Yves Denizot

Flexible Long-Range Loops in the VH Gene Region of the Igh Locus Facilitate the Generation of a Diverse Antibody Repertoire

The immunoglobulin heavy-chain (Igh) locus undergoes large-scale contraction in pro-B cells, which facilitates VH-DJH recombination by juxtaposing distal VH genes next to the DJH-rearranged gene segment in the 3 proximal Igh domain. By using high-resolution mapping of long-range interactions, we demonstrate that local interaction domains established the three-dimensional structure of the extended Igh locus in lymphoid progenitors. In pro-B cells, these local domains engaged in long-range interactions across the Igh locus, which depend on the regulators Pax5, YY1, and CTCF. The large VH gene cluster underwent flexible long-range interactions with the more rigidly structured proximal domain, which probably ensures similar participation of all VH genes in VH-DJH recombination to generate a diverse antibody repertoire. These long-range interactions appear to be an intrinsic feature of the VH gene cluster, because they are still generated upon mutation of the Eμ enhancer, IGCR1 insulator, or 3 regulatory region in the proximal Igh domain.

Self-Restrained B Cells Arise following Membrane IgE Expression

Among immunoglobulins (Igs), IgE can powerfully contribute to antimicrobial immunity and severe allergy despite its low abundance. IgE protein and gene structure resemble other Ig classes, making it unclear what constrains its production to thousand-fold lower levels. Whether class-switched B cell receptors (BCRs) differentially control B cell fate is debated, and study of the membrane (m)IgE class is hampered by its elusive inxa0vivo expression. Here, we demonstrate a self-controlled mIgE+ B cell stage. Primary or transfected mIgE+ cells relocate the BCRs into spontaneously internalized lipid rafts, lose mobility to chemokines, and change morphology. We suggest that combined proapoptotic mechanisms possibly involving Hax1 prevent mIgE+ memory lymphocyte accumulation. By uncoupling inxa0vivo IgE switching from cytokine and antigen stimuli, we show that these features are independent from B cell stimulation and instead result from mIgE expression per se. Consequently, few cells survive IgE class switching, which might ensure minimal long-term IgE memory upon differentiation into plasma cells.

Elevated plasma phospholipase A2 and platelet-activating factor acetylhydrolase activity in colorectal cancer

This clinical study reports that blood levels of the pro-inflammatory mediator platelet-activating factor (PAF) did not change in colorectal cancer patients. In contrast, plasma levels of two enzymatic activities, one implicated in PAF production (i.e. phospholipase A2) and one in PAF degradation (i.e. PAF acetylhydrolase activity) were significantly elevated.

Functional anatomy of the immunoglobulin heavy chain 3΄ super-enhancer needs not only core enhancer elements but also their unique DNA context

Abstract Cis-regulatory elements feature clustered sites for transcription factors, defining core enhancers and have inter-species homology. The mouse IgH 3΄ regulatory region (3’RR), a major B-cell super-enhancer, consists of four of such core enhancers, scattered throughout more than 25 kb of packaging ‘junk DNA’, the sequence of which is not conserved but follows a unique palindromic architecture which is conserved in all mammalian species. The 3’RR promotes long-range interactions and potential IgH loops with upstream promoters, controlling class switch recombination (CSR) and somatic hypermutation (SHM). It was thus of interest to determine whether this functional architecture also involves the specific functional structure of the super-enhancer itself, potentially promoted by its symmetric DNA shell. Since many transgenic 3’RR models simply linked core enhancers without this shell, it was also important to compare such a ‘core 3’RR’ (c3’RR) with the intact full-length super-enhancer in an actual endogenous IgH context. Packaging DNA between 3’RR core enhancers proved in fact to be necessary for optimal SHM, CSR and IgH locus expression in plasma cells. This reveals that packaging DNA can matter in the functional anatomy of a super-enhancer, and that precise evaluation of such elements requires full consideration of their global architecture.

3′RR targeting in lymphomagenesis: a promising strategy?

During precursor B-cell differentiation, heavy (H) and light chain genes of an immunoglobin (Ig) molecule are somatically assembled from germline DNA. This V(D)J recombination process occurs in the bone marrow prior to antigenic challenge (Fig. 1). In germinal centers, variable (V) regions become the target of somatic hypermutation (SHM) in activated B-cells generating high-affinity Ig (Fig. 1). In mature B-cells, class-switch recombination (CSR) deletes the constant (C) μ region and replaces it with a downstream CH gene. This enables B-cells to express various Ig isotypes without affecting antigen specificity (Fig. 1). Once activated, B-cells differentiate into Ig-secreting plasma cells. During B-cell development, V(D)J recombination, SHM, CSR and Ig synthesis are coupled with transcriptional accessibility of the IgH loci. IgH transcription is controlled by complex functional interactions of multiple enhancers, promoters and insulators spread among the 2.5 megabases of the locus. Among them, the 3′ regulatory region (3′RR) stands out as a major player during late stages of B-cell maturation (i.e., SHM, CSR and Ig synthesis).1,2 n n n nFigure 1. n nDNA recombination and mutations occur during B-cell maturation. RAG-induced (during VDJ recombination) and AID-induced (during CSR and SHM) DNA breaks are potential sites of oncogene translocations. The IgH 3′RR may act as an oncogene deregulator ... n n n nOngoing recombination and mutation all along B-cell development make the IgH locus a hotspot for translocations. Numerous lymphomas are marked by proto-oncogene translocation into the IgH locus. Cyclin D1 and Bcl-2 translocations, found respectively in mantle cell lymphoma and follicular lymphoma, take place during the V(D)J recombination. Cyclin D1/D3 or c-maf translocations observed in myeloma are obviously linked to CSR. c-myc translocation, the typical hallmark of Burkitt lymphoma, is linked to SHM or CSR. The mouse 3′RR, located downstream of the IgH Cα gene, shares a strong structural homology with the regulatory regions located downstream of each human IgH Cα gene (Cα1 and Cα2). Mouse models exploring the role of the 3′RR in B-cell physiology and in B-cell malignancies should provide useful indications about the pathophysiology of human B-cell proliferations. Convincing demonstration of the key contribution of the 3′RR in mature B-cell lymphomagenesis has been done by transgenic animal models. c-myc-3′RR transgenics developed Burkitt lymphoma-like proliferation,3 and the knock-in of a 3′RR cassette upstream of the endogenous c-myc gene induced B-cell lymphomas.4 Interestingly the phenotype of lymphoproliferations induced by the c-myc-3′RR transgene is affected by the presence of associated mutations in key cell cycle dependent genes such as p53 or Cdk4.5 n nKnock-out models have clarified the functions of the 3′RR as essential for high-rate IgH transcription at the plasma cell stage.2 3′RR may thus be a potent activator of IgH-translocated oncogene transcription, even when the breakpoints lie several hundred kb away from the 3′RR. Long-range interactions between the 2 regions of chromatin, through formation of a loop structure, constitute an important mechanism of normal and abnormal gene transcription regulation by the 3′RR. Studies have reported interactions between the 3′RR and the IgH variable region in normal and lymphomagenetic contexts. Therefore, targeted inhibition of the 3′RR could theoretically provides a therapeutic strategy for the treatment of a wide range of mature B-cell lymphomas. However, the first step before considering the 3′RR as a potential suitable target for anti-lymphoma pharmacological therapy is to demonstrate the innocuousness of such an approach, and notably the absence of potent adverse effect on normal immune and inflammatory B-cell mediated responses. Induction of negative alterations on the physiological anti-lymphoma immune or inflammatory networks would be obviously a counterproductive approach. We have recently tested this prerequisite by investigating the in vivo pristane-induced inflammatory response in 3′RR-deficient BALB/c and wt BALB/c mice.6 The lack of the 3′RR in BALB/c mice has no significant effect on the incidence, the kinetic of development and the cellular composition (IgM+IgD+ B-cells, CD4+ T cells, CD8+ T cells, monocyte/macrophage cells) of peritoneal ascites. Moreover ascite pro- and anti-inflammatory cytokine levels (IL-6, IL-21, IL-12/23, TNF-α IL-10, interferon-γ) are unaffected by the 3′RR-deficiency. Thus, a fully functional 3′RR is dispensable for the efficient recruitment of immune cells and the development of a normal inflammatory response in the in vivo pristane-induced inflammatory model. n nIn conclusion, the 3′RR is considered as a major lymphoma oncogene deregulator,3-5 and its deletion has no effect on immune and inflammatory responses in the pristane mouse model. It is, thus, tempting to speculate that the 3′RR might be considered as a potential suitable target for anti-lymphoma pharmacological therapy without significant impact on the normal immune and inflammatory networks. Previous results have reported that the 3′RR is a sensitive immunological target and that 3′RR activation and transcriptional activity can be altered by a diverse range of chemicals, including ones with anti-carcinogenic properties such as isothiocyanates,7 strengthening the hypothesis that altering 3′RR activity by chemicals could modulate the occurrence and severity of lymphomas. A limitation of the pristane mouse model is that inflammation is restricted to the peritoneal cavity. It is of evidence that other mouse models of inflammatory reactions must be tested before definitive validation of this hypothesis. Finally given the strong sequence homology between human and mouse 3′RR enhancers, mouse models could reveal useful tools for an in vivo study of lymphoma treatments based on IgH 3′RR down-regulation.

No evidence for a putative involvement of platelet-activating factor in systemic lupus erythematosus without active nephritis.

BACKGROUND: Platelet-activating factor (PAF) seems to be implicated in systemic lupus erythematosus (SLE) patients with associated renal diseases. AIMS: In this study, we ensured the role of PAF in SLE patients without renal complications. METHODS: Blood PAF and acetylhydrolase activity, plasma soluble phospholipase A(2), and the presence of antibodies against PAF were investigated in 17 SLE patients without active nephritis and in 17 healthy controls. RESULTS: Blood PAF levels were not different (p=0.45) between SLE patients (6.7+/-2.8 pg/ml) and healthy subjects (9.6+/-3.1 pg/ml). Plasma acetylhydrolase activity (the PAF-degrading enzyme) was significantly (p=0.03) elevated in SLE patients (57.8+/-6.4 nmol/min/ml) as compared with controls (37.9+/-2.6 nmol/min/ml). Plasma soluble phospholipase A(2) (the key enzyme for PAF formation) was not different (p=0.6) between SLE patients (59.1+/-5.1 U/ml) and controls (54.7+/-2.4 U/ml). Antibodies against PAF were detected only in 3/17 SLE patients. Flow cytometry analysis did not highlight PAF receptors on circulating leukocytes of SLE patients. CONCLUSION: This clinical study highlights no evidence for a putative important role of PAF in SLE patients without active nephritis.

Uracil-DNA glycosylase is not implicated in the choice of the DNA repair pathway during B-cell class switch recombination

Mature B-cells express membrane IgM and IgD (of same specificity) through alternative splicing of a pre-mRNA encompassing constant (C)μ and Cδ genes. After encountering antigen, Bcells undergo class switch recombination (CSR) that substitutes the Cμ gene with Cγ, Cε, or Cα, thereby generating IgG, IgE, and IgA antibodies with the same antigenic specificity but new effector functions. DNA-editing enzyme activation-induced deaminase (AID) is essential for CSR by targeting switch (S) regions preceding Cμ (namely, the Sμ donor region) and the Cγ, Cε, and Cα genes (namely, the Sγ,ε,α acceptor regions). 1, 2 CSR is controlled in cis by IgH locus super-enhancers and in trans by a wide spectrum of enzymes and proteins. 2 Among them, the role of the uracil DNA glycosylase (UNG) remains controversial. UNG is a key enzyme of base excision repair, which carries out faithful repair. Some authors estimate that during CSR, the UNG enzymatic activity removes the AID-induced dC to dU converted base of singlestrand DNA, generating abasic sites and leading to DNA strand breaks. For other authors, the role of UNG is to stabilize the S–S synapse and to recruit DNA repair factors that facilitate the endjoining process. 5 Thus, the classical non-homogenous end joining pathway would be increased over the alternative end joining (A-EJ) pathway in UNG-deficient mice, suggesting an intriguing role of UNG in promoting the A-EJ pathway. These results were based on the analysis of several Sμ–Sγ1 sequences obtained in UNG-deficient conditions due to technical limitations (conventional Sanger sequencing) to investigate S–S junction molecular signatures. We recently reported a new computational tool (CSReport) for automatic analysis of CSR junctions sequenced by high-throughput sequencing. We used this tool to analyze the rare Sμ–σδ junctions formed during IgD CSR 7, 8 and Sμ–Sγ3, Sμ–Sγ1, and Sμ–Sα junctions in wild-type (wt) mice. 9 We thus used CSReport and high-throughput sequencing to analyze the molecular signature of Sμ–Sγ3, Sμ–Sγ1, and Sμ–Sα junctions in UNG-deficient mice in detail. Our research has been approved by our local ethics committee review board (Comité Régional d’Ethique sur l’Expérimentation Animale du Limousin, Limoges, France) and carried out according to the European guidelines for animal experimentation. Wt mice and UNG-deficient mice (a gift from Dr. Tomas Lindahl, UK) were used. Single-cell suspensions of spleen cells were cultured for 4 days at 1 × 10 cells/ml in RPMI 1640 with 10% fetal calf serum and 5 μg/ml LPS, with (CSR toward IgG1) or without (CSR toward IgG3) addition of 20 ng/ml IL-4 or 2 ng/ml TGFβ (PeproTech, Rocky Hill, NJ) (CSR toward IgA). 9, 10 Splenocytes were washed in PBS and stained with various antibodies: anti-B220-PC5, anti-CD138APC, anti-IgM PECy7, anti-IgG3-FITC, anti-IgG1-FITC, and anti-IgAFITC. Cells were analyzed on a Fortessa LSR2 (Beckman Coulter). In parallel experiments, stimulated splenocyte DNA was extracted for investigation of Sμ–Sγ3, Sμ–Sγ1, and Sμ–Sα junctions. As previously described in detail, junctions were amplified with PCR. Libraries of 200 bp were prepared from the 1–2 kb PCR products of Sμ–Sγ1, Sμ–Sγ3, and Sμ–Sα amplified for Ion Proton sequencing (“GénoLim platform” of the Limoges University, France). Sequenced reads were then mapped to the Sμ, Sγ1, Sγ3, and Sαregions using BLAST algorithms. The computational tool developed for experiments performs junction assembly; identifies breakpoints in Sμ, Sγ1, Sγ3, and Sα; identifies junction structure (blunt, micro-homology, largehomology, or junction with insertions) and outputs a statistical summarization of the identified junctions. The UNG deficiency markedly reduced CSR toward IgG3 (mean 0.1 vs 6.9%), IgG1 (mean 1.9 vs 14.7%), and IgA (mean 1.4 vs 2.5%) compared to that of wt mice (Fig. 1a). After DNA extraction, the molecular signatures of the Sμ–Sγ1, Sμ–Sγ3, and Sμ–Sα junctions were investigated. The structural profiles of the Sμ–Sγ1, Sμ–Sγ3, and Sμ–Sα junctions (blunt, micro-homology, large-homology, or junction with insertions) for UNG-deficient and wt mice are shown in Fig. 1b. The distributions of the IgG1, IgG3, and IgA junctions in terms of distance from the forward PCR primer in Sμ and from one of the reverse primers in Sγ3, Sγ1, and Sα are reported in Fig. 1c. Localizations of breakpoints within AID hotspots (AGCT, WRCY, RGYW) and other motifs are shown in Fig. 1d (displayed along specifically targeted segments within S regions). The data indicated that if UNG deficiency markedly affected CSR efficiency, it did not significantly affect the pattern of blunt vs micro-/large-homology remaining junctions. Similarly, the positions of IgG3, IgG1, and IgA junctions (in term of distance from Sμ and Sγ1,γ3,α) and their colocalization with AID-attack motifs were not affected. Summarization of these 1568 independent junctions provided a description of the CSR sequences in UNG-deficient mice at an unprecedented level. To date, only a few had been reported after cloning and subsequent sequencing by Sanger’s method. In this study, the panel of wt junctions was markedly different (in terms of blunt vs micro-homology junctions) from ours and those of Panchakshari and colleagues. Our results undoubtedly demonstrate that CSR in UNG-deficient conditions did not affect the balance between N-HEJ and A-EJ. These results also showed that μ–γ1, μ–γ3, and μ-α CSR in UNG-deficient mice is regulated and that double-strand breaks for IgG1, IgG3, and IgA CSR are not random breaks. In conclusion, our results strengthen the hypothesis that during AID-induced CSR, UNG in association

3′RR and 5′E μ immunoglobulin heavy chain enhancers are independent engines of locus remodeling

By their impact on nuclear organization, enhancers are master regulators of cell fate. The immunoglobulin heavy chain (IgH) locus undergoes numerous changes as B cells differentiate. Among them, transcription and accessibility for V(D)J recombination, class switch recombination (CSR), and somatic hypermutation (SHM) are the most notable. The IgH locus carries two potent enhancers that are separated by 200 kb of distance. Eμ and the 3′ regulatory region (3′RR), at both ends of the constant gene cluster, control locus remodeling as B cells differentiate. Previous studies reported long-range interactions (of still unclear functional significance) between the Eμ and 3′RR enhancers during B-cell maturation. The question of a mutual transcriptional cross talk between these two enhancer entities remains open. We thus investigated if they were independent engines of locus remodeling or if their functions were more intimately intermingled. In this study, we developed ΔEμ-RAG-deficient and Δ3′RR-RAG-deficient mice to investigate the potential transcriptional cross talk between Eμ and 3′RR enhancers at the immature B-cell maturation stage. RAG-deficient mice, double Eμ-RAG-deficient mice, and double 3′RR-RAG-deficient mice were developed in our animal facility (free of specified pathogenic organisms). Our research was approved by our local ethics committee review board (Comité Régional d’Ethique sur l’Expérimentation Animale du Limousin, Limoges, France) and carried out according to the European guidelines for animal experimentation. Femoral pro-B cells were recovered with the EasySep mouse B-cell isolation Kit (STEMCELL Technologies, France), which was designed to isolate B cells from single-cell suspensions by negative selection. Cells from RAGdeficient, ΔEμ-RAG-deficient, and Δ3′RR-RAG-deficient mice (8–12 weeks old, males and females) were used. RNA was extracted using Trizol (ThermoFisher Scientific) according to the manufacturer’s instructions. Two pooled RNA (with four to six mice) were obtained for each genotype. RNA libraries were obtained using TruSeq Stranded Total RNA with Ribo-Zero Gold (Illumina), according to the manufacturer instructions. Libraries were sequenced on a NextSeq 500 sequencer, using NextSeq 500/ 550 High Output Kit (Illumina). Illumina NextSeq 500 paired-end 2 × 150 nt reads were mapped with STAR release v2.4.0a versus mm10 with gene model information from Ensembl release 77 with default parameters. RNAseq experiments were done in the genomics platform of Nice Sophia Antipolis, as previously reported. Data were deposited in Gene Expression Omnibus under the accession number, GSE117449. Femoral pro-B cells were isolated from RAG-deficient, ΔEμRAG-deficient, and Δ3′RR-RAG-deficient mice to investigate the potential transcriptional cross talk between Eμ and 3′RR enhancers in immature B cells. A schematic representation of the IgH locus is reported in Fig. 1a. Non-coding RNAs (ncRNAs) contribute to chromosomal looping. Among these ncRNAs, enhancer RNAs (eRNAs) are transcribed from enhancer DNA sequences, including the 3′RR, and contribute to their enhancer function. RNAseq experiments did not highlight any 3′RR eRNA in the pro-B cells of RAG mice (Fig. 1b), confirming results from a previous study with specific reverse trascription-quantitative PCR (RT-QPCR). The absence of 3′RR eRNA in pro-B cells is in agreement with studies reporting that 3′RR has no direct role in V(D)J recombination. As a positive control, 3′RR eRNA was evident in lipopolysaccharidestimulated B splenocytes (Fig. 1b). Excepted for Cμ (Fig. 1c), the RNAseq experiments showed no transcription in the Cγ, Cε, and Cα constant genes of the IgH locus (data not shown). If genomic deletion of the Eμ enhancer reduced sense transcription around its location (including Cμ transcription), genomic deletion of the 3′RR paradoxically enhanced both sense and especially antisense transcription of the D and J segments, as well as Eμ and Cμ transcription (Fig. 1c, d). Peak transcription levels were specifically found to originate from the DQ52 promotor (D4-1) and the Eμ enhancer (known as μo and Iμ sense transcripts, respectively). The concept of the 3′RR-mediated transcriptional silencing activity was first reported by Braikia et al., with RT-QPCR analysis. In contrast, with the present study, the μo and Iμ sense transcripts were not reportedly altered by the 3′RR deletion. Deletion of the 5′Eμ enhancer markedly lowered B-cell V(D)J recombination without affecting SHM and CSR. In contrast, deletion of the 3′RR enhancer affects B2 B-cell fate, SHM, and conventional CSR. If the 3′RR deletion also affects B1 B-cell fate and SHM, then it has no evident effect on B1 B-cell IgA CSR. If the roles of these two IgH enhancers have been observed during B-cell fate and maturation, then few data were available concerning their synergy, cooperation, and transcriptional cross talk. Analysis of chromatin marks, eRNA, and accessibility in the ΔEμ and Δ3′RR mice shows in mature activated B cells that the 3′

Pre-germinal center origin for mature mouse B cell lymphomas: a major discrepancy with human mature lymphomas.

During the different stages of their development, B lymphocytes undergo several genetic events such as V(D)J recombination, somatic hypermutation (SHM) and class switch recombination (CSR) in their immunoglobulin light (IgL) and heavy (IgH) chain loci. Ongoing recombination and mutations make the Ig loci hotspots for oncogene translocation. Indeed, many lymphomas are induced by oncogenes (bcl2, cyclin D1, cyclin D3, c-maf, c-myc. . .) translocations to the Ig loci. These loci are controlled by cis-regulatory elements that regulate both transcription of the Ig genes and induction of the genetic events, making them critical elements in the context of lymphomagenesis. Among them, the most important oncogen-driven cis-regulatory elements are the IgH 50Em enhancer, the 4 enhancers of the IgH 30 regulatory regions (30RR), and the Eik enhancer of the Igk locus. Although less studied, transcriptional enhancers are also present in the Igl locus. Animal models demonstrated that translocation of an oncogene under the transcriptional control of one of these enhancers promotes B cell malignancies. However, these translocations are not sufficient to induce transformation since tumors require additional “hits” to develop. The role of the activation-induced cytidine deaminase (AID) that target thousand loci on the B cell genome is suspected. In human mature B cell lymphomas both germinal center-associated genes such as bcl6 and the translocated oncogene bear stigmata of AID targeting. By using a mouse model (l-MYC mice) mimicking the translocation of a human c-myc under the transcriptional control of mouse Igl enhancers we have suggested that these additional hits would be influenced by the 30RR,10,11 through its AID-driven action during SHM, and CSR. Indeed, l-MYC-derived lymphomas exhibited a marked reduction of their degree of maturity in a 30RR-deficient background. To assess this hypothesis we have investigated in our collection of mouse lymphomas the occurrence of SHM in c-myc (both endogenous and translocated) and bcl6 genes, 2 well known AID-targeted genes in human B cell lymphomas. We first looked for mutations that could explain the transcriptional burst of the tranlocated c-myc in l-MYC mice. Immature (B220CCD117C) and mature (B220CIgMCIgDC) B cell lymphomas were investigated and compared with pre-malignant mature splenic B cells (B220CIgMC). Using high throughput sequencing (primers are located in Fig. 1A), we could not find a significant amount of mutation neither in the translocated human c-myc nor in the endogenous mouse c-myc both in immature and mature lymphomas; irrespectively of the presence of a functional 30RR (Fig. 1B). We next searched for mutations in bcl6, a germinal center associated protein frequently AID-mutated in human B cell lymphomas. Surprisingly, we found that bcl6 was unmutated in mouse B cell lymphomas (even with a mature phenotype) (Fig. 1B) and that bcl6 transcript levels were dramatically lowered compared to wt mature splenic B cell levels (Fig. 1C). The lack of a functional 30RR did not affect B cell lymphoma bcl6 transcript levels. Taken altogether these results evidenced no AID-induced stigmata in mouse B cell lymphomas. Strengthening this hypothesis, lowered AID transcript levels were found in both immature and mature mouse B cell lymphomas (with or without a 30RR) as compared to wt splenic B cells (Fig. 1D). The present data indicate that the marked reduction of the degree of maturity in B cell lymphomas of 30RR-deficient l-MYC mice is not linked to reduced AID-induced myc and AID-induced bcl6 mutations. The molecular mechanisms controlling this maturity shift remain unclear. Recently, the 30RR was reported to control m chain transcription and the B cell fate. Differences in BCR signaling (strength, timing and/or cooperation with accessory molecules) might be suggested to explain the change in lymphoma maturity in 30RR-deficient mice. Strengthening this hypothesis the class-specific BCR tonic signal was reported to modulate