GON4L is required to activate the gene-expression programs that drive erythropoiesis and somite formation. In contrast, our data suggest that GON4L is not required to put central aspects of the B cell developmental program in motion, because Justy B cell progenitors express PAX5 at normal levels and undergo Igh gene rearrangement. This conclusion is also supported by data from our previous study, which showed that critical B-lineage genes (e.g., Ebf1, Cd19, Cd79a, and Cd79b) are expressed at normal levels in Justy B cell progenitors (32). Thus, our findings argue that GON4L-deficient pro-B cells can establish specific parts of the B-lineage gene program but fail to develop further, potentially as a result of defects in downstream pathways that affect gene expression and cell division.

The B cell transcription factor network is critical for proliferation by B cell progenitors. Key players are IL-7 signaling, STAT5, and EBF1 (21, 27). At the cellular level, the developmental arrest associated with GON4L deficiency closely resembles that caused by the absence of IL-7 signaling or EBF1 (7, 25). However, in Justy cells, IL-7 signaling and Ebf1 mRNA expression are normal (32). Of note, proliferation and development of B cell progenitors deficient for IL-7R can be partially rescued by enforced expression of EBF1 (25). However, overexpression of EBF1 in Justy progenitors had only a slight effect on B cell development (J.Y. Barr and J.D. Colgan, unpublished observations), suggesting that defects due to GON4L deficiency are distinct from those caused by disruption of IL-7 signaling. Together, these findings suggest that GON4L acts downstream of IL-7 signaling, STAT5, and EBF1.

GON4L deficiency in zebrafish embryos causes a G₂/M arrest in primitive erythroid progenitors (34). A similar, but more modest, G₂/M arrest appears to occur in cells within the tail of the embryo, which fails to undergo extension in the absence of GON4L (37). Consistent with a defect at the G₂/M phase, apoptosis in specific tissues is dramatically increased in GON4L-deficient zebrafish embryos (37). These defects were attributed to disruption of cell cycle regulation and activation of p53-dependent pathways for G₂/M arrest or apoptosis, a conclusion supported by data showing that p53 knockdown in zebrafish embryos partially rescues erythropoiesis, somite formation, and tail extension and also suppresses apoptosis (34, 37). One implication of these findings is that the absence of GON4L causes DNA damage or some other genotoxic stress that induces p53-dependent pathways.

Consistent with the findings described above, our data show that GON4L deficiency in B cell progenitors disrupts the cell cycle and causes apoptosis. However, our results suggest that p53-dependent pathways are not an important factor in the defects caused by GON4L deficiency in B cell progenitors. It remains unclear why different effects were observed in our studies and those of zebrafish. One possible explanation is that GON4L-deficient B cell progenitors undergo apoptosis via p53-independent pathways that supersede p53 activation and G₂/M arrest, thus lessening the role of p53.

Our data suggest that GON4L acts at the G₁/S phase of the cell cycle. Particularly compelling are our results showing that pro-B cell development from Justy cells is partially rescued by enforced expression of the G₁/S regulators cyclin D3, E, and E2F2. Previous studies demonstrated that cyclin D3 is essential for B cell development (23, 24), but the role of the E-type cyclins has not been addressed. We imagine that enforced expression of cyclin D3 or cyclin E affects similar, or overlapping, pathways that promote progression through the G₁/S transition. In support of this conclusion, defects caused by cyclin D1 deficiency are rescued by cyclin E (67), indicating functional overlap between the D- and E-type cyclins. Enforced E2F2 expression had a more modest effect relative to that caused by overexpressing cyclin D3 or E. This result implies that the function of GON4L goes beyond acting as a cofactor for E2F2. More likely is the possibility that GON4L controls proliferation by functioning together with (or in parallel to) the G₁/S cyclins.

Acute B-lymphoblastic leukemia (B-ALL) originates from early-stage B cell progenitors (68). The generation and growth of B-ALL cells likely rely on the proliferative capacity of B cell progenitors. Our data argue that GON4L is a key component of the machinery that controls B cell progenitor proliferation. Given this, it would be of interest to determine the importance of GON4L for the transformation and expansion of cells that give rise to B-ALL. A central role for GON4L in B-ALL, and a better understanding of how GON4L functions, could lead to novel therapeutic strategies for treating B-ALL.

We thank Dr. Mike Knudson (University of Iowa) for providing the BCL-2–expressing retroviral vector. We also thank the University of Iowa Flow Cytometry Facility and the Genomics Division of the University of Iowa Institute of Human Genetics for technical assistance and advice.

The authors have no financial conflicts of interest.